Gtx 980 vs gtx 980ti: GTX 980 Ti vs GTX 980 Game Performance Benchmarks (i7-4790K vs i7-4770K)

NVIDIA GeForce GTX 970 vs NVIDIA GeForce GTX 980 Ti

The following benchmarks stem from our benchmarks of review laptops. The performance depends on the used graphics memory, clock rate, processor, system settings, drivers, and operating systems. So the results don’t have to be representative for all laptops with this GPU. For detailed information on the benchmark results, click on the fps number.

For more games that might be playable and a list of all games and graphics cards visit our Gaming List

log 29. 23:00:29

#0 checking url part for id 6279 +0s … 0s

#1 checking url part for id 7325 +0s … 0s

#2 not redirecting to Ajax server +0s … 0s

#3 did not recreate cache, as it is less than 5 days old! Created at Tue, 27 Sep 2022 17:27:11 +0200 +0s … 0s

#4 linkCache_getLink using $NBC_LINKCACHE +0.044s … 0.045s

#5 linkCache_getLink using $NBC_LINKCACHE +0. 004s … 0.049s

#6 linkCache_getLink using $NBC_LINKCACHE +0s … 0.049s

#7 linkCache_getLink using $NBC_LINKCACHE +0s … 0.049s

#8 linkCache_getLink using $NBC_LINKCACHE +0s … 0.049s

#9 linkCache_getLink using $NBC_LINKCACHE +0s … 0.049s

#10 linkCache_getLink using $NBC_LINKCACHE +0s … 0.049s

#11 composed specs +0s … 0.049s

#12 did output specs +0s … 0.049s

#13 start showIntegratedCPUs +0s … 0.049s

#14 linkCache_getLink using $NBC_LINKCACHE +0.031s … 0.079s

#15 linkCache_getLink using $NBC_LINKCACHE +0s … 0.08s

#16 getting avg benchmarks for device 6279 +0.001s … 0.08s

#17 linkCache_getLink using $NBC_LINKCACHE +0.001s … 0.081s

#18 got single benchmarks 6279 +0.008s … 0.089s

#19 getting avg benchmarks for device 7325 +0s … 0.089s

#20 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.089s

#21 got single benchmarks 7325 +0.006s … 0.095s

#22 got avg benchmarks for devices +0s … 0.095s

#23 linkCache_getLink no uid found +0s … 0.095s

#24 linkCache_getLink no uid found +0s … 0.095s

#25 linkCache_getLink no uid found +0s … 0.096s

#26 linkCache_getLink no uid found +0s … 0.096s

#27 linkCache_getLink no uid found +0s … 0.096s

#28 linkCache_getLink no uid found +0s … 0.097s

#29 linkCache_getLink no uid found +0s … 0.097s

#30 linkCache_getLink no uid found +0s … 0.097s

#31 linkCache_getLink no uid found +0s … 0.097s

#32 linkCache_getLink no uid found +0s … 0.098s

#33 linkCache_getLink no uid found +0s … 0.098s

#34 linkCache_getLink no uid found +0s … 0.098s

#35 linkCache_getLink no uid found +0s … 0.098s

#36 linkCache_getLink no uid found +0s . .. 0.098s

#37 linkCache_getLink no uid found +0s … 0.098s

#38 linkCache_getLink no uid found +0s … 0.099s

#39 linkCache_getLink no uid found +0s … 0.099s

#40 linkCache_getLink no uid found +0s … 0.099s

#41 linkCache_getLink no uid found +0s … 0.099s

#42 linkCache_getLink no uid found +0s … 0.1s

#43 linkCache_getLink no uid found +0s … 0.1s

#44 linkCache_getLink no uid found +0.001s … 0.1s

#45 linkCache_getLink no uid found +0s … 0.1s

#46 linkCache_getLink no uid found +0s … 0.1s

#47 linkCache_getLink no uid found +0s … 0.101s

#48 linkCache_getLink no uid found +0s … 0.101s

#49 linkCache_getLink no uid found +0s … 0.101s

#50 linkCache_getLink no uid found +0s … 0.101s

#51 linkCache_getLink no uid found +0s … 0.101s

#52 linkCache_getLink no uid found +0s . .. 0.102s

#53 linkCache_getLink no uid found +0s … 0.102s

#54 linkCache_getLink no uid found +0s … 0.102s

#55 linkCache_getLink no uid found +0s … 0.103s

#56 linkCache_getLink no uid found +0s … 0.103s

#57 linkCache_getLink no uid found +0s … 0.104s

#58 linkCache_getLink no uid found +0s … 0.104s

#59 linkCache_getLink no uid found +0.001s … 0.105s

#60 linkCache_getLink no uid found +0s … 0.105s

#61 linkCache_getLink no uid found +0s … 0.105s

#62 linkCache_getLink no uid found +0s … 0.106s

#63 linkCache_getLink no uid found +0s … 0.106s

#64 linkCache_getLink no uid found +0s … 0.106s

#65 linkCache_getLink no uid found +0s … 0.107s

#66 linkCache_getLink no uid found +0s … 0.107s

#67 linkCache_getLink no uid found +0s … 0.107s

#68 linkCache_getLink no uid found +0s . .. 0.108s

#69 linkCache_getLink no uid found +0s … 0.108s

#70 linkCache_getLink no uid found +0s … 0.109s

#71 linkCache_getLink no uid found +0s … 0.109s

#72 linkCache_getLink no uid found +0s … 0.109s

#73 linkCache_getLink no uid found +0s … 0.11s

#74 linkCache_getLink no uid found +0s … 0.11s

#75 linkCache_getLink no uid found +0s … 0.11s

#76 linkCache_getLink no uid found +0s … 0.111s

#77 linkCache_getLink no uid found +0s … 0.111s

#78 linkCache_getLink no uid found +0s … 0.111s

#79 linkCache_getLink no uid found +0s … 0.111s

#80 linkCache_getLink no uid found +0s … 0.112s

#81 linkCache_getLink no uid found +0s … 0.112s

#82 linkCache_getLink no uid found +0s … 0.112s

#83 linkCache_getLink no uid found +0s … 0.113s

#84 linkCache_getLink no uid found +0s . .. 0.113s

#85 linkCache_getLink no uid found +0s … 0.114s

#86 linkCache_getLink no uid found +0s … 0.114s

#87 linkCache_getLink no uid found +0s … 0.114s

#88 linkCache_getLink no uid found +0s … 0.115s

#89 min, max, avg, median took s +0s … 0.115s

#90 before gaming benchmark output +0s … 0.115s

#91 Got 475 rows for game benchmarks. +0.014s … 0.129s

#92 composed SQL query for gamebenchmarks +0s … 0.129s

#93 linkCache_getLink using $NBC_LINKCACHE +0s … 0.129s

#94 linkCache_getLink using $NBC_LINKCACHE +0s … 0.129s

#95 linkCache_getLink using $NBC_LINKCACHE +0s … 0.129s

#96 linkCache_getLink using $NBC_LINKCACHE +0s … 0.129s

#97 linkCache_getLink using $NBC_LINKCACHE +0s … 0.129s

#98 linkCache_getLink using $NBC_LINKCACHE +0s … 0.129s

#99 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.129s

#100 linkCache_getLink using $NBC_LINKCACHE +0s … 0.129s

#101 linkCache_getLink using $NBC_LINKCACHE +0s … 0.129s

#102 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#103 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#104 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#105 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#106 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#107 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#108 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#109 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#110 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#111 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#112 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#113 linkCache_getLink using $NBC_LINKCACHE +0s … 0. 13s

#114 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#115 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#116 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#117 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#118 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#119 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#120 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#121 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#122 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#123 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#124 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#125 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#126 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#127 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#128 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.13s

#129 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#130 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#131 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#132 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#133 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#134 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#135 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#136 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#137 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#138 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#139 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#140 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#141 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#142 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#143 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.13s

#144 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#145 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#146 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#147 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#148 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#149 linkCache_getLink using $NBC_LINKCACHE +0s … 0.13s

#150 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#151 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#152 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#153 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#154 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#155 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#156 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#157 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.131s

#158 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#159 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#160 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#161 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#162 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#163 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#164 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#165 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#166 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#167 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#168 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#169 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#170 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#171 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.131s

#172 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#173 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#174 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#175 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#176 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#177 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#178 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#179 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#180 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#181 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#182 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#183 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#184 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#185 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.131s

#186 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#187 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#188 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#189 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#190 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#191 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#192 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#193 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#194 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#195 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#196 linkCache_getLink using $NBC_LINKCACHE +0s … 0.131s

#197 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#198 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#199 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.132s

#200 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#201 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#202 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#203 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#204 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#205 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#206 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#207 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#208 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#209 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#210 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#211 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#212 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#213 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.132s

#214 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#215 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#216 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#217 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#218 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#219 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#220 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#221 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#222 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#223 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#224 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#225 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#226 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#227 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.132s

#228 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#229 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#230 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#231 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#232 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#233 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#234 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#235 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#236 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#237 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#238 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#239 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#240 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#241 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.132s

#242 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#243 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#244 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#245 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#246 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#247 linkCache_getLink using $NBC_LINKCACHE +0s … 0.132s

#248 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#249 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#250 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#251 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#252 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#253 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#254 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#255 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.133s

#256 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#257 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#258 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#259 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#260 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#261 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#262 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#263 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#264 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#265 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#266 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#267 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#268 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#269 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.133s

#270 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#271 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#272 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#273 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#274 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#275 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#276 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#277 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#278 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#279 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#280 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#281 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#282 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#283 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.133s

#284 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#285 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#286 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#287 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#288 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#289 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#290 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#291 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#292 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#293 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#294 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#295 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#296 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#297 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.133s

#298 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#299 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#300 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#301 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#302 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#303 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#304 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#305 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#306 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#307 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#308 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#309 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#310 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#311 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.133s

#312 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#313 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#314 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#315 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#316 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#317 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#318 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#319 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#320 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#321 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#322 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#323 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#324 linkCache_getLink using $NBC_LINKCACHE +0s … 0.133s

#325 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.134s

#326 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#327 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#328 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#329 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#330 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#331 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#332 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#333 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#334 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#335 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#336 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#337 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#338 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#339 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.134s

#340 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#341 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#342 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#343 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#344 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#345 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#346 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#347 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#348 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#349 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#350 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#351 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#352 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#353 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.134s

#354 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#355 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#356 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#357 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#358 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#359 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#360 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#361 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#362 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#363 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#364 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#365 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#366 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#367 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.134s

#368 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#369 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#370 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#371 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#372 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#373 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#374 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#375 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#376 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#377 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#378 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#379 linkCache_getLink using $NBC_LINKCACHE +0s … 0.134s

#380 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#381 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.135s

#382 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#383 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#384 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#385 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#386 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#387 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#388 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#389 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#390 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#391 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#392 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#393 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#394 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#395 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.135s

#396 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#397 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#398 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#399 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#400 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#401 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#402 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#403 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#404 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#405 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#406 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#407 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#408 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#409 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.135s

#410 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#411 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#412 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#413 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#414 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#415 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#416 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#417 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#418 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#419 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#420 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#421 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#422 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#423 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.135s

#424 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#425 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#426 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#427 linkCache_getLink using $NBC_LINKCACHE +0s … 0.135s

#428 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#429 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#430 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#431 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#432 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#433 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#434 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#435 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#436 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#437 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.136s

#438 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#439 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#440 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#441 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#442 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#443 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#444 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#445 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#446 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#447 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#448 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#449 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#450 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#451 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.136s

#452 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#453 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#454 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#455 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#456 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#457 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#458 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#459 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#460 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#461 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#462 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#463 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#464 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#465 linkCache_getLink using $NBC_LINKCACHE +0s . .. 0.136s

#466 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#467 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#468 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#469 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#470 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#471 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#472 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#473 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#474 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#475 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#476 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#477 linkCache_getLink using $NBC_LINKCACHE +0s … 0.136s

#478 got data and put it in $dataArray +0.003s … 0.14s

#479 benchmarks composed for output. +0.064s . .. 0.204s

#480 calculated avg scores. +0s … 0.204s

#481 linkCache_getLink using $NBC_LINKCACHE +0.001s … 0.206s

#482 return log +0.002s … 0.208s

GTX 980 Ti Linux CUDA performance vs Titan X and GTX 980

Always look at the date when you read a hardware article. Some of the content in this article is most likely out of date, as it was written on June 12, 2015. For newer information, see our more recent articles.

The latest addition to the NVIDIA GeForce line, the GTX 980 Ti, is a significant improvement over the GTX 980 for single precision compute tasks using CUDA. It rivals, and is very similar to, the astounding performance of the of the Titan X. The most significant difference is that the GTX 980 Ti’s 6GB of memory is half that of the Titan X’s 12GB (but 2GB more than the 980’s 4GB). For computational work the extra memory of the Titan X may be important to you. It can make life easier when you are working on new code and haven’t optimized the memory buffering to keep the GPU loaded. In general, more memory is always a good thing. The Titan X is so fast for single precision calculations that the extra memory can help keep large workloads flowing. However, many CUDA accelerated programs have excellent memory buffering from host to card and 6GB will be enough to keep those CUDA cores loaded.

I’ll be adding the GTX 980 Ti to my GPU computing testing results from here out. I should have a post up on some Molecular Modeling applications soon. I’ve been working with GROMACS lately and I suspect that the 980 Ti may be the card of choice for this. For now I just want to show a simple CUDA benchmark on Linux so you can see what to expect for serious compute performance from the 980 Ti.

For some reason the 980 Ti was more of a problem to get working under Linux than the Titan X when it was first released. The Titan X «just worked» under Linux even before there was a functioning Windows driver! The 980 Ti required the latest NVIDIA beta driver!

System setup;

Puget Systems Peak Tower

• Dual 10 core Xeon 2687W v3 @ 3. 1 GHz
• 64GB DDR4-2133 Reg ECC
• NVIDIA Titan X 12GB X16 1000MHz core clock 3072 cuda cores
• NVIDIA GTX 980 Ti 6GB X16 1000MHz core clock 2816 cuda cores
• NVIDIA GTX 980 4GB X16 1126MHz core clock 2048 cuda cores
• Linux CentOS 7
• NVIDIA kernel driver modules version 352.09 (beta)

Make careful note of the NVIDIA driver version! I had to use the most recent beta driver to get the GTX 980 Ti to work. That means I had to use the downloaded shell archive .run file which means the driver install is outside of the normal package management and all of the usual difficulties with that situation apply. I generally prefer to have the NVIDIA driver installed from the «package managed» cuda repo.

The driver I used was,

http://us.download.nvidia.com/XFree86/Linux-x86_64/352.09/NVIDIA-Linux-x86_64-352.09.run

I suggest that you look at my post Install NVIDIA CUDA on Fedora 22 with gcc 5.1 for an approach you might want to consider taking to get CUDA and the beta driver installed. [ I do a CUDA install from the NVIDIA repo that fails with the driver install (but sets up everything for it and CUDA) and then install the NVIDIA shell archive .run file — crazy but simple and effective! ]

The NVIDIA display driver from the current cuda 7.0-28 repo does not support the GTX 980 Ti. I had cuda 7 installed on the system already when I added the 980 Ti to the mix. It is device 1 in the following nvidia-smi output.

[kinghorn@tower Downloads]$ nvidia-smi 
Mon Jun  1 15:31:35 2015       
+------------------------------------------------------+                       
| NVIDIA-SMI 346.46     Driver Version: 346.46         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX 980     Off  | 0000:03:00. 0     N/A |                  N/A |
| 26%   37C    P8    N/A /  N/A |    116MiB /  4095MiB |     N/A      Default |
+-------------------------------+----------------------+----------------------+
|   1  ERR!                Off  | 0000:04:00.0     N/A |                  N/A |
| 22%   36C    P8    N/A /  N/A |     20MiB /  6143MiB |     N/A      Default |
+-------------------------------+----------------------+----------------------+
|   2  Graphics Device     Off  | 0000:81:00.0     Off |                  N/A |
| 22%   31C    P8    15W / 250W |     23MiB / 12287MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

Note: Device 2 is the Titan X which shows up as «Graphics Device» with the 346.46 display driver. Device 1 with the «ERR!» is the GTX 980 Ti.

After the 352.09 beta driver install we have,

[kinghorn@tower ~]$ nvidia-smi 
Mon Jun  1 15:44:00 2015       
+------------------------------------------------------+                       
| NVIDIA-SMI 352. 09     Driver Version: 352.09         |                       
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX 980     Off  | 0000:03:00.0     N/A |                  N/A |
| 27%   46C    P2    N/A /  N/A |     81MiB /  4095MiB |     N/A      Default |
+-------------------------------+----------------------+----------------------+
|   1  Graphics Device     Off  | 0000:04:00.0     N/A |                  N/A |
| 22%   45C    P8    N/A /  N/A |     20MiB /  6143MiB |     N/A      Default |
+-------------------------------+----------------------+----------------------+
|   2  GeForce GTX TIT...  Off  | 0000:81:00.0     Off |                  N/A |
| 22%   39C    P8    16W / 250W |     23MiB / 12287MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

Yea! The GTX 980 Ti, device 1, is working and now has the distinguished title «Graphics Device» and the Titan X gets it’s proper name.

For a quick test of the CUDA single precision floating point performance here’s the results from running benchmark mode with the nbody simulation from the cuda samples.

The GTX 980 Ti is going to be a great card for single precision compute loads!

Following is some more detailed output from the nbody simulation and deviceQuery output for your enjoyment 🙂

[kinghorn@tower release]$ nvidia-smi -L
GPU 0: GeForce GTX 980 (UUID: GPU-477f0fd5-9db5-a015-e8b3-15ac96a06920)
GPU 1: Graphics Device (UUID: GPU-42f87dff-242a-17b2-7faf-2b4e18aec0d8)
GPU 2: GeForce GTX TITAN X (UUID: GPU-f195b1fa-16ec-ea58-8a17-146c0f93930e)

[kinghorn@tower release]$ ./nbody -benchmark -numbodies=256000 -device=0
...
gpuDeviceInit() CUDA Device [0]: "GeForce GTX TITAN X
> Compute 5.2 CUDA device: [GeForce GTX TITAN X]
number of bodies = 256000
256000 bodies, total time for 10 iterations: 3598. 562 ms
= 182.117 billion interactions per second
= 3642.344 single-precision GFLOP/s at 20 flops per interaction

[kinghorn@tower release]$ ./nbody -benchmark -numbodies=256000 -device=1
...
gpuDeviceInit() CUDA Device [1]: "Graphics Device
> Compute 5.2 CUDA device: [Graphics Device]
number of bodies = 256000
256000 bodies, total time for 10 iterations: 3906.858 ms
= 167.746 billion interactions per second
= 3354.921 single-precision GFLOP/s at 20 flops per interaction

[kinghorn@tower release]$ ./nbody -benchmark -numbodies=256000 -device=2
...
gpuDeviceInit() CUDA Device [2]: "GeForce GTX 980
> Compute 5.2 CUDA device: [GeForce GTX 980]
number of bodies = 256000
256000 bodies, total time for 10 iterations: 5129.597 ms
= 127.761 billion interactions per second
= 2555.210 single-precision GFLOP/s at 20 flops per interaction

[kinghorn@tower release]$ ./deviceQuery
. /deviceQuery Starting...

 CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 3 CUDA Capable device(s)

Device 0: "GeForce GTX TITAN X"
  CUDA Driver Version / Runtime Version          7.5 / 7.0
  CUDA Capability Major/Minor version number:    5.2
  Total amount of global memory:                 12288 MBytes (12884705280 bytes)
  (24) Multiprocessors, (128) CUDA Cores/MP:     3072 CUDA Cores
  GPU Max Clock rate:                            1076 MHz (1.08 GHz)
  Memory Clock rate:                             3505 Mhz
  Memory Bus Width:                              384-bit
  L2 Cache Size:                                 3145728 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(65536), 2D=(65536, 65536), 3D=(4096, 4096, 4096)
  Maximum Layered 1D Texture Size, (num) layers  1D=(16384), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(16384, 16384), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  2048
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 2 copy engine(s)
  Run time limit on kernels:                     No
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 129 / 0
  Compute Mode:
      Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) 

Device 1: "Graphics Device"
  CUDA Driver Version / Runtime Version          7. 5 / 7.0
  CUDA Capability Major/Minor version number:    5.2
  Total amount of global memory:                 6144 MBytes (6442254336 bytes)
  (22) Multiprocessors, (128) CUDA Cores/MP:     2816 CUDA Cores
  GPU Max Clock rate:                            1076 MHz (1.08 GHz)
  Memory Clock rate:                             3505 Mhz
  Memory Bus Width:                              384-bit
  L2 Cache Size:                                 3145728 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(65536), 2D=(65536, 65536), 3D=(4096, 4096, 4096)
  Maximum Layered 1D Texture Size, (num) layers  1D=(16384), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(16384, 16384), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  2048
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 2 copy engine(s)
  Run time limit on kernels:                     No
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 4 / 0
  Compute Mode:
      Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) 

Device 2: "GeForce GTX 980"
  CUDA Driver Version / Runtime Version          7. 5 / 7.0
  CUDA Capability Major/Minor version number:    5.2
  Total amount of global memory:                 4095 MBytes (4294246400 bytes)
  (16) Multiprocessors, (128) CUDA Cores/MP:     2048 CUDA Cores
  GPU Max Clock rate:                            1216 MHz (1.22 GHz)
  Memory Clock rate:                             3505 Mhz
  Memory Bus Width:                              256-bit
  L2 Cache Size:                                 2097152 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(65536), 2D=(65536, 65536), 3D=(4096, 4096, 4096)
  Maximum Layered 1D Texture Size, (num) layers  1D=(16384), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(16384, 16384), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  2048
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 2 copy engine(s)
  Run time limit on kernels:                     Yes
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 3 / 0
  Compute Mode:
      Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) 
> Peer access from GeForce GTX TITAN X (GPU0) -> Graphics Device (GPU1) : No
> Peer access from GeForce GTX TITAN X (GPU0) -> GeForce GTX 980 (GPU2) : No
> Peer access from Graphics Device (GPU1) -> Graphics Device (GPU1) : No
> Peer access from Graphics Device (GPU1) -> GeForce GTX 980 (GPU2) : No
> Peer access from Graphics Device (GPU1) -> GeForce GTX TITAN X (GPU0) : No
> Peer access from Graphics Device (GPU1) -> Graphics Device (GPU1) : No
> Peer access from GeForce GTX 980 (GPU2) -> GeForce GTX TITAN X (GPU0) : No
> Peer access from GeForce GTX 980 (GPU2) -> Graphics Device (GPU1) : No

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 7. 5, CUDA Runtime Version = 7.0, NumDevs = 3, Device0 = GeForce GTX TITAN X, Device1 = Graphics Device, Device2 = GeForce GTX 980
Result = PASS

Happy computing! —dbk
>/p>

Nvidia GeForce GTX 1080 vs 980

Nvidia’s GTX 1080 is the world’s most powerful consumer GPU. A couple of years ago, the same could be said of the GTX 980, so now it’s time to see what’s changed.

Comparing a new card with its predecessor might seem strange, but there’s some logic at play here: the GTX 1080 still costs nearly £600, while the GTX 980 can now be found for half of that amount. Is Nvidia’s older card turning into a high-end bargain?

Trusted has individual reviews of both cards, so click here to read our GTX 1080 and GTX 980 reviews.

Related: Best graphics cards to buy in 2016

Nvidia GeForce GTX 980 v Nvidia GeForce GTX 1080: Specs and technology

There are huge differences between the older GTX 980 and the factory-fresh GTX 1080, but that’s no surprise – they use different architectures and manufacturing processes.

For starters, the GTX 980 uses a 28nm manufacturing process and the Maxwell architecture, while the GTX 1080 and its new Pascal architecture have slimmed down to 16nm. That’s a big difference that allows Nvidia to push the envelope: it can now fit more transistors into the same amount of room, and it can now build them onto fins rather than in a single horizontal plane.

That move to 16nm and its greater transistor density means Nvidia can boost the output of its graphics cores, while fitting them into a smaller space, with increased electrical efficiency. The GTX 980 used 5.2 billion transistors on a 398mm2 piece of silicon, but the GTX 1080 has two billion more transistors inside a 314mm2 die. The GTX 980 required a peak power draw of 165W, while the new card needs just 15W more.

Related: Best gaming PC specs to build yourself

The increase in transistors has led to improvements elsewhere. The GTX 1080 has 2,560 CUDA cores while the GTX 980 made do with 2,048, and the new card’s 1,607MHz core clock outpaces the older GPU’s 1,126MHz speed.

Those CUDA cores benefit from Nvidia’s new Pascal architecture, too. It alters the organisation of the cores, bonding them into smaller groups to improve task management and delegation. That architecture also brings better asynchronous computing-style calculations to the table alongside Lens Matched Shading, which is far pickier about rendering pixels that the user won’t see.

There’s also simultaneous multi-projection, which can render a single scene from sixteen different viewpoints – a future-proofed inclusion with two eyes on VR headsets.

Then there’s the memory. The GTX 980 made do with 4GB of bog-standard GDDR5, while the GTX 1080 deploys 8GB of GDDR5X, which delivers huge speed and bandwidth improvements and is currently only found on Nvidia’s flagship card.

Video: GTX 1080 review

Nvidia GeForce GTX 980 v Nvidia GeForce GTX 1080: Testing Methodology

I benchmarked the GTX 980 and GTX 1080 on TrustedReviews’ test rig, which has been designed to match the specifications of a standard enthusiast gaming PC. It features the following components:

• Motherboard: Asus Z170-Deluxe
• Processor: Intel Core i5-6600K (not overclocked)
• RAM: Corsair Vengeance 2666MHz, 16GB DDR4
• Cooler: Corsair H60 liquid cooler
• PSU: Corsair CX750M
• SSD: Samsung 850 EVO
• OS: Windows 10 Pro 64-bit

I tested the cards with four top-end games. Each title was run at its highest settings at 1,080p, 1,440p and 2,160p (4K).

I also used a selection of synthetic tests, with 3D Mark’s demanding Fire Strike test alongside benchmarks to measure power consumption and peak temperature.

DIRT Rally
Ultra at 1,920 x 1,080 average

GTX 980: 111.2

GTX 1080: 154.8
Ultra at 2,560 x 1,440, average

GTX 980: 91.3

GTX 1080: 122.1
Ultra at 3,840 x 2,160, average

GTX 980: 39.7

GTX 1080: 63.2

The fast-paced texture changes and demanding lighting of this racing game can prove testing for GPUs, but this is still the easiest title in the slate of games. That makes it a good base-level benchmark, and I always test with every option maximised.

Both cards rattled through the game’s 1080p test with averages beyond 100fps, and the GTX 1080 continued that good form at 1440p – although here the GTX 980 could “only” manage 91.3fps.

This is an easier game, so it’s no surprise that both cards handled its benchmark at 3,840 x 2,160. However, the GTX 1080’s improved architecture and shrunken manufacturing process helped it open up a huge lead over its predecessor: it rattled through with a butter-smooth average of 63.2fps, while the GTX 980 only managed 39.7fps. That’s playable, but it’s a long way behind.

Middle-earth: Shadow of Mordor
Ultra at 1,920 x 1,080 average

GTX 980: 98

GTX 1080: 160.3
Ultra at 2,560 x 1,440, average

GTX 980: 75.3

GTX 1080: 112.3
Ultra at 3,840 x 2,160, average

GTX 980: 32.7

GTX 1080: 54. 9

This game is a more challenging test thanks to a wealth of particle and fire effects. It’s a short run, but it pushes GPUs as the camera sweeps through Middle-earth.

Middle Earth’s tougher test saw the GTX 980 begin to falter at my highest test resolution. Its 4K average of 32.7fps is just about playable, but its minimum framerate will be a long way below that 30fps barrier – which means that the game will stutter and slow down, especially in frantic moments.

It’s a better bill of health for the GTX 1080, which averaged 54.9fps in the same test – an easily playable pace.

The difference between old and new silicon was illustrated in other Shadow of Mordor tests. The GTX 980 was playable at 1440p and 1080p, but it was around 30fps behind the GTX 1080 at 2,560 x 1440 and around 60fps behind at 4K.

Hitman
Max Settings at 1,920 x 1,080, average

GTX 980: 67.7

GTX 1080: 81.5
Max Settings at 2,560 x 1,440, average

GTX 980: 66. 4

GTX 1080: 84.1
Max Settings at 3,840 x 2,160, average

GTX 980: 28.8

GTX 1080: 47.7

This benchmark is dense with NPCs, objects and complex lighting, and it’s one of the tougher tests in this suite of benchmark games.

This test really demonstrated the GTX 980’s strengths and weakness – it’s an excellent card for 1080p and 1440p gaming, but its performance falls down when it’s confronted with anything more demanding.

The GTX 980 maintained good speeds at lesser resolutions with average framerates of 67.7fps and 66.4fps – so that’s barely any change when jumping from 1080p to 1440p. However, when the game ran at 3,840 x 2,160 the GTX 980 averaged just 28.8fps – not good enough for smooth gameplay.

Nvidia’s newer card encountered no such issues. It handled 4K at 47.7fps, and averaged more than 80fps at lesser resolutions. This is a tough game, but Nvidia’s latest GPU won’t struggle.

Rise of the Tomb Raider
Max Settings at 1,920 x 1,080 average

GTX 980: 84. 7

GTX 1080: 135.8
Max Settings at 2,560 x 1,440, average

GTX 980: 56.1

GTX 1080: 92.5
Max Settings at 3,840 x 2,160, average

GTX 980: 29.9

GTX 1080: 48.3

Few games have weather effects quite like Tomb Raider, which makes it a great tool when it comes to benchmarking – only top cards will be able to handle this game smoothly, especially at high resolutions.

This is another game where the GTX 980 is good – but, when it comes to 4K, not good enough. Its 29.9fps average is only a tenth of a frame behind where I’d consider the title playable, but an inevitably lower minimum framerate ensures that the game won’t run smoothly.

No such problems for the GTX 1080, which eased through the 4K benchmark at 48.3fps and then almost doubled its pace to 92.5fps when I dropped the resolution to 2,560 x 1,440.

Both cards handled 1080p with ease, although the gulf between Maxwell and Pascal is clear: the GTX 980 blitzed through at 84. 7fps, but the new card averaged 135.8fps.

Synthetic Benchmarks, Power and Heat

3D Mark Fire Strike Ultra

GTX 980: 3286

GTX 1080: 4884
Peak power:

GTX 980: 270W

GTX 1080: 286W
Peak temperature:

GTX 980: 64°C

GTX 1080: 75°C

3D Mark Fire Strike clearly demonstrated the difference between the GTX 980 and GTX 1080 without game engines and different drivers playing havoc with results.

The GTX 980’s score of 3,286 points pales in comparison to the GTX 1080’s 4,884 points. That GTX 1080 result is also higher than the GTX 980 Ti, and the GTX 980 can’t beat the 3,953 scored by the GTX 1070 either.

Nvidia has developed this huge performance leap while consuming barely any extra electricity. With the GTX 980 installed the test rig demanded a maximum wattage of 270W, and this figure only rose to 286W with the GTX 1080 in place. Considering the performance gains, that’s a phenomenal result.

There’s a wider gap in temperature results, where the GTX 1080’s maximum level of 75°C is nine degrees higher than the GTX 980. That said, it’s always worth placing an asterisk next to temperature figures as so much of the score is down to heatsink and cooler designs.

Performance Conclusions

There’s no getting away from the fact that the GTX 1080 is the world’s best consumer graphics card right now – and it’s left the GTX 980 in the dust.

The difference between the two is stark. The older card might handle lesser resolutions with aplomb but the Pascal-powered GPU is still far faster. And it uses that extra oomph to deliver smooth framerates in 4K tests. That bodes well for VR and multi-monitor, too, which are both areas where the GTX 980 will begin to show its age.

Don’t write off the older card, though. It’s still a capable core for 1080p and 1440p gaming, with impressive framerates in every game I tested at these resolutions. And while it doesn’t have quite the power needed for current games at 4K, it’ll handle older titles just fine at 3,840 x 2,160.

If you can afford it, the GTX 1080 is the better card by some distance. But the GTX 980 is a viable option for those on a budget, not least those who don’t want to play at 4K.

Overclocking and SLI

Nvidia has developed GPU Boost 3.0 alongside its new cards, and it says the software delivers better control overclock speed – and that its new GPUs have ample overclocking headroom.

I’ve no reason to doubt that approach. After a few minutes of tinkering with the GTX 1080 I’d managed to add 200MHz and 189MHz to the core and memory clocks, which are both deeply impressive for a card that already pushes the envelope. Those tweaks delivered about five extra frames per second.

The SLI situation is a bit murkier. I haven’t been able to test two GTX 1080s together yet, and Nvidia won’t be letting customers use three- and four-card SLI unless they register with the firm and ask for an unlock code.

That said, I can’t imagine many people will need to run three- and four- card SLI when one or two of these GPUs will deliver such huge performance.

Buy Now: GTX 1080 at Amazon.com from $609.00

Other Things to Consider

First things first: don’t bother with Nvidia’s Founders Edition cards. They’re expensive models that don’t offer anything extra aside from the brownie points of having invested in one of the first GPUs to leave the factory.

The third-party card situation is far more compelling. Prices for board partner versions range between £590 and £750, with core and memory overclocks common – cores often now sit at 1,700MHz, with memory tweaked to 10,500MHz.

There aren’t any smaller versions of the GTX 1080 available, but the market is stuffed with good-looking GPUs with a multitude of fan and heatsink designs – and even water-cooled versions.

The GTX 980 is in a similarly good situation. Prices begin at £275, and there’s just as much variety when it comes to board partner designs. Overclocks are commonplace, too.

Buy Now: GTX 980 at Amazon.com from $369.99

Verdict

The GTX 980 may be far cheaper than the GTX 1080, but it’s worse in numerous key areas. It’s slower at every resolution, and doesn’t really have the power to handle 4K – and it’ll only get worse as games become more demanding.

The older card also doesn’t have the new features of its successor, nor does it have the efficiency improvements. If you can’t spend as much on a new GPU, I’d drop further back to the GTX 970: it’s even cheaper, and it also handles 1080p and 1440p gaming smoothly.

Related: Best graphics cards to buy in 2016
 

Nvidia GeForce GTX 980 Ti Review

Our Verdict

Nearly as fast as the Titan X at a better price point, the GTX 980 Ti is the new choice for high-end gamers. The ball is now in AMD’s court.

The «value» alternative for high-end PCs

We all know the type: the ultra-cool guy that has the best hair, wears the right clothes, and drives the hottest car. The problem is that trying to make friends with such a person is difficult… and it will cost you. But did you know that the coolest guy in town has a little brother who’s much more down to earth? Oh, he still dresses right and drives a nice car, but he’s not quite so ostentatious. His name is GeForce GTX 980 Ti, son of Nvidia and brother to the GeForce GTX Titan X.

There’s not a whole lot of surprise with the announcement of the GeForce GTX 980 Ti (henceforth the 980 Ti). We saw the writing on the wall as soon as the GTX Titan X was launched: “Great, here’s the halo $1,000 luxury GPU that most can only dream of owning; now show us something we might actually buy!” The same thing happened with the original Titan and the GTX 780 Ti, though this time there shouldn’t be any instances where the less expensive part actually offers superior performance. There’s also no improved FP64 mode on Titan X, so saving 30 percent and getting roughly 95 percent of the performance is hardly a bad thing, but let’s not jump ahead.

Meanwhile, we’re still waiting for AMD’s next-generation GPU to make the scene. That means that for the time being, the 980 Ti is going to be even more expensive than the already costly GTX 980. AMD’s Fiji GPUs can’t possibly come soon enough. Last time, when NVIDIA launched the GTX 780 Ti, it was intended to reclaim the performance crown from AMD’s just-released R9 290X, and the price of the GTX 780 dropped $150 in response. We’ll have to see what AMD has in store and whether they can help bring some balance—and much-needed competition—to the high-end graphics market in the coming weeks.

To give you a high-level overview of the current competitive landscape, here’s a quick rundown of specifications for most of the high-end GPUs:

Graphics Card Specifications

Titan X

GTX 980 Ti

GTX 980

GTX 970

GTX 780 Ti

GTX 780

R9 290X

Generation

GM200

GM200

GM204

GM204

GK110

GK110

Hawaii

Core Clock (MHz)

1,000

1,000

1,216

1,088

876

889

«up to» 1GHz

Boost Clock (MHz)

1,075

1,075

1,317

1,228

928

941

N/A

VRAM Clock (MHz)

7,010

7,010

7,000

7,000

7,000

6,000

5,000

VRAM Amount

12GB

6GB

4GB

4GB

3GB

3GB

4GB

Bus

384-bit

384-bit

256-bit

256-bit

384-bit

384-bit

512-bit

ROPs

96

96

64

56

48

48

64

TMUs

192

176

128

104

240

192

176

Shaders

3,072

2,816

2048

1,664

2,880

2,304

2816

SMMs

24

22

16

13

15

12

N/A

TDP (watts)

250

250

165

145

250

250

290

Launch Date

March 2015

June 2015

Sept 2014

Sept 2014

Nov 2013

May 2013

Oct 2013

Launch Price

$999

$649

$549

$329

$649

$699

$549

Current Price

$1,049

$649

$499

$329

N/A

N/A

$299

Comparing the two GM200 cards, other than the drop in the number of shaders/SMMs, the 980 Ti looks a lot like the Titan X. It still has 96 ROPs and a 384-bit GDDR5 interface, though “only” with 6GB VRAM this time. We’re already seeing signs of 4GB VRAM being insufficient, particularly for higher resolution modes like 4K, but with very few cards sporting more than 4GB VRAM most games are unlikely to need 12GB VRAM. 6GB on the other hand, yes, thank you, we can definitely use that! It’s also important to note that there are none of the 3.5GB + 512MB segmented memory shenanigans of GTX 970 at play with the 980 Ti; this is a fully enabled GM200 chip as far as the memory interface and cache are concerned. That means in situations where games don’t need more than 6GB VRAM, the 980 Ti should be no worse than eight percent slower than the Titan X—and in many cases it will be less than that.

We’ve already covered most of the other interesting tidbits for GM200 in the Titan X review. It has eight billion transistors compared to six billion in the GM204 (and 5.2 billion in GK110). The chips are divided into SMMs—“streaming multi-processors”—each of which contains 128 CUDA cores subdivided into four warp schedulers of 32 cores each. Each SMM also has 16 Texture Units (TMUs), a 96KB memory pool, and another 48KB of shared L1/texture cache. The main difference here is that 980 Ti disables two of the SMMs while keeping everything else in place.

While the basic design of Nvidia’s first- and second-generation Maxwell parts is the same, it’s important to note that only second-generation Maxwell processors support several new features: DSR (Dynamic Super Resolution), third-generation Delta Color Compression, Multi-Pixel Program Sampling, VXGI (Voxel Global Illumination), VR Direct, Multi-Projections Acceleration, and MFAA (Multi-Frame Sampled Anti-Aliasing). HDMI 2.0 is also supported on second-generation Maxwell, which means you can do things like drive a 4Kp60 display without resorting to any compression algorithms.

Most of these new features are targeted at high-end gaming, so while it’s unfortunate that not all Maxwell GPUs have the same feature set, in practice it’s not quite so bad—running a VR device off a GeForce 750 Ti wasn’t going to be a great experience with or without VR Direct, and 4K gaming is generally out of reach of even powerful GPUs like the GTX 970, never mind the 750 Ti. We’ll look at some of the other new technology Nvidia is announcing today in separate articles, but here our focus is purely on the 980 Ti.

There are some other interesting tidbits at play here. For example, even though the 980 Ti has fewer cores than the 780 Ti, the Maxwell 2.0 cores are more efficient. GTX 980 is already able to match or slightly exceed the performance of the 780 Ti with 2048 cores, so accounting for clock speeds the 980 Ti should be at around 15–20 percent faster than the GTX 980. In situations where the VRAM amounts come into play (e.g., at 4K resolutions), the performance delta will potentially be even greater. Nvidia claims that the 980 Ti is as much as 65 percent faster than the 780 Ti at 4K resolutions, and compared to the GTX 680 they’re claiming up to a 3x increase in performance. Those are lofty claims, and realistically, that “up to” phrase is pretty nebulous; on average, the increase will often be less substantial, except in cases where you run out of VRAM.

The GTX 980 Ti reference design retains Nvidia’s proven cooling solution from the Titan X and earlier GPUs. It has worked well and manages to provide good cooling without generating a ton of noise, so it’s no surprise to see it appear yet again. Unlike Titan X, however, add-in board manufacturers are free to experiment a bit more with 980 Ti, so while our test GPU looks exactly like Titan X, other than the silver shroud instead of black and the 980 Ti logo emblazoned on its side, we’re likely to see open-air coolers as well as blowers from companies like EVGA, Asus, etc.

Test Hardware and Software: We’ve Got Upgrades

Before we get to the benchmarks, let’s also talk about our new GPU test bed. An introduction might also be in order, in case you missed it. The long and short of it is that we’ve got a new senior editor, and along with that we’ve upgraded our GPU test bed. You see, our old Sandy Bridge-E system was feeling a bit out of date. We looked at the options and decided that with a new editor it was a good time to upgrade systems. Given the choice between Haswell, Haswell-E, and Broadwell, Haswell-E continues to deliver the best overall combination of performance and features, and the availability of 40 PCI Express lanes can be important. The extra lanes are most useful when you’re running multiple GPUs, which we have done with several of our graphics cards. So, we grabbed our recent Turbo build and tossed in a CPU overclock to further reduce any potential CPU bottlenecks, and then got busy testing GPUs on the new platform. Here’s our test system:

Maximum PC 2015 GPU Test Bed
CPU	Intel Core i7-5930K (4.2GHz Overclock)
Mobo	Gigabyte GA-X99-UD4
GPUs	Nvidia GeForce GTX Titan X Nvidia GeForce GTX 980 Ti 2x EVGA GeForce GTX 980 SC 2x Zotac GeForce GTX 970 Nvidia GeForce GTX 770 2x AMD Radeon R9 290X
SSD	2x Samsung 850 Evo 250GB
HDD	Seagate Barracuda 3TB 7200RPM
PSU	EVGA SuperNOVA 1300 G2
Memory	G. Skill Ripjaws 16GB DDR4-2666
Cooler	Cooler Master Nepton 280L
Case	Cooler Master CM Storm Trooper
Drivers	Nvidia 352.90 Beta AMD Catalyst 15.5 Beta

We’ve also updated our benchmarks a bit. In the interest of keeping things mostly consistent, we’ve kept all of the existing benchmarks from the Titan X review, but we’ve added two new games: Grand Theft Auto V and The Witcher 3. Both titles are quite demanding, and GTAV is interesting in that it will actually show how much VRAM the game needs to store all assets on the GPU. Hint: with all settings maxed (which is how we roll at Maximum PC), only 1920×1080 is able to fit within 4GB VRAM; 2560×1440 requires around 4.2GB VRAM, while 3840×2160 comes in just under 6GB. This should provide some interesting results with the different amounts of VRAM on the cards we’re testing.

We’ve also started running 3DMark’s Fire Strike Ultra and Extreme tests, though the gaming results will continue to be our primary emphasis. Finally, we’ll be reporting both average and minimum (or in some cases, 97th percentile) values for the games we benchmark. Average frame rates are great for giving you a general idea of how well a game runs, but performance consistency is also important—a game that runs at 60+fps most of the time but drops below 20fps on occasion would provide a somewhat choppy gaming experience. The minimum frame rates will be particularly telling when it comes to VRAM limitations, where texture thrashing can occur on cards that don’t have enough memory.

Enough talk. We know Titan X is a fast card, and there’s every reason to expect the 980 Ti will be a very close second. Let’s hit the benchmarks.

OMG… graphs! Nine games and applications, three resolutions, and nine GPUs. Yes, that is a lot of benchmarks to run, but we like to be thorough. There’s a lot to digest here, with individual game results as well as the average performance across all tested games. We’ve separated the charts into two galleries: single-GPU comparisons and comparisons against our three SLI/CrossFire configurations. We’ll start with the single-GPU charts.

Just as expected, the GTX 980 Ti is a force to be reckoned with. If you just look at the top graph, you’ll find everything you really need to know about overall performance. The GeForce 980 Ti ends up being pretty darn close to equal with the more expensive Titan X, particularly at lower resolutions. How close does it come? The largest difference in average frame rates is less than five percent, and across all eight games tested it’s less than a one percent gap. Minimum frame rates are slightly more in favor of Titan X, with an average difference of seven percent at 4K, but only 2-3 percent at 1080p and 1440p. If you want Titan X performance for $350 less, the 980 Ti delivers.

For those with an older GPU like the GTX 770 (which is about 10 percent faster than a GTX 680), the GTX 980 Ti represents an astounding doubling of performance at our 1080p and 1440p settings; if you happen to run 4K games at maxed-out settings, we actually averaged more than a 3x improvement in performance. That’s mostly thanks to a couple of games that did very well—Tomb Raider shows a 4.7x improvement and The Witcher 3 shows a 5.8x improvement. Then again, both Batman: Arkham Origins and Grand Theft Auto V failed to run on the GTX 770 with our 4K settings, so in that sense the improvement is even better. Practically speaking, it means you can go from an unplayable single-digit gaming experience at 4K maximum quality on the 770 to a reasonably playable 30+fps experience. Of course, you can get far better 4K performance if you disable AA and turn a few settings down a notch.

If you already have a good GPU, 980 Ti isn’t going to provide nearly as much benefit. Compared to the 970, it’s a reasonable bump in performance of 55-65 percent, but compared to the GTX 980 it’s a much smaller 20–25 percent improvement. The R9 290X continues to show its potential, especially at higher resolutions where on average it’s only 10 percent slower than the GTX 980. That makes the 980 Ti 35–40 percent faster than the 290X, but you could buy two AMD R9 290X GPUs instead of the 980 Ti and still have money left for lunch. That’s the difficulty Nvidia needs to overcome when it comes to convincing gamers to buy their latest “affordable” powerhouse. But since you can buy two R9 290X cards or two GTX 970 SLI cards for the price of a single GTX 980 Ti, it’s worth looking at how that actually plays out.

GeForce GTX 980 Ti Takes on Multi-GPU

[Ed—We have tested with a second 980 Ti in SLI, for those that are interested.]

Raise your hand if you were expecting the GTX 970 SLI setup to crush the GTX 980 Ti. Sure, there are definitely cases where the SLI 970 cards win, but there are also times when that single high-end GPU is the better approach. Batman and Witcher don’t appear to scale as well with multiple GPUs, but even in the best scenarios, the 970 SLI configuration is only about 20 percent faster than a 980 Ti. Overall, the average performance advantage of 970 SLI is a paltry 4–8 percent, depending on resolution, with higher resolutions scaling slightly better with SLI.

290X CF is at least a clear winner in every benchmark, with no instances where a single GPU beats the two Hawaii cores. AMD has some particularly strong showings at 4K, leading to an overall 30 percent performance improvement at that resolution over the newcomer 980 Ti. Note that at 1080p the benefits of CrossFire are often outweighed by greater CPU usage to hit high frame rates, so if you’re using a 1080p display, CrossFire and SLI probably aren’t necessary.

However, looking at average frame rates only tells half the story; the minimum frame rate can be just as important. Minimum frame rates are also why we generally prefer a single fast GPU over two slower GPUs running in SLI/CF. There’s CPU and system overhead associated with SLI/CF, so while it can provide clear benefits to average frame rates, minimum frame rates will often drop.

Case in point: We just mentioned that on average 970 SLI is roughly six percent faster than the 980 Ti and 290X CF is up to 30 percent faster. Switch over the minimum frame rates and the story changes: On average 970 SLI is almost 10 percent slower than a single 980 Ti—16 percent slower at 4K. 290X CF likewise shows an overall decrease in minimum frames per second of 10 percent, and over 20 percent at 1080p, where CPU overhead is creating a bigger bottleneck.

It might seem a bit surprising how little impact the extra 2GB VRAM seems to have on performance, but minimum frame rates as well as the benchmarks themselves are the explanation. If a game doesn’t use more than 4GB VRAM at our benchmark settings, having more VRAM does nothing for you—it’s why the 12GB VRAM in the Titan X feels like overkill, at least for now. Since many of our games are a bit older, it’s no surprise that they don’t need more than 4GB—or even 3GB. 6GB is really forward looking, but it will become increasingly important over the coming year.

GTAV is the poster child for being able to use 6GB VRAM, but it requires some pretty strenuous settings to get there. Max out all the Graphics and Advanced Graphics settings and run at 4K, and the VRAM use is very close to 6GB. That means all of the 4GB cards should show signs of thrashing, and they do. The R9 290X in particular has trouble, and while the 18fps of the 980 Ti is nothing extraordinary, the 4–5 minimum frames per second results for 290X (and 290X CF) show how bad things can get.

Overclocking and Cooling

We haven’t had time to fully investigate overclocking, but in limited testing it looks like a 15 percent overclock of the GPU core should be attainable, and at least a 5 percent GDDR5 overclock. That’s enough to more than close the gap with Titan X, though Titan X has similar overclocking potential. [Update: We’ve finished our overclocked GTX 980 Ti testing.]

We’ll be running some additional tests over the coming week, including power and noise testing, but there’s nothing particularly noteworthy to discuss in those areas. The 980 Ti has a 250W TDP, and it will make use of it. The cooler meanwhile continues to be the gold standard for a blower-style graphics card, which is really the only way to go (outside of liquid cooling) if you want to run 3-way or 4-way SLI.

Conclusion

The two-word summary of the GTX 980 Ti is this: It’s awesome! We didn’t provide benchmarks of every game out there, but other titles exist where having 4GB VRAM can cause problems—Assassin’s Creed: Unity comes to mind. With the Xbox One and PlayStation 4 both sporting 8GB total RAM, we’re seeing many titles push beyond the 2GB mark that has been around for so long. Couple that with 4K displays and we’ve rapidly gone from 2GB being “enough” and 4GB being “future-proof” to a new level where at least 6GB is necessary if you want to enable maximum quality. Like the Titan X, the 980 Ti ensures you have enough VRAM, and then it tosses in a bunch of bandwidth and shaders as a bonus.

But as much as we like everything about the GTX 980 Ti, we also have one other word of advice: Wait!

Yes, the GTX 980 Ti performs better than any other GPU besides the Titan X, but AMD’s Fiji with HBM (High Bandwidth Memory) should be here within the month. We don’t have details on the memory configuration for AMD’s Fiji parts, and it’s possible they’ll stay at 4GB with HBM, but as the saying goes: The proof is in the eating of the pudding. Assuming you didn’t jump on Titan X when it launched, waiting another few weeks to see where AMD’s cards fall won’t hurt. It may end up faster than 980 Ti at a better price, it may be faster with a higher price, or it could end up slower but with a better price. Whatever happens, we can expect Nvidia to respond appropriately. Unless you absolutely have to upgrade right now, in which case the 980 Ti looks primed for the future.

If the new GTX 980 Ti doesn’t tickle your fancy, check out recent choices in our guide to the best graphics cards on the market.

Follow Jarred on Twitter.

Read our review policy

Reviews

Nearly as fast as the Titan X at a better price point, the GTX 980 Ti is the new choice for high-end gamers. The ball is now in AMD’s court.

Jarred’s love of computers dates back to the dark ages when his dad brought home a DOS 2.3 PC and he left his C-64 behind. He eventually built his first custom PC in 1990 with a 286 12MHz, only to discover it was already woefully outdated when Wing Commander was released a few months later. He holds a BS in Computer Science from Brigham Young University and has been working as a tech journalist since 2004, writing for AnandTech, Maximum PC, and PC Gamer. From the first S3 Virge ‘3D decelerators’ to today’s GPUs, Jarred keeps up with all the latest graphics trends and is the one to ask about game performance. 

Nvidia GTX 980 Ti review: All the power of the Titan X for $650

Tech —

No, it’s not cheap, but the 980 Ti is still an outstanding graphics card for the price.

Mark Walton
— Jun 1, 2015 1:08 pm UTC

After an endless stream of rumours, renders, and cheeky tweets, AMD is finally on the cusp of releasing a brand new flagship graphics card. We know that it’ll feature a huge 1024-bit memory bus featuring AMD’s innovative high bandwidth memory, and that it’ll only have 4GB of it. We also know roughly what it’ll look like, if the renders and tweets are to be believed. What we don’t have is a release date, even if an official announcement at Computex or E3 is likely. The Radeon R9 390X, or whatever it ends up being called, is an important product for a company that desperately needs a win right now to claw back the market share it’s lost to rival Nvidia—AMD’s last flagship single-GPU launched back in October of 2013.

It’s an exciting time for AMD, then. Typically enough, though, Nvdia’s here to rain on AMD’s parade. There’s no shiny new tech or architecture change this time, but the brand new GTX 980 Ti is one of Nvidia’s most aggressive graphics card launches to date—even more so than the exceptionally good value GTX 970, which got you most of the GTX 980’s performance at nearly half the price.

The GTX 980 Ti is priced at £550 ($650/€740), features 6GB of GDDR5 memory matched to a 384-bit bus, and is based on the same full-fat GM200 Maxwell chip as the £870 ($1000) Titan X. The card will be available worldwide at all the usual retailers this week.

Advertisement

While a few CUDA cores and texture units are missing (Nvidia needed to do something with all those binned chips that didn’t make cut, right?), the 980 Ti is, for all intents and purposes, a Titan X with less memory and a far more palatable price tag. Performance-wise, the 980 Ti is almost identical to the benchmark-topping Titan X. Yesterday, getting the best single-GPU performance on the market meant spending £870; today, it can be had for £550. Who says a bit of healthy competition is a bad thing, eh?

Yes, it’s pretty much a Titan X

When the original Titan launched, it was a monster graphics card, but it didn’t actually make use of the full (GK110) Kepler architecture it was based on. It wasn’t until the release of the 780 Ti that we saw fully utilized GK110 chips. The Titan X and 980 Ti reverse that situation, with the 980 Ti featuring 2816 CUDA cores and 176 texture units versus the 3072 CUDA cores and 192 texture units of the Titan X.

That’s not much of a cut considering how much cheaper the 980 Ti is, and those few missing CUDA cores make little difference to real-world performance.

Specs at a glance:	GTX Titan X	GTX 980 Ti	GTX 980	GTX 970	GTX 780 Ti
CUDA Cores	3072	2816	2048	1664	2880
Texture Units	192	176	128	104	240
ROPs	96	96	64	56	48
Core Clock	1000MHz	1000MHz	1126MHz	1050MHz	875MHz
Boost Clock	1050MHz	1050MHz	1216MHz	1178MHz	928MHz
Memory Bus Width	384-bit	384-bit	256-bit	256-bit	384-bit
Memory Speed	7GHz	7GHz	7GHz	7GHz	7GHz
Memory Bandwidth	336GB/s	336GB/s	224GB/s	196GB/s	336GB/sec
Memory Size	12GB GDDR5	6GB GDDR5	4GB GDDR5	4GB GDDR5	3GB GDDR5
TDP	250W	250W	165W	145W	250W

Stepping down to 6GB of memory does little to harm performance either. The Titan X’s 12GB was always overkill, being more suited to the world of workstations and VFX production (where entire scenes must be stored in graphics memory), rather than the humble gaming desktop.

6GB of GDDR5 running at 7GHz attached to a 384-bit bus for 336GB/sec of memory bandwidth is more than enough for running games. That’s certainly less than what you’ll get with AMD’s HBM, but Nvidia’s memory-compression technology has worked wonders so far, resulting in a pixel fill rate that’s miles ahead of the competition. As for whether 6GB is «enough,» even the likes of GTA V and Shadow of Mordor—two games that are notoriously memory hungry—top out at 6GB, even with ultra HD texture packs and mods in tow.

Advertisement

Still, those wanting to squeeze the maximum amount of life out of their GPU—or are into the aforementioned rendering work—may want to stick with the Titan X for the time being. Even then, though, with the likes of the upcoming DirectX 12 and Vulkan APIs set to support memory pooling across multiple graphics cards, perhaps the wiser investment is in two 980 Ti cards for a full-on 4K gaming rig at just a small premium over the cost of a single Titan X.

Enlarge / Unlike with the Titan X, partners will be able to create custom cooling solutions like Zotac did with its Amp Extreme Core Edition GTX 970.

The rest of the 980 Ti specs are standard Nvidia fare. For the reference design at least, it features the same excellent all-metal cooler that launched with the original GTX Titan, complete with great cooling performance and a snazzy light-up GeForce logo for all those windowed case users out there. While blower-style coolers (those that force hot air outside the case) often get a bad rap for being noisy, Nvidia’s reference cooler fares pretty well. While it was audible under load, it wasn’t horribly distracting like many other blower cards we’ve used in the past. The blower design is also great for smaller cases, where the reduced space doesn’t work so well venting hot air internally. You’ll still have to make room for the 980 Ti’s dual-slot design and 267mm length.

Power comes courtesy of one 8-pin and one 6-pin PCIe connector, with Nvidia recommending a minimum 600W PSU to handle the 980 Ti’s 250W TDP. Nvidia says that still leaves plenty of headroom for overclocking, with the company claiming that some cards can be boosted as high as 1400MHz.

Port-wise, you get 3x DisplayPort, 1x HDMI 2.0, and 1x dual-link DVI. If the reference cooler or port loadout doesn’t appeal, thankfully—unlike with the Titan X—Nvidia is allowing partners to create their own custom designs. Expect the likes of the excellent Asus Strix and MSI Twin Frozr coolers to make an appearance on the 980 Ti sooner rather than later.

Nvidia GeForce GTX 980 Ti review

We knew it was coming, of course — it was just the speed of its arrival that took us off-guard. Nvidia’s astonishing Titan X graphics card, based on the 8bn transistor GM200 processor, redefined the boundaries of single-GPU performance when we reviewed it back in March. Imagine the firm’s previous flagship — the GTX 980 — combined with a mainstream GTX 960 in a single package, backed by a ridiculous 12GB of 7gbps GDDR5 and you have a remarkable technological achievement. The only problem was its price — $999. The new GTX 980 Ti is a mildly cut-back version of the same product, offering around 98 per cent of the raw performance at 65 per cent of the price.

That’s not a bad outcome bearing in mind the extent of the compromises Nvidia has made. The 3072 cores of the Titan X get cut-down by a factor of nine per cent, two shader clusters disabled to give us a final tally of 2816 stream processors. By extension, we lose some texture mapping units too — dropping from 196 in Titan X to 176 in the GTX 980 Ti. However, the biggest cutback is perhaps the most inconsequential of all, bearing in mind the current gaming landscape. The mammoth 12GB of GDDR5 found in the top-end flagship is pared back to a still lavish, but more reasonable 6GB. However, memory speeds remain the same, core and boost clocks are identical, and unlike the GTX 970, Nvidia’s cutbacks have not come at the expense of ROPs or bandwidth — both are identical to Titan X. To be clear, this time there are no split-memory shenanigans.

Graphics cards are parallel by nature, distributing work across however many cores are available. That being the case, it’s a totally reasonable assumption to suggest that losing around nine per cent of the computational power should result in a similar drop to overall performance. However, raw processing power is just one element of the equation and in a card which is seemingly already hitting bottlenecks elsewhere within the system (be it the driver, the DX11 API or the CPU), the reality is that the GTX 980 Ti is virtually interchangeable with its much more expensive sibling.

Indeed, at a couple of points during our testing, the 980 Ti actually seemed a tiny bit faster than Titan X — theoretically impossible of course, but a practical reality when two cards with so much computational power butt up against limitations elsewhere within the system. That’s no mean feat considering our test system uses a Core i7 4790K overclocked to 4.6GHz along with 16GB of 1600MHz DDR3 RAM. On top of that, in common with the entire Maxwell line of Nvidia GPUs, the GTX 980 Ti is a very decent overclocker — not only can you increase speeds to match Titan X, you can move considerably beyond its stock performance too.

Caption

Attribution

• Order the GeForce GTX 980 Ti 6GB from Amazon with free shipping

Similarities with the top-end card don’t end at the specs either. Aside from a colour scheme more in keeping with the rest of the GeForce line, the GTX 980 Ti boasts the same high-quality chassis and excellent cooler found on Titan X. TDP is the same at 250W, meaning that the six and eight-pin power input arrangement remains in place, while identical video outputs remain in place — one dual-link DVI, three DisplayPorts, plus an HDMI 2.0 port, allowing for a 4K pixel throughput at 60Hz.

All of which is fascinating of course, but what you’re really here for is the performance — validation that you can enjoy Titan X levels of rendering power without paying quite so much for the privilege. We kick off with our standard gameplay test, upgraded for this new top-tier of graphics hardware. We’re running Crysis 3 at its highest visual quality, with top-tier textures engaged and SMAA T2x anti-aliasing in play. Up until the arrival of Titan X, we’ve carried out this test at 1080p resolution, but it’s clear that this is not going to stress this new wave of Nvidia hardware at all, so we move on to 2560×1440.

A standard benchmark involves turning v-sync off and running the game at the absolute fastest refresh it can manage (and if that’s your bag, we’ve got more than enough testing of that ilk further on down the page). However, in our opinion, the optimal gameplay experience is to match the native resolution of your display and its refresh rate. In short, we want something as close to locked, consistent 60fps as possible, while maintaining 1440p resolution on this highly demanding game.

The results? Well, there’s just a 0.79 per cent variance from the target frame-rate between the new card and the more capable Titan X — margin of error stuff, really. The GTX 980 Ti fails to hit the target for just 3.35 per cent of the duration, up against the somewhat more noticeable 20.66 per cent of the older GTX 980 — but Titan X remains top of the pile with a mere 2.56 per cent of dropped frames. Bearing in mind that the playthroughs seen here are not totally like-for-like, it’s clear that we’re going to need a more forensic look at game engines in motion to separate the two GM200 cards.

In our Crysis 3 gameplay tests, the idea isn’t to run at the absolute fastest frame-rates possible, but rather to synchronise as closely as possible with the display in terms of resolution and refresh — in this case, operating at 2560×1440 at 60Hz. We’re using the top-end very high preset, with very high textures, SMAA T2x anti-aliasing and v-sync.

Crysis 3 1440p60/V-Sync Gameplay	GTX 980 Ti	GTX Titan X	GTX 980
Lowest Frame-Rate	44.0fps	44.0fps	31.0fps
Dropped Frames (from 18650 total)	624 (3. 35%)	477 (2.56%)	3852 (20.66%)

Order the graphics cards tested against the GTX 980 Ti from Amazon with free shipping:

• Radeon R9 290 4GB
• Radeon R9 290X 4GB
• GeForce GTX 970 4GB
• GeForce GTX 980 4GB
• Titan X 12GB

GTX 980 Ti wasn’t really designed to run games at 1080p resolution, and as we noted in our Titan X review, this is where the like-for-like gains over the GTX 980 are at their lowest. Across all nine games we tested, there’s an average boost of just 24.8 per cent in performance compared to the reference GTX 980 — a figure likely to fall significantly if you bought (or are in the market for) a third-party board such as the excellent GTX 980 Strix from Asus or the equally impressive MSI Gaming 4G model.

Our biggest gains come from well-optimised games with a fair amount of CPU overhead left over: Ryse and Shadow of Mordor are 29 per cent faster, and Tomb Raider and Crysis 3 aren’t too far behind. Worst performer of the bunch is Far Cry 4 with a mere 15.8 per cent boost — which may have something to do with its over-reliance on single thread CPU performance.

However, for many, it’s the comparison with Titan X that is of most relevance. At stock settings, an average of all titles sees the GTX 980 Ti hand in 98.6 per cent of the Titan X’s performance and not one title fell below the 97 per cent threshold. We’ll go into overclocking in a touch more depth later, but as you can see, the GTX 980 Ti easily boosts past the thousand dollar flagship’s stock performance (of course, you can overclock that too). That said, you don’t get the best value from your OC at 1080p — there’s an average increase of 11.2 per cent, the biggest improvements coming from Battlefield 4 and Shadow of Mordor at 14 per cent. Bearing in mind we implemented a 220MHz core clock boost, along with an additional 400MHz on the RAM, that’s not exactly a great return.

1080p is the most popular gaming resolution for those using discrete graphics cards, but GTX 980 Ti and Titan X are under-utilised here, often bottlenecked by CPU and driver overheads. That’s bad news for those looking to use these cards for driving higher refresh full HD displays.

1920×1080 (1080p)	R9 290	R9 290X	GTX 970	GTX 980	GTX 980 Ti	GTX 980 Ti OC	Titan X
The Witcher 3, Ultra, HairWorks Off, Custom AA	46.4	50.5	56.0	65.8	82.6	92.1	84.4
Battlefield 4, Ultra, 4x MSAA	64.3	68.0	74.6	86.5	109.9	125.4	112.4
Crysis 3, Very High, SMAA	65.4	70.2	71.4	81.5	104.0	115.5	105.2
Assassin’s Creed Unity, Ultra High, FXAA	38.0	42.6	51.7	62.4	74.4	84.3	74.7
Far Cry 4, Ultra, SMAA	72.2	75.0	77.4	87.4	101. 2	103.0	101.4
COD Advanced Warfare, Extra, FSMAA	90.5	92.9	117.8	128.0	156.8	173.1	159.9
Ryse: Son of Rome, High, SMAA	69.2	74.8	65.1	75.8	97.8	109.5	99.2
Shadow of Mordor, Ultra, High Textures, FXAA	82.8	88.3	80.8	91.7	118.5	135.4	119.0
Tomb Raider, Ultimate, FXAA	84.1	89.0	102.4	118.2	150.1	168.2	150.3

At 2560×1440, the gap with GTX 980 opens up still further as the advantages of the 384-bit memory bus come further into play. Increases to performance compared to GTX 980 still vary though, with some fairly disappointing results from Assassin’s Creed Unity, Advanced Warfare and Far Cry 4 — all in the 22-23 per cent area. However, BF4, Crysis 3 and Ryse all comfortably break the 30 per cent threshold, with The Witcher 3 and Tomb Raider getting extremely close.

1440p resolution also sees our overclock hits its optimal level of performance, averaging out across all nine games with an impressive 15.6 increase to frame-rates (ACU faring worst at 12.6 per cent, with Tomb Raider achieving an impressive 18.4 per cent boost). Bearing in mind that the GTX 980’s effectiveness declines as you raise resolution, the 980 Ti’s showing here isn’t unexpected, but it is welcome nonetheless.

The fact that the new card features the exact same level of memory bandwidth as Titan X may also explain the stand-out result here, as Shadow of Mordor sees its performance matched exactly between the two GM200 cards. Across all games, the GTX 980 Ti handed in 98 per cent of the Titan X’s performance — very, very slightly down compared to the 1080p results, but once again not a single game managed less than 97 per cent of the more expensive card’s performance.

In terms of pixel count, quality settings and overclocking potential, there’s a strong argument that 2560×1440 is the sweet spot for the GTX 980 Ti’s capabilities.

2560×1440 (1440p)	R9 290	R9 290X	GTX 970	GTX 980	GTX 980 Ti	GTX 980 Ti OC	Titan X
The Witcher 3, Ultra, HairWorks Off, Custom AA	35.3	38.2	39.2	47.5	61.7	70.9	63.3
Battlefield 4, Ultra, 4x MSAA	43.4	46.8	48.9	57.0	75.0	86.7	76.1
Crysis 3, Very High, SMAA	42.0	45.1	43.0	50.0	66.2	75.7	68.0
Assassin’s Creed Unity, Ultra High, FXAA	26.6	29.5	32.8	39.4	48.3	54.4	49.6
Far Cry 4, Ultra, SMAA	53.2	57.8	53.9	61.3	75.4	86.9	77.0
COD Advanced Warfare, Extra, FSMAA	77.2	81. 5	87.0	98.2	121.3	139.4	123.2
Ryse: Son of Rome, High, SMAA	51.8	55.6	46.3	54.1	71.2	83.3	72.8
Shadow of Mordor, Ultra, High Textures, FXAA	61.3	65.7	57.0	66.0	87.2	101.8	87.2
Tomb Raider, Ultimate, FXAA	56.9	62.4	65.5	76.7	99.2	117.2	101.9

Moving on to 4K resolution, here’s where the GM200 cards comfortably pull ahead of the GTX 980 — the 384-bit memory really coming its own, and in many ways, taking centre-stage. Disappointments in terms of performance increases are fewer, Assassin’s Creed Unity’s mere 22 per cent the outlier in an area where boosts in excess of 30 per cent are much more common. The average performance increase up across all games up against GTX 980 just about hits 30 per cent, but more important than the raw numbers is the gameplay experience: with some settings tweakage, all of our test titles bar ACU are eminently playable at this extreme resolution, some of them even approaching 60fps.

The importance of memory bandwidth is reflected in the fact that GTX 980 Ti seems to be even closer to Titan X performance-wise than at 1440p, to the point where occasionally we see the less capable card very slightly outperforming the full GM200 in our tests — margin of error stuff on titles where CPU bottlenecks can cause an issue, and indicative of just how close these two GPUs are in actual real-life conditions. Also noteworthy is that our overclock offers diminishing returns at 4K — just a 10 per cent boost to frame-rates on average.

Regardless, it’s at the high end that GTX 980 Ti hits its stride. Combine the new card with one of the upcoming 4K G-Sync monitors and we could be onto a winning combination, but the recent announcement of a 34-inch, curved 21:9 screen from Acer — the G-Sync enabled XR341CKA — could well see the GM200 GPUs find their true calling. It features a 3440×1440 resolution — 60 per cent of 4K resolution — but with a wider fielder of view and less extreme pixel density, making it great for productivity as well as gaming (and side-stepping Windows’ lacklustre DPI scaling). Other ‘3.4K’ panels are available of course, but as far as we know, only the Acer packs in G-Sync too.

The importance of memory bandwidth eclipses shader power at 4K, making the gap between Titan X and GTX 980 Ti even less of an issue.

3840×2160 (4K)	R9 290	R9 290X	GTX 970	GTX 980	GTX 980 Ti	GTX 980 Ti OC	Titan X
The Witcher 3, High, HairWorks Off, Custom AA	23.6	26.0	22.8	27.7	36.9	40.7	37.5
Battlefield 4, High, Post-AA	36.9	39.4	39.5	46.8	61.0	69.6	61.3
Crysis 3, High, SMAA	33.9	36.0	31.9	39.0	52.5	59.7	52.4
Assassin’s Creed Unity, Very High, FXAA	16.6	18.1	18.4	21.8	26.5	29. 0	27.4
Far Cry 4, Very High, SMAA	33.3	36.0	30.0	36.1	47.1	50.9	46.7
COD Advanced Warfare, Console Settings, FXAA	60.2	62.1	59.6	69.0	86.9	96.9	90.8
Ryse: Son of Rome, Normal, SMAA	31.2	34.0	25.7	31.5	41.7	45.6	42.2
Shadow of Mordor, High, High Textures, FXAA	41.5	44.8	35.3	42.3	54.8	59.7	54.8
Tomb Raider, Ultra, FXAA	37.6	41.0	39.0	47.0	61.3	66.0	64.6

With the performance testing done and dusted, let’s discuss heat, acoustics and overclocking. As expected, GTX 980 Ti is very similar indeed to Titan X, right down to the intermittent (but not obtrusive) ‘coil whine’ and its tendency to level out at temperatures of around 80 degrees Celsius under load. At stock speeds, the card is very quiet with the reference cooler, only becoming noticeable once you start to dabble with overclocking.

In this area, we found that the 980 Ti is able to hit and sustain core speeds in the 1400MHz area with boost active. We managed this by pushing max power draw up to its maximum 110 per cent, adding 220MHz to the core clock, and 400MHz to the RAM (taking us up to 7.8gbps in total). During our Titan X testing, we didn’t have access to over-volting, but MSI Afterburner has that functionality now and it works with GTX 980 Ti too. Unfortunately, 1400MHz or thereabouts does seem to be the core’s limit — we couldn’t push further without obvious instability. Remember, GM200 is a relatively massive chip, and typically, the larger the processor the less overclocking headroom is available. Additionally, our memory overclock was slightly less effective than our tests with Titan X, where we could hit 465MHz.

Power consumption remains impressive, and again, pretty much identical to the more expensive GM200 card in Nvidia’s line-up: we hit a peak of 362W drawn from the wall during the Metro Last Light benchmarking sequence running at 1440p on max settings. That’s almost identical to the much less capable Radeon R9 290X, once again highlighting Maxwell’s excellent power efficiency, even on the top of the line, no holds barred products.

	R9 290	R9 290X	GTX 970	GTX 980	GTX 980 Ti	GTX 980 Ti OC	Titan X
Peak System Power Draw	340W	363W	265W	265W	362W	396W	364W

The arrival of the GTX 980 Ti causes a mild ripple effect to prices elsewhere in Nvidia’s high-end GPU line-up. The GTX 980 — still a remarkably good card, by the way — drops down to $499, but the value-rific GTX 970 remains at its previous $330. In a world where price doesn’t tend to scale in a linear fashion with performance at the premium end of the market, at least with the 980 Ti you are getting an approximate 30 per cent boost in frame-rates for an equivalent amount of cash compared to the older 980. However, the fact remains that GTX 980 and the new 980 Ti are still priced somewhat at the extreme side. How long they will stay there all depends on just how fast AMD’s upcoming new Radeon actually is, and how much the firm intends to charge for it.

In the meantime, the new GTX 980 Ti will remain out of reach for many gamers — perhaps frustratingly so. But it’s worth bearing in mind that today’s halo product is tomorrow’s more mainstream product. There was a time when the original Titan and its derivatives were just as expensive as Nvidia’s current brace of ultra high-end cards. Fast forward to the arrival of the second-gen Maxwell architecture (and in particular the seismic debut of the GTX 970) and prices across the stack nose-dived. Doubtless, we’ll see the same thing happen here in the fullness of time.

In the meantime, while the development of truly generational leaps in CPU processing power seem few and far between, it’s great to see that the push for ever more powerful graphics hardware remains in full effect. Question marks remain over whether these new GPUs are being fully utilised (we really need DX12), and we still can’t help but think that GM200 should be capable of more, but that can’t take away from the many plus points of the product.

So, should you buy a GTX 980 Ti? As Ferris Bueller might say, «if you have the means, I highly recommend picking one up», but do bear in mind the more limited returns at lower resolutions, and perhaps hold fire until we see the performance level — and asking price — of AMD’s upcoming alternative.

0024

100W below TDP?
150W vs 250W

247MHz faster memory speed?
2000MHz vs 1753MHz

988MHz higher effective clock speed?
8000MHz vs 7012MHz

33.33% more VRAM?
8GB vs 6GB

Supports ray tracing?

607MHz faster GPU turbo speed?
1683MHz vs 1076MHz

Smaller 12nm semiconductors?
16nm vs 28nm

• 80. 5GB/s more memory bandwidth?
  336.5GB/s vs 256GB/s
• 128bit wider memory bus?
  384bit vs 256bit
• 896 more stream processors?
  2816 vs 1920
• 800million more transistors?
  8000 million vs 7200 million
• 56 more texture units (TMUs)?
  176 vs 120
• 32 more ROPs?
  96 vs 64

Which comparisons are the most popular?

Nvidia GeForce GTX 1070

vs

Nvidia GeForce RTX 2060

Nvidia GeForce GTX 980 Ti

vs

Nvidia GeForce RTX 3070

Nvidia GeForce GTX 1070

vs

Nvidia Geforce GTX 1660 Super

Nvidia GeForce GTX 980 Ti

vs

Nvidia GeForce GTX 1060

Nvidia GeForce GTX 1070

vs

Nvidia GeForce RTX 3050 Laptop

Nvidia GeForce GTX 980 Ti

vs

Nvidia GeForce RTX 3060

Nvidia GeForce GTX 1070

vs

AMD Radeon RX 580 9003

0004 Nvidia GeForce GTX 980 Ti

vs

MSI GeForce GTX 1050 Ti

Nvidia GeForce GTX 1070

vs

Nvidia GeForce GTX 1650 Ti Laptop

Nvidia GeForce GTX 980 Ti

vs

Nvidia Geforce GTX 1660 Super

Nvidia GeForce GTX 1070

vs

Nvidia GeForce GTX 1660

Nvidia GeForce GTX 980 Ti

vs

Nvidia GeForce GTX 1050

30003

vs

Nvidia GeForce RTX 3050 Ti Laptop

Nvidia GeForce GTX 980 Ti

vs

Zotac GeForce GTX 1660 Ti AMP

Nvidia GeForce GTX 1070

vs

Nvidia GeForce GTX 1660 Ti

Nvidia GeForce GTX 980 Ti

vs

Nvidia GeForce GTX Titan Black

Nvidia GeForce GTX 1070

vs

Nvidia GeForce GTX 1060

Nvidia GeForce GTX 980 Ti 90

80 Ti

0. 0 /10

0 Reviews of users

Functions

Price and quality ratio

/10

3 VOTES

Reviews are not

9000 10

3 Votes

Reviews not yet

performance

/10

3 Votes

003

9.7 /10

3 Votes

Reviews not yet

Reliability

10.0 /10

3 VOTES

Reviews yet not

performance

1. TCTIC FREEMEN

1506MHz

1000MHz

The graphics processing unit (GPU) has a higher clock speed.

2.turbo GPU

1683MHz

1076MHz

When the GPU is running below its limits, it can jump to a higher clock speed to increase performance.

3.pixel rate

96.4 GPixel/s

96 GPixel/s

The number of pixels that can be displayed on the screen every second.

4.flops

5.78 TFLOPS

5.63 TFLOPS

FLOPS is a measure of GPU processing power.

5.texture size

180.7 GTexels/s

176 GTexels/s

Number of textured pixels that can be displayed on the screen every second.

6.GPU memory speed

2000MHz

1753MHz

Memory speed is one aspect that determines memory bandwidth.

7.shading patterns

Shading units (or stream processors) are small processors in a video card that are responsible for processing various aspects of an image.

8.textured units (TMUs)

TMUs accept textured units and bind them to the geometric layout of the 3D scene. More TMUs generally means texture information is processed faster.

9 ROPs

ROPs are responsible for some of the final steps of the rendering process, such as writing the final pixel data to memory and for performing other tasks such as anti-aliasing to improve the appearance of graphics.

Memory

1.memory effective speed

8000MHz

7012MHz

The effective memory clock frequency is calculated from the memory size and data transfer rate. A higher clock speed can give better performance in games and other applications.

2.max memory bandwidth

256GB/s

336.5GB/s

This is the maximum rate at which data can be read from or stored in memory.

3.VRAM

VRAM (video RAM) is the dedicated memory of the graphics card. More VRAM usually allows you to run games at higher settings, especially for things like texture resolution.

4.memory bus width

256bit

384bit

Wider memory bus means it can carry more data per cycle. This is an important factor in memory performance, and therefore the overall performance of the graphics card.

5. GDDR memory versions

Later versions of GDDR memory offer improvements such as higher data transfer rates, which improve performance.

6. Supports memory debug code

✖Nvidia GeForce GTX 1070

✖Nvidia GeForce GTX 980 Ti

Memory debug code can detect and fix data corruption. It is used when necessary to avoid distortion, such as in scientific computing or when starting a server.

Functions

1.DirectX version

DirectX is used in games with a new version that supports better graphics.

2nd version of OpenGL

The newer version of OpenGL, the better graphics quality in games.

OpenCL version 3

Unknown. Help us offer a price. (Nvidia GeForce GTX 980 Ti)

Some applications use OpenCL to use the graphics processing unit (GPU) for non-graphics computing. Newer versions are more functional and better quality.

4. Supports multi-monitor technology

✔Nvidia GeForce GTX 1070

✔Nvidia GeForce GTX 980 Ti

The graphics card has the ability to connect multiple displays. This allows you to set up multiple monitors at the same time to create a more immersive gaming experience, such as a wider field of view.

5. GPU temperature at boot

Unknown. Help us offer a price. (Nvidia GeForce GTX 1070)

Lower boot temperature — this means that the card generates less heat and the cooling system works better.

6.supports ray tracing

✔Nvidia GeForce GTX 1070

✖Nvidia GeForce GTX 980 Ti

Ray tracing is an advanced light rendering technique that provides more realistic lighting, shadows and reflections in games.

7.Supports 3D

✔Nvidia GeForce GTX 1070

✔Nvidia GeForce GTX 980 Ti

Allows you to view in 3D (if you have a 3D screen and glasses).

8.supports DLSS

✖Nvidia GeForce GTX 1070

✖Nvidia GeForce GTX 980 Ti

DLSS (Deep Learning Super Sampling) is an AI based scaling technology. This allows the graphics card to render games at lower resolutions and upscale them to higher resolutions with near-native visual quality and improved performance. DLSS is only available in some games.

9. PassMark result (G3D)

Unknown. Help us offer a price. (Nvidia GeForce GTX 1070)

This test measures the graphics performance of a graphics card. Source: Pass Mark.

Ports

1.has HDMI output

✔Nvidia GeForce GTX 1070

✔Nvidia GeForce GTX 980 Ti

Devices with HDMI or mini HDMI ports can stream HD video and audio to an attached display.

2.HDMI connectors

Unknown. Help us offer a price. (Nvidia GeForce GTX 980 Ti)

More HDMI connectors allow you to connect multiple devices such as game consoles and TVs at the same time.

HDMI 3.Version

HDMI 2.0

Unknown. Help us offer a price. (Nvidia GeForce GTX 980 Ti)

Newer versions of HDMI support higher bandwidth for higher resolutions and frame rates.

4. DisplayPort outputs

Allows connection to a display using DisplayPort.

5.DVI outputs

Allows connection to a display using DVI.

Mini DisplayPort 6.outs

Allows connection to a display using Mini DisplayPort.

Price Match

Cancel

Which graphics cards are better?

Compare NVIDIA GeForce GTX 1650 Ti and NVIDIA GeForce GTX 980 Ti

Comparative analysis of video cards NVIDIA GeForce GTX 1650 Ti and NVIDIA GeForce GTX 980 Ti by all known characteristics in the categories: General Information, Specifications, Video Outputs and Ports, Compatibility, Dimensions, Requirements, API Support, Memory, Technology Support.
Analysis of video card performance by benchmarks: PassMark — G3D Mark, PassMark — G2D Mark, Geekbench — OpenCL, CompuBench 1.5 Desktop — Face Detection (mPixels/s), CompuBench 1.5 Desktop — Ocean Surface Simulation (Frames/s), CompuBench 1.5 Desktop — T -Rex (Frames/s), CompuBench 1.5 Desktop — Video Composition (Frames/s), CompuBench 1.5 Desktop — Bitcoin Mining (mHash/s), GFXBench 4.0 — Car Chase Offscreen (Frames), GFXBench 4.0 — Manhattan (Frames), GFXBench 4.0 — T-Rex (Frames), GFXBench 4.0 — Car Chase Offscreen (Fps), GFXBench 4.0 — Manhattan (Fps), GFXBench 4.0 — T-Rex (Fps), 3DMark Fire Strike — Graphics Score.

NVIDIA GeForce GTX 1650 Ti

versus

NVIDIA GeForce GTX 980 Ti

Benefits

Reasons to choose NVIDIA GeForce GTX 1650 Ti

• Newer graphics card, 4 year(s) 10 month(s) difference in release dates by about 3 5%

more: 1350 MHz vs 1000 MHz

• Boost core clock 38% more: 1485 MHz vs 1075 MHz
• 540x more texturing speed: 95. 04 GTexel/s vs 176 billion / sec
• Newer manufacturing process of the video card makes it more powerful, but with lower power consumption: 12 nm vs 28 nm
• 5 times less power consumption: 50 Watt vs 250 Watt
• Memory frequency 214.3 times more: 1500 MHz (12000 MHz effective) vs 7.0 GB/s
• About 6% more performance in Geekbench — OpenCL benchmark: 42452 vs 40025
• Performance in CompuBench 1.5 Desktop — Face Detection (mPixels) benchmark /s) about 4% more: 151.562 vs 145.843
• About 7% better performance in CompuBench 1.5 Desktop — Ocean Surface Simulation (Frames/s) benchmark: 1843.045 vs 1722.98
• 3.1x better performance in CompuBench 1.5 Desktop — Video Composition (Frames/s): 116.231 vs 37.16
• About 16% more performance in GFXBench 4.0 — Car Chase Offscreen (Frames) benchmark: 12180 vs 10471
• 2.4x more performance in GFXBench 4.0 — Manhattan (Frames) benchmark: 8926 vs 3695
• 2.4 times more performance in GFXBench 4. 0 — T-Rex (Frames) benchmark: 8062 vs 3338
• About 16% more performance in GFXBench 4.0 — Car Chase Offscreen (Fps) benchmark: 12180 vs 10471
• 2.4 times more performance in GFXBench 4.0 — Manhattan (Fps) benchmark: 8926 vs 3695
• 2.4 times more performance in GFXBench 4.0 — T-Rex (Fps) benchmark: 8062 vs 3338 9027

Issue date	2 Apr 2020 vs 2 June 2015
Core clock	1350 MHz vs 1000 MHz
Boost core clock	1485 MHz vs 1075 MHz
Texturing speed	95.04 GTexel/s vs 176 billion / sec
Process	12 nm vs 28 nm
Power consumption (TDP)	50 Watt vs 250 Watt
Memory frequency	1500 MHz (12000 MHz effective) vs 7. 0 GB/s
Geekbench — OpenCL	42452 vs 40025
CompuBench 1.5 Desktop — Face Detection (mPixels/s)	151.562 vs 145.843
CompuBench 1.5 Desktop — Ocean Surface Simulation (Frames/s)	1843.045 vs 1722.98
CompuBench 1.5 Desktop — Video Composition (Frames/s)	116.231 vs 37.16
GFXBench 4.0 — Car Chase Offscreen (Frames)	12180 vs 10471
GFXBench 4.0 — Manhattan (Frames)	8926 vs 3695
GFXBench 4.0 — T-Rex (Frames)	8062 vs 3338
GFXBench 4. 0 — Car Chase Offscreen (Fps)	12180 vs 10471
GFXBench 4.0 — Manhattan (Fps)	8926 vs 3695
GFXBench 4.0 — T-Rex (Fps)	8062 vs 3338

Reasons to choose NVIDIA GeForce GTX 980 Ti

• 2.8x more shader processor(s): 2816 vs 1024
• About 50% more max memory size: 6 GB vs 4 GB
• PassMark benchmark performance — G3D Mark about 85% more: 13852 vs 7502
• PassMark performance — G2D Mark 2.1 times greater: 852 vs 398
• CompuBench 1.5 Desktop performance — T-Rex (Frames/s) approx. at 9% more: 11.661 vs 10.685
• About 22% more performance in CompuBench 1.5 Desktop — Bitcoin Mining (mHash/s): 787.997 vs 644.011
• About 58% more performance in 3DMark Strike — Graphics Score vs Fire: 5794 3677

Number of shaders	2816 vs 1024
Maximum memory size	6 GB vs 4 GB
PassMark — G3D Mark	13852 vs 7502
PassMark — G2D Mark	852 vs 398
CompuBench 1. 5 Desktop — T-Rex (Frames/s)	11.661 vs 10.685
CompuBench 1.5 Desktop — Bitcoin Mining (mHash/s)	787.997 vs 644.011
3DMark Fire Strike — Graphics Score	5794 vs 3677

Benchmark comparison

GPU 1: NVIDIA GeForce GTX 1650 Ti
GPU 2: NVIDIA GeForce GTX 980 Ti

PassMark — G3D Mark	GPU 1 GPU 2
PassMark — G2D Mark	GPU 1 GPU 2
Geekbench — OpenCL	GPU 1 GPU 2
CompuBench 1. 5 Desktop — Face Detection (mPixels/s)	GPU 1 GPU 2	151.562 145.843
CompuBench 1.5 Desktop — Ocean Surface Simulation (Frames/s)	GPU 1 GPU 2	1843.045 1722.98
CompuBench 1.5 Desktop — T-Rex (Frames/s)	GPU 1 GPU 2	10. 685 11.661
CompuBench 1.5 Desktop — Video Composition (Frames/s)	GPU 1 GPU 2	116.231
CompuBench 1.5 Desktop — Bitcoin Mining (mHash/s)	GPU 1 GPU 2	644.011 787.997
GFXBench 4. 0 — Car Chase Offscreen (Frames)	GPU 1 GPU 2
GFXBench 4.0 — Manhattan (Frames)	GPU 1 GPU 2
GFXBench 4.0 — T-Rex (Frames)	GPU 1 GPU 2
GFXBench 4.0 — Car Chase Offscreen (Fps)	GPU 1 GPU 2
GFXBench 4. 0 — Manhattan (Fps)	GPU 1 GPU 2
GFXBench 4.0 — T-Rex (Fps)	GPU 1 GPU 2
3DMark Fire Strike — Graphics Score	GPU 1 GPU 2

Name	NVIDIA GeForce GTX 1650 Ti	NVIDIA GeForce GTX 980 Ti
PassMark — G3D Mark	7502	13852
PassMark — G2D Mark	398	852
Geekbench — OpenCL	42452	40025
CompuBench 1. 5 Desktop — Face Detection (mPixels/s)	151.562	145.843
CompuBench 1.5 Desktop — Ocean Surface Simulation (Frames/s)	1843.045	1722.98
CompuBench 1.5 Desktop — T-Rex (Frames/s)	10.685	11.661
CompuBench 1.5 Desktop — Video Composition (Frames/s)	116.231	37.16
CompuBench 1.5 Desktop — Bitcoin Mining (mHash/s)	644.011	787.997
GFXBench 4.0 — Car Chase Offscreen (Frames)	12180	10471
GFXBench 4. 0 — Manhattan (Frames)	8926	3695
GFXBench 4.0 — T-Rex (Frames)	8062	3338
GFXBench 4.0 — Car Chase Offscreen (Fps)	12180	10471
GFXBench 4.0 — Manhattan (Fps)	8926	3695
GFXBench 4.0 — T-Rex (Fps)	8062	3338
3DMark Fire Strike — Graphics Score	3677	5794

Feature comparison

	NVIDIA GeForce GTX 1650 Ti	NVIDIA GeForce GTX 980 Ti
Architecture	Turing	Maxwell 2. 0
Codename	TU117	GM200
Issue date	2 Apr 2020	2 June 2015
Place in the rating	168	172
Type	Laptop	Desktop
Price at first issue date		$649
Price now		$679.99
Price/performance ratio (0-100)		23.43
Boost core clock	1485 MHz	1075MHz
Core clock	1350 MHz	1000 MHz
Process	12nm	28nm
Peak Double Precision (FP64) Performance	95. 04 GFLOPS (1:32)
Peak Half Precision (FP16) Performance	6.083 TFLOPS (2:1)
Peak Single Precision (FP32) Performance	3.041 TFLOPS
Number of shaders	1024	2816
Pixel fill rate	47.52GPixel/s
Texturing speed	95.04 GTexel/s	176 billion / sec
Power consumption (TDP)	50 Watt	250 Watt
Number of transistors	4700 million	8,000 million
Number of CUDA conveyors		2816
Floating point performance		6. 060 gflops
Video connectors	No outputs	1x DVI, 1x HDMI, 3x DisplayPort, Dual Link DVI-I, HDMI 2.0, 3x DisplayPort 1.2
Audio input for HDMI		Internal
G-SYNC support
HDCP
Maximum resolution VGA		2048×1536
Multi-monitor support
Interface	PCIe 3. 0 x16	PCIe 3.0 x16
Additional power connectors	None	6-pin + 8-pin
Tire		PCI Express 3.0
Height		4.376″ (11.1 cm)
Length		10.5″ (26.7cm)
Recommended power supply		600 Watt
SLI support		4x
DirectX	12.1	12. 0 (12_1)
OpenCL	1.2
OpenGL	4.6	4.5
Shader Model	6.5
Vulcan
Maximum memory size	4GB	6GB
Memory bandwidth	192.0 GB/s	336.5 GB/s
Memory bus width	128bit	384 Bit
Memory frequency	1500 MHz (12000 MHz effective)	7. 0 GB/s
Memory type	GDDR6	GDDR5
Shared memory		0
3D Vision
Adaptive Vertical Sync
CUDA
GameStream
GameWorks
GeForce Experience
GeForce ShadowPlay
GPU Boost
SLI
Surround

Review Nvidia GeForce GTX 980 Ti

Pros

• Fantastic, stable 4K performance
• £250 cheaper than Titan X
• Pretty Cool Operation

Cons

• Overkill for 1080p and 1440p games

Key features

• Review Price: £550. 00
• Core clock 1000 MHz
• 1753 MHz 6 GB GDDR5 memory
• 8 billion transistors
• 2816 stream processors
• Requires 1 x 6-pin and 1 x 8-pin power connectors
• Manufacturer: Nvidia

What is Nvidia GeForce GTX 980 Ti?

Updated: The GeForce GTX 980 Ti is the second most powerful graphics card in Nvidia’s latest generation of Maxwell graphics cards, sitting under the Titan X in terms of price and performance. With 4K capability and VR Ready certification, this is an amazing graphics card that keeps getting cheaper.

Since our initial review of the 980 Ti, Nvidia has released its first Pascal-based GPU, the GeForce GTX 1080. The GTX 980 Ti is one of the few GTX 1080 alternatives that actually looks good compared to the new card. Some of this is down to its price; You can buy a used GTX 980Ti for £400 or a new model for £550 and prices will certainly drop. Performance in the latest games running at 4K is still excellent, especially with an aftermarket card like the overclocked version released by EVGA. You can see the full list of the latest benchmarks in our GTX 1080 review, which also tests the aftermarket GTX 980 Ti.

With all that said, if the 980 Ti prices don’t drop in huge volumes in the coming months, it’s going to look like a much poorer deal, so it’s worth looking for a better price if you want to save some money.

Below is our original review written in July 2015.

Nvidia GeForce GTX 980 Ti — Under the Hood

This card’s Ti suffix says it shares DNA with the cheaper GTX 980, but it doesn’t — GTX 9The 80 Ti is built around the same GM200 Maxwell core found inside the mighty Titan X.

This means the new card’s specification is much closer to Nvidia’s best GPU than the cheaper stock GTX 980. It has 2816 stream processors, second only to Titan 256, and six GPU clusters and 24 streaming multiprocessors. That last number is only half that of the Titan X, but six more than that of the GTX 980.

0759

Reworking the GM200 core means the GTX 980 Ti also has 8 billion transistors and 601mm ² die — nearly double the chip inside the GTX 980. The MHz is the same as the Titan, though the 1126MHz GTX 980 takes the lead in this area.

However, there is one area where the GTX 980 Ti lags behind the more expensive Titan X: memory. Nvidia’s Titan X has 12GB of GDDR5 RAM, but GTX 980 Ti costs half as much. It still runs at the same 1753MHz, and it’s still plenty — most cards don’t even come close to 6GB, let alone double that number.

In fact, this may matter, but it is unlikely to be noticeable. This is more for use in high-performance computing applications, while the 980 Ti is more suitable for gaming.

So, in practical terms, expect performance closer to Titan X. On paper, this makes 9The 80 Ti seems like a £550 bargain: much closer to the £400 GTX 980 than the Titan X, which typically retails for over £800. Both feature Nvidia’s aluminum cooler design, and both require single six- and eight-pin power connectors.

GTX 980 Ti needs a larger case: Nvidia’s reference model is 267mm long. Unlike AMD Fury X is much shorter as it uses a separate liquid cooler.

TrustedReviews Awards 2015: Winners Announced

AMD’s older line has nothing to compete with the GTX 980 Ti, and its closest competitor is nowhere near even the GTX 980. However, the new R9 series is expected to fight harder .

The closest contender will be the AMD Radeon R9 Fury X, which will cost around £510. It’s going to be an intriguing battle: the AMD card has just 4GB of RAM, but it has more stream processors and transistors than the Nvidia card, as well as a faster core.

Nvidia GeForce GTX 980 Ti — how we tested

We loaded five games for this GPU test. Battlefield 4, BioShock Infinite and Crysis 3 are all returning from our previous reviews, and we’ve added Metro: Last Light and Batman: Arkham Origins to the mix. We tested at 2560 x 1440 and 3840 x 2160 to see how the GTX 980 Ti performs with high resolution single screens. We haven’t tested at 1080p as we know this card is powerful enough to handle any game at this low resolution. 9SEE ALSO: Best Gaming Headsets coolest and most economical.

Our test rig consists of an Asus X79-Deluxe motherboard, an Intel Core i7-4960X processor, 16GB of RAM and a 1TB hard drive.

For prices, we visited the scanning website and wrote down the cheapest stock card we could find, although we will refer to the various overclocked and optimized models available for each GPU — which will be more expensive — later in the review.

Mike has worked as a technology journalist for over a decade and has written for most of the UK’s most prominent websites and magazines. While writing articles about technology, he developed obsessions…

Unlike other sites, we thoroughly test every product we review. We use standard industry benchmarks to properly compare features. We will always tell you what we find. We never, ever accept money for a product review.
Tell us what you think — send your letters to the editor.

Review NVIDIA GTX 1080 vs GTX 980 Ti, Pascal vs Maxwell, GP104 vs GM200 (EVGA GeForce GTX 980 Ti HYBRID 06G-P4-1996-KR)

Yesterday we published a detailed test of the first video card based on Pascal — NVIDIA GeForce GTX 1080. The new process technology, technologies and architecture showed impressive results. NVIDIA managed to make a noticeable breakthrough, leaving AMD solutions far behind, including cards with new HBM memory, and a noticeable advantage over MAXWELL solutions of the GTX 9 series was also announced00. Moreover, this time the opposition was carried out at the presentation not only from the GTX 980 to the GTX 1080, but even TITAN X. The prospects are tempting, given that it still costs comparable to the GTX 980 Ti.

We have NVIDIA GeForce GTX 1080 Founder’s Edition (review) and EVGA GeForce GTX 980 Ti HYBRID (review) . Today we offer to compare these two video cards using a system built on the Intel Core i7 6700K processor, which reveals the full potential of video cards. As part of this material, we will not repeat the innovations of the NVIDIA GeForce GTX 1080 Founder’s Edition, we offer you to look at its separate review. Let us recall only the relevant moments.

NVIDIA GTX 1080 vs GTX 980 Ti review

The NVIDIA GeForce GTX 1080 Founder’s Edition is based on the GP104 chip with GDDR5X memory. The EVGA GeForce GTX 980 Ti HYBRID is based on the GM200 chip with GDDR5 memory. With a smaller bus width, 256 bits versus 384 bits, the NVIDIA GeForce GTX 1080 has more bandwidth, and new color compression technologies must be taken into account here. 6 vs. 8 GB of graphics memory.

The transition to a new process technology 16 nm from 28 nm, with a smaller die area, has a similar number of CUDA cores 2560 versus 2816. Peak computing power is 8873 GFLOPs versus 6144 GFLOPs. There are differences in the number of rasterization units: GP104 — 64 units, GM200 — 96 units. This difference will be reflected in the level of performance when working with 4K resolution.

The EVGA GeForce GTX 980 Ti HYBRID draws up to 250W, requiring power through the 8- and 6-pin connectors. The NVIDIA GeForce GTX 1080 Founder’s Edition has modest appetites, drawing 180 watts through a single 8-pin connector.

At the time of the release of this comparative review, no versions with proprietary cooling were announced. But we can assume that between the GTX 980 Ti and GTX 1080 will have a lot of similarities, both in appearance and cooling design.

Among the differences is NVIDIA’s bet on reference cooling, while it will cost $100 more than the cheapest options from partners.

Test NVIDIA GTX 1080 and GTX 980 Ti

Model	Data
Case 905 Define8 S905 Design80589
Operating system	Windows Pro 10

Game data is captured in Full HD resolution, similar settings for each of the games with ultra textures and maximum anti-aliasing. For a visual comparison, we present video clips, we offer a visual comparison of the image in games and trace the dynamic change in FPS.

3DMark11

3DMark Fire Strike Extreme

3DMark Fire Strike Ultra

All measurements in games are carried out in Full HD resolution with ultra graphics settings.

DirectX 12: Ashes of the Singularity

Ashes of the Singularity is a strategy game from Stardock developed specifically for DirectX 12 and 64-bit systems.

Crysis 3

Hitman Absolution

Tomb Raider

Batman: Arkham Knight

Middle-earth: Shadow of Mordor

GTA 5

Assassins Creed Syndicate

FarCry 4

Tom Clancy The Division

NVIDIA GTX 1080 or GTX 980 Ti?

NVIDIA GeForce GTX 1080 Founder’s Edition in Full HD resolution showed an advantage over EVGA GeForce GTX 980 Ti HYBRID in both modern games and synthetic benchmarks. It is inferior only in a couple of games: GTA 5 and FarCry 4. In the near future we will compare these cards in 4K, here, as we can assume, the advantage will be on the side of the GTX 980 Ti, as it will be in reality, we will find out further. In the meantime, we can state that NVIDIA engineers do not eat their «bread» in vain. Implemented 16nm process technology, increased core and memory frequencies, new image processing engine optimized for multiple monitors and VR, SLI with memory pooling, compression technologies, reduced power consumption, improved asynchronous computing.

Review and testing of ASUS ROG POSEIDON GTX 980 Ti video card GECID.com. Page 1

::>Video cards
>2015
> ASUS ROG POSEIDON GTX 980 Ti

10/28/2015

Page 1
Page 2
One page

ASUS ROG series graphics cards have always attracted users with their high performance, innovation, quality and originality. ASUS has decided to expand this line with a model based on one of the highest performance graphics cards on the market — NVIDIA GeForce GTX 980 Ti, the key features of which you can find in a separate article.

Video card ASUS ROG POSEIDON GTX 980 Ti (POSEIDON-GTX980TI-P-6GD5) which will be appreciated not only by gamers, but also by those who like to get maximum performance through manual overclocking. It is equipped with a DirectCU h3O hybrid cooler that can work as an air cooling system or as part of a CBO circuit, which allows you to achieve much higher performance when overclocking.

Specification table for ASUS ROG POSEIDON GTX 980 Ti:

Model	ASUS ROG POSEIDON GTX 980 Ti (ASUS POSEIDON-GTX980TI-P-6GD5)
Graphics core	NVIDIA GM200-310 (Maxwell)
Number of CUDA cores	2816
Supported APIs and Technologies	DirectX 12 (Shader Model 5. 0), OpenGL 4.5, NVIDIA 3D Vision Surround, NVIDIA PhysX, NVIDIA TXAA, NVIDIA FXAA, NVIDIA MFAA, NVIDIA 3D Vision, NVIDIA Adaptive Vertical Sync, NVIDIA SLI, NVIDIA CUDA, NVIDIA GPU Boost 2.0, NVIDIA GameWorks, NVIDIA GameStream
Rated / dynamic frequency of the graphics core, MHz	1114 / 1203
Memory frequency (effective), MHz	1753 (7010)
Memory capacity, GB	6
Memory type	GDDR5
Memory bus width, bit	384
Memory bandwidth, GB/s	336.6
Tire type	PCI Express 3.0 x16
Maximum resolution	Digital — up to 5120 x 3200 Analog — up to 2048 x 1536
Image output interfaces	1 x DVI-I 1 x HDMI 2. 0 3 x DisplayPort
Minimum power supply, W	600
Additional PCIe 9 power connectors0003	2 x 8-pin
Dimensions from the official website (measured in our test lab), mm	287 x 137 x 40.6 (295 x 138)
Drivers	Latest drivers can be downloaded from the ASUS website or the GPU manufacturer’s website
Manufacturer website	ASUS

Since we have an almost completely rethought modification of the NVIDIA GeForce GTX 980 Ti, we can’t speak of any correspondence to the reference model. Only operating frequencies can be directly compared. In this case, the graphics core of the video card boasts a noticeable level of factory overclocking — up to 1114 MHz (an increase of 11. 4% relative to the recommended 1000 MHz), and taking into account the work of NVIDIA GPU Boost 2.0 technology, the average value is 1203 MHz (instead of 1075 MHz for the reference sample) . The video memory runs at the recommended 7010 MHz.

Packing and contents

The ASUS GeForce GTX 980 Ti POSEIDON PLATINUM video card comes in a large box corresponding to its flagship status, made of thick cardboard and decorated with laconic and stylish printing with the ASUS ROG line logo. In addition to the name of the manufacturer and device model, we note the mention of support for NVIDIA G-SYNC, NVIDIA GameWorks, NVIDIA GameStream and DirectX 12 technologies.

On the reverse side of the hinged lid, under which there is a small viewing window, there is a description of the main advantages of the novelty, namely: and ensures the rejection of aggressive chemicals in the production process. In addition, it is based on a high-quality element base (Super Alloy Power II), which makes it possible to further increase the reliability of the novelty.

ROG LED Light – Modding enthusiasts will appreciate the ASUS ROG series logo on the side of the graphics card, which is equipped with LED lighting.

The reverse side of the box bears an image of the interface panel and a description of the hybrid cooling system: composition of the SVO circuit.

G1/4 — fittings in the water block have a G1/4″ threaded connection. This is the most popular standard, so it will be easy to select the appropriate fittings for connecting the CBO circuit.

Dual dust-proof fans — the cooling system includes two fans with dust-proof technology, which increases the life of the cooler by 25%.

The list of system requirements for the computer where you plan to install the video card is located on one of the sides of the box. Based on the recommendations, the power supply in such a system must have a power of at least 600 W and deliver at least 42 A along the + 12V line. The tested model also requires two 8-pin PCIe cables.

Included with the graphics adapter, we found the standard documentation, software CD and power adapter.

To output an image on the tested novelty, a reference set of interfaces is used:

• 1 x DVI-I;
• 1 x HDMI 2.0;
• 3 x DisplayPort.

The following resolutions are supported:

• digital — up to 5120 x 3200;
• analog — up to 2048 x 1536.

Appearance and element base

The ASUS POSEIDON-GTX980TI-P-6GD5 model is made on an original black printed circuit board with a reference layout principle. The used element base corresponds to the Super Alloy Power II proprietary concept and includes exclusively high-quality components: solid-state and tantalum-polymer capacitors, improved MOSFETs and ferrite core chokes. This improves the stability and reliability of the graphics adapter as a whole, and also extends its service life.

The novelty is powered by a reinforced 10-phase scheme, where 8 phases are responsible for powering the graphics core, and 2 are for the video memory subsystem. Recall that the reference NVIDIA GeForce GTX 980 Ti is powered by an 8-phase scheme.

The ASP1500U chip is used as a digital PWM controller for the power subsystem.

The graphics adapter under test is powered by a PCI Express x16 slot and two 8-pin PCIe connectors located on the side of the board. Due to its large size, the cooler makes it a little difficult to disconnect PCIe cables. In turn, the reference version manages only one reinforced 8-pin and one 6-pin connector.

To ensure the operation of NVIDIA SLI technology, two connectors are used to connect the corresponding bridges, which allow you to combine up to three video accelerators for joint calculation of graphic effects.

The reverse side of the novelty, originally covered by a base plate, is notable for a number of electrical components.

The tested model is based on the NVIDIA GM200-310 (Maxwell) GPU, manufactured using the 28-nm process technology. It includes 2816 CUDA cores, 96 rasterization units and 176 texture units. As we have already said, the video card is distinguished by factory overclocking of the graphics core, the operating frequencies of which were 1114 and 1203 MHz.

The memory of the ASUS POSEIDON-GTX980TI-P-6GD5 model, with a total capacity of 6 GB, is assembled using 12 SK hynix H5GQ4h34MFR-R2C chips with a capacity of 4 Gb each. According to the documentation, their effective frequency is 7000 MHz. The exchange of data between the GPU and memory is carried out through a 384-bit bus, which is capable of passing 336.6 GB of information per second.

Cooling system

The video card with DirectCU h3O cooling system installed occupies two expansion slots and has a total length of 287 mm according to the official ASUS website (295 mm according to measurements in our test lab).

The cooler consists of a fairly massive heatsink, which uses 44 longitudinal aluminum plates, and two 86-mm axial fans mounted on a plastic casing.

The turntables themselves are manufactured by FirstDO and are marked «FDC10h22S9-C». The nominal voltage of their operation is 12 V, and the current strength is 0.35 A, which gives a total power of 4.2 watts.

Three copper heat pipes of different diameters are used to evenly distribute heat over the entire area of ​​the radiator: two 6 mm and one 8 mm. The heatsink fins and heatpipes are covered with a layer of nickel, which is designed to reduce the drop in CO efficiency during operation caused by the oxidation of metals.

At the base of the heatsink there is a fairly large copper evaporation chamber fixed inside the base plate, on which a thermal interface is applied to cool the memory chips, power subsystem elements and the PWM controller.

The operation of the CBO is implemented using a 6 mm water channel passing through the copper evaporation chamber.