AMD Vs Intel | HP® Tech Takes
Choosing a Central Processing Unit (CPU) is the most significant decision you can make when purchasing a new laptop or PC. While you can quickly get lost in the technical details of CPU benchmarks, specifications, and architectures, your choice of manufacturers is a binary one: AMD vs. Intel®.
Statistics
show that AMD and Intel combine for the entire global x86 CPU market. The two companies have been locked in competition for so long that they have developed loyal fanbases online, the red team (AMD) and blue team (Intel), named after the color of each company’s logo.
However, the vast majority of us are not wedded to a particular CPU manufacturer. Therefore, choosing between AMD vs. Intel depends on your specific needs and budget. So, let’s break down the key differences between AMD and Intel CPUs and help you find the best processor for you.
What is a CPU?
Before we go any further, though, let’s quickly explain what a CPU is.
The CPU is the core component within the computer responsible for executing instructions. It is often referred to as the computer’s brain, performing the actual computations to run all of its software and interact with hardware.
A CPU is a silicon chip built of billions and billions of tiny transistors rapidly flipping back and forth between 0 and 1, crunching the numbers to get a computer to respond to the user’s input, output results, and write data to the computer’s memory and storage. The calculations the CPU performs translate into everything you do with your device, whether watching a video, writing an email, or browsing the internet for a new CPU.
In all these instances, the CPU takes instructions from various pieces of software (video player, email client, web browser, etc.) and outputs the desired response. If you’re on a windows laptop or desktop and want to know what your CPU is doing right now, press «CTRL» + «SHIFT» + «ESC» to open the task manager. You’ll see a list of all the processes currently running, along with their CPU usage (%).
The exact design of CPU chips defines how it operates. AMD and Intel’s research and development teams are always looking for new and improved ways to pack more transistors into a smaller space and boost computation power.
Each CPU comprises multiple cores, allowing it to focus on more than one task simultaneously. Therefore the more cores a CPU has, the more powerful it is. In addition, many processors utilize multithreading that further splits individual cores into virtual ones that can work on multiple processes simultaneously.
Today, there are also different CPU cores, including E cores (efficiency) and P cores (Performance). CPUs require a lot of power to run. This eats into battery life and can cause overheating issues (no one like a loud fan). To better trade-off these factors against overall performance, modern CPUs can utilize E cores for low computing power, often background applications, and leave P cores for intensive processes.
The overall performance of a CPU is measured by «clock speed.» This parameter, measured in gigahertz (GHz), tells you how many operations or calculations a processor can perform each second.
AMD processors
Sometimes seen as the underdog in the industry, AMD began life as a licensed chip manufacturer for larger companies, including Intel. However, in the early 2000s, AMD started to make a name for itself by releasing its own CPUs and growing its market share, particularly with the Ryzen™ range.
AMD Ryzen, based on its Zen architecture, was first released in 2017 and quickly became a popular CPU option. There are now four generations of Ryzen:
Ryzen 3: AMD’s most affordable range with up to 4-core processors
Ryzen 5: A great trade-off between performance and price with up to 6-core processors
Ryzen 7: High-power CPUs with up to 8-core processors
Ryzen 9: Top-of-the-range CPUs with up to 16-core processors
When comparing processors, the largest barrier is often getting your head around AMD and Intel’s naming conventions.
When purchasing one of the four Ryzen generations, you will also be met with a model number, for example, AMD Ryzen 9 5950X.
While it is a simplification, generally, the higher the model number, the faster the CPU is. The first digit of the model number corresponds with its series and the architecture it is based on:
- 2000 series Zen+ processors were first released in 2018
- 3000 series Zen 2 processors were first released in 2019
- 5000 series Zen 3 processors were first released in 2020
Additional suffixes following the model number offer further information. For example, «G» means the CPU includes Vega video processing, and «X» means it is a slightly faster version of the model.
The Ryzen 9 generation is not the fastest AMD offers. They also have a line of AMD Ryzen Threadripper CPUs, which AMD calls the most powerful desktop processor in the world with 64 cores and 128 threads.
The chips are also much larger in size to pack the extra power in.
Ryzen Threadripper chips are for creative professionals who need tremendous power for high-core count applications like video editing and rendering.
Intel processors
The older and larger company of the two, Intel has been around since 1968 and is the largest semiconductor company in the world by revenue. While Intel has been making processors for over 50 years, its range now comes down to four main brands:
Core™ i3: A significant step up from previous Pentium processors. Core i3 processors offer a great mix of price and performance with between 2 and 4 cores.
Core i5: Jump in quality that allows users to dip their toes into intensive processes, such as photo or video editing. Available with up to 6 cores.
Core i7: High-performance processors for intensive workloads. Available with up to 10 cores.
Core i9: Intel’s top processor range, generally meant for specialist machines required to perform challenging tasks.
Available with up to 16 cores.
Again, to know what you’re looking at, you need to go beyond the brand and learn the naming conventions for the specific series and models.
Intel processor names look something like «Core i5-1145G7.» The first two digits after the hyphen are the generation number. Therefore, the example above is the 11th generation. Each generation also has a codename. The latest 12th-generation chips are known as «Alder Lake.»
The following two digits are the position in the range, and bigger is better. Finally, the suffix tells you additional information. For example, the «G7» from above means that the processor has a built-in graphics chip. However, there are a range of other suffixes corresponding to different capabilities, such as «T» for smaller chips with slightly lower clock speeds and «K» for higher performance with a small increase in clock speed.
Comparing AMD vs. Intel processors
In the 2010s, Intel’s Core series became the dominant CPU on the market.
However, AMD has made an impressive comeback with Ryzen. The best AMD Ryzen processors could go head-to-head with Intel, often coming out on top.
But a company as big as Intel isn’t staying quiet for long, recently releasing the new Alder Lake 12th generation CPUs. So now the question remains AMD Ryzen or Intel, which is better?
Looking at pure processing power, Intel’s Alder Lake now sets the standard. However, the top-end clock speeds are only relevant for people needing the best performance possible, regardless of price. The question of Ryzen vs. Intel is more nuanced and depends on user requirements, budget, and value for money.
Pricing
When you compare CPU ranges, AMD has had a strong case for offering the best value at budget-friendly prices. AMD also has advantages with overclocking its models and board compatibility. AMD’s chips come with the AM4 CPU socket, which is compatible with most motherboards.
In comparison, many Intel CPUs require purchasing newer, more expensive motherboards to overclock, which can increase your outlay significantly.
When you compare the companies’ CPU ranges and chips, generally speaking, the AMD Ryzen comes at a slightly lower price. However, this gap has narrowed considerably, and prices vary greatly depending on the seller. As a result, it is now not unusual to also see an AMD chip going for more than the Intel equivalent.
Going back to the top end, even if you consider the cost of a new motherboard, Intel’s Alder Lake CPUs outperform the best Ryzen chips on value (performance per dollar).
Laptop CPUs
For laptop CPUs, Intel has a significant lead compared to AMD. While you can find AMD laptops, given Intel’s more extensive portfolio, it is more common to find laptops with Intel processors of various generations, often combined with integrated graphics.
So which processor is best for laptops in terms of performance?
Again, the 12th generation Alder Lake Intel CPUs take the prize. Looking at CPU benchmark tests, the best processor for laptops are currently:
- Intel Core i7-12800HX
- Intel Core i9-12900HX
- Intel Core i9-12950HX
While Intel currently has the most powerful CPUs, that doesn’t mean there aren’t many great AMD options.
Both companies have an excellent range of laptop options, and it is essential to look into the specifications you need to handle your specific needs.
Until you reach the high end, many AMD and Intel chips have equivalent performance for the same price.
Summary
Both Intel and AMD have been doing this for a long time, and with a little bit of investigation, you can find the perfect CPU for your needs at a reasonable price point.
But the CPU market is never static, and both companies are continually working on their next-generation chips. The Intel Raptor Lake and AMD Zen 4 processors are set to launch in 2022, ensuring that the AMD vs. Intel battle will likely continue.
About the Author
Arthur Smalley is a science and technology writer based in the UK.
Popular HP Desktops:
Intel Core i5-13600K vs. AMD Ryzen 5 7600X
The Core i5-13600K is the most affordable model in Intel’s new Raptor Lake (K-SKU) range. We’re very excited about the prospects of this CPU because the 12600K was a personal favorite of ours, so with the 13600K promising even better gaming performance, it should be a real winner.
Intel lists an MSRP of $330 (per 1,000 unit price) and we’re happy to report that retailers have been able to list it for $330. That makes the Core i5-13600K just $30 more expensive than the Ryzen 5 7600X and $70 cheaper than the 5800X3D.
As for specifications, there are some noteworthy changes when compared to the 12600K. The P-cores have seen an increase in L2 cache capacity to 2 MB, while the E-cores see an even bigger upgrade, not only are there twice as many at 8, but the L2 cache has been upgraded from 2 MB per cluster to 4 MB for a total of 8 MB. Even the L3 cache capacity has seen a bump, going from 20 MB to 24 MB.
In total, the i5-13600K packs a massive 49% more cache than the 12600K. Not only that, but the peak core clock speed has also increased by 4% from 4.9 GHz to 5.1 GHz. Moreover, DDR5 memory support has been improved from DDR5-4800 with 12th-gen to DDR5-5600 for 13th-gen, though DDR4 sees the same 3200 support. There’s also 16 lanes of PCI Express 5.
0 and 4 lanes of PCIe 4.0 from the CPU.
As with previous reviews of Raptor Lake chips, a word about power limits, or lack thereof. It’s important to note that although Intel claims a 181 watt max turbo power for the 13600K, that’s not enforced (and by Intel’s own guidance it’s not meant to be enforced). This means all Z690 and Z790 boards run 12th and 13th-gen Core series processors without any power limits in place.
This has been an Intel issue for a long time and it’s become an AMD issue as well. The Zen 4 CPUs ran far more efficiently when power limited, sacrificing only a small amount of performance for significantly lower power consumption which results in lower operating temperatures. But this isn’t the out of the box performance as both AMD and Intel are gunning for top spot.
Anyway, we didn’t power limit the AMD Zen 4 CPUs, so we’re not going to for Intel either, rather we’ll focus on showing you what the default behavior is for these chips.
For testing and benchmarking we have some Intel 10th-gen and 12th-gen Core processors including popular Core i5, i7 and i9 models for comparison in case you are planning to upgrade.
All 12th and 13th-gen CPUs were tested using DDR4-3600 dual-rank CL14 memory and DDR5-6400 single-rank CL32 memory. All of this testing has been updated for this review, so all the gaming data is fresh.
From AMD we’ve also got a range of Zen 2 and Zen 3 CPUs tested using 32GB of DDR4-3600 CL14 dual-rank memory on the MSI MPG X570S Carbon Max WiFi motherboard running the latest BIOS. The Zen 4 AM5 test system is based on the MSI MEG X670E ACE using 32GB of DDR5-6000 CL30 memory, we have included all the Zen 4 CPUs released thus far.
All our CPU gaming data has been updated using the Nvidia GeForce RTX 4090, testing at 1080p and 1440p, so there’s a lot to go over. Also please note we’re using Windows 11 and resizable bar has been enabled for all configurations.
Clock Speeds
First, here’s a look at the i5-13600K clock behavior in Cinebench R23. After an hour of load testing the 13600K maintained an all-core frequency of 4.9 GHz for the P-cores and 3.9GHz for the E-cores, this is down from 5.
1 GHz for the P-core when the CPU is cool, so we still saw a 4% decline in performance due to thermal throttling.
On that note, this frequency was sustained using the MSI CoreLiquid S360 360mm liquid cooler installed inside the MSI Prospect 700R case.
Then for single core workloads, the 13600K appeared to maintain a clock frequency of 5.1 GHz underload, so the advertised frequency was achieved in this test.
Application Benchmarks
Starting with the Cinebench R23 multi-core results, we see that the 13600K is good for a score of roughly 24,000 pts after a 10 minute loop, making it a whopping 54% faster than the 7600X, an embarrassing margin there for AMD, and a quick indication that productivity is going to be a one-sided battle.
The single core performance of those P-cores is also very good as the 13600K managed to nudge ahead of the 12900K and 7700X, making it a few percent faster than the 7600X. This again suggests that best case the 7600X will only match the 13600K in productivity benchmarks, and even then this will be a rare occurrence.
Firing up the 7-Zip File Manager compression benchmark sees the 13600K easily dominating the 7600X, delivering 36% greater performance and that places it alongside the likes of the 5950X and 12700K, not bad for the new $330 processor, not bad at all.
The decompression results are more favorable for AMD, but even so the new Core i5 still easily defeats the 7600X by a convincing 20% margin and is only 7% slower than the 7700X.
Moving over to the Blender Open Data benchmark we find that the 13600K is almost 40% faster than the 7600X, completing this workload in just 625 seconds opposed to 862 seconds. Yet another comfortable win for Intel’s new Core i5 processor.
It’s even worse in Corona as the 13600K fairs even better in short burst workloads, here it was 51% faster than the 7600X, completing the task in just 61 seconds, making it slightly faster than the 12700K, and roughly on par with the Ryzen 9 5900X.
Even in Premiere Pro the 13600K led the 7600X, this time by an 11% margin and that saw it match the 5950X with a score of 1084 pts.
Here is a rare example of the 7600X outpacing the 13600K in a productivity workload, though performance overall was about the same with a small 4% difference in it. Both delivered 12700K-like performance, so needless to say both CPUs are very fast.
The 13600K jumped ahead of the After Effects benchmark, leading the 7600X by a reasonably slim 6% margin.
The last application benchmark we have for you looks at code compilation performance and here the 13600K is comfortably back out in front of the 7600X, winning this test by a 16% margin, coming in just behind the 5900X.
Gaming Benchmarks
Time for some gaming benchmarks and we’ll start with Factorio as we’ve been doing as of late. Here the 13600K and 7600X are surprisingly close, but we guess that makes sense as the E-cores aren’t particularly useful here. This allowed the 7600X to take the lead, albeit by a very slim margin.
Next up we have Watch Dogs: Legion at 1080p and please note we have recorded all the 1440p data, but rather than go over that here as it is quite similar to the 1080p results, we’ll just show the 1440p average after we’ve gone over all the games individually.
Here we see that the 13600K is good for 163 fps on average using DDR5 memory or 137 fps with DDR4, making the DDR5 memory 19% faster which is a significant difference. This also means when using DDR5 the 13600K was 5% faster than the 7600X, not a huge margin but still a win for Intel here. The 13600K also wasn’t much slower than the 7700X as it managed to match the previous generation flagship 12900K, so very impressive stuff.
In Rainbow Six Extraction DDR5 memory is of little benefit but even so the 7600X was 8% faster than the 13600K and a massive 21% faster when comparing 1% lows.
Still as disappointing as those margins are, the performance on offer from the 13600K is mighty, I doubt many will scoff at over 400 fps on average.
The 7600X also came out ahead in Hitman 3, though only by a slim margin, we’re talking just 4%. So performance is much of a muchness in this title, and the 13600K was only 6% slower than the 13900K.
The 13600K did get the better of the 7600X in Far Cry 6, boosting 1% lows by 8% despite both delivering the same 180 fps on average.
So again it won’t matter which of these CPUs is used for playing this title.
The 13600K also edged out the 7600X in F1 22, boosting the average frame rate by 3% and the 1% lows by a much more significant 10%.
Of course, both CPUs delivered more than enough performance in this title, and it was impressive to see the 13600K matching the 12900K.
Despite the 13900K dominating in Spider-Man Remastered, the 13600K wasn’t nearly as impressive with 118 fps on average, meaning it could only roughly match the 7600X, though truth be told it wasn’t much slower than the 13700K, so overall a very good result.
The new Core i5 was a bit more dominant in Shadow of the Tomb Raider, and although it only knocked off the 7600X by a 4% margin when comparing the average frame rate, it was a much more convincing 12% faster when looking at the 1% lows.
The 13600K gets steamrolled in Horizon Zero Dawn by a 28% margin in favor of the 7600X, it’s bizarre how well the Zen 4 CPUs perform in this game, it’s certainly an outlier in our limited 12 game testing, but a strong result for AMD all the same.
The 13600K and 7600X were far more competitive in Cyberpunk 2077, basically delivering identical performance, making them comparable with the 7950X, 7900X, 7700X and 12900K.
The 7600X did outgun the 13600K by a slim margin in ACC, we’re only talking about a 4% margin though so overall they’re very similar. In short the 13600K was a fraction faster than the 12900K, and only slightly slower than the 13700K.
The 13600K wins The Riftbreaker benchmark when armed with DDR5 memory, beating the 7600X by a 6% margin with an average of 195 fps on average, making it just a fraction slower than the 7700X, 13700K and 13900K.
Finally, we have the CS:GO results and a bit like Horizon Zero Dawn this is a game where the 7600X easily beats the 13600K, though the margins aren’t nearly as extreme at 13%.
The 13600K was slightly slower than the 12900K, but faster than the 12700K, so a good result when compared to the previous generation Core i7 and i9 parts.
12 Game Average
Here’s our look at the gaming average at 1080p, and the Ryzen 5 7600X managed to come out on top, at least when comparing the average frame rate where it was a mere 3% faster. The 1% lows were identical at 175 fps though, and the takeaway should be that the overall gaming experience is virtually identical using either CPU.
The 13600K did drop off a little with the DDR4-3600 CL14 memory, but we’re only talking about a mere 6% decline for the average frame rate and 8% for the 1% lows. Of course, at times the margins were significant, so it will depend on the games you play.
At 1440p we see similar results, where the R5 7600X was 4% faster when comparing the average frame rate, but not faster for the 1% lows.
In short, the 13600K was comparable to the 12900K and a bit faster than the 12700K, which is a very positive comparison for the new Core i5 processor.
Power Consumption
For all-core workloads, and similar to all other 13th-gen Core CPUs, the 13600K is a bit of a pig on power, pushing total system usage to 315 watts, which certainly isn’t crazy, but it’s also a little over 70 watts more than the 5900X for slightly less performance.
It’s also slightly more than the 7900X despite the Zen 4 part being almost 90% faster. So power efficiency isn’t good, but it’s manageable and that’s about the most positive thing you can say.
Cooling Performance
Using the MSI CoreLiquid S360 360mm liquid cooler, the Core i5-13600K peaked at TjMAX after less than a minute of all-core load, so 100c. That said, despite running extremely hot, this only reduced the all core frequency by 200 MHz, from 5.1 GHz at the start of our testing to 4.9 GHz once at TjMAX.
So while annoying, it doesn’t have a significant impact on performance. Oddly this is slightly worse than the result we received with the 13700K and we should note that by default the 13600K ran at a sustained package power of 186 watts, with a peak package power of 200 watts.
If we load the 181 watt profile in the MSI BIOS, this reduces the PL1 and PL2 rating to 181 watts, though interesting the package power maxed out at 154w typically hovered about 145 watts for long duration testing.
This lowered the sustained P-core frequency from 4.9 GHz to 4.7 GHz, a 4% drop, and an 8% drop from the peak of 5.1 GHz when running without power limits. This lower power target did drop the peak operating temperature to 87c which is a nice improvement over stock and for those running a lot of all-core workloads this is a more optimal configuration in our opinion, but of course, you are leaving some performance on the table.
Cost vs. Performance
Moving on to cost per frame analysis, we’ll start by comparing just the cost of the CPU. This gives the 7600X an advantage over the 13600K as it provides the same level of gaming performance, while costing $30 less.
But for a more accurate comparison we need to include platform costs, so let’s do that…
Now when factoring in the cost of the memory and motherboard, things look different. The 13600K is clearly better value than the 7600X right now. With both running quality DDR5 memory, the Core i5 processor came in at a cost reduction of 7% per frame, which is a reasonable margin.
Perhaps more crucial is the fact that this made the 13600K the best value part on the market right now, muscling out the 5800X3D and 5600X.
As mentioned in the 13900K review, we’re aware you can get much cheaper DDR4 memory, but it will also reduce performance, which kind of breaks the cost per frame analysis. We’d have to revisit parts like the 13600K and 5800X3D with sweet spot DDR4-3600 CL16 memory to see how they stack up.
What We Learned
As we expected, the Core i5-13600K is a much more practical and better value CPU than not just the 13900K, but also the 13700K, particularly when focusing on gaming.
At the high-end we feel strong arguments can be made for both Intel Raptor Lake and AMD Zen 4, but towards the mainstream segment it looks as though Intel is on a comfortable position over AMD.
Sure, the power consumption isn’t great and without some power limiting the thermals are a bit ugly, but for gaming a 150w package power limit shouldn’t have a big impact, realistically you should see similar results to what’s shown in this review, if not the exact same results.
To be clear, the 7600X and 13600K are extremely evenly matched when it comes to performance, there’s certainly no clear winner here. But the 13600K package is cheaper overall, largely thanks to the availability of already established Intel Z690 motherboards.
Although you can save even more money by opting for DDR4, and perhaps DDR4-3600 CL16 is the way to go there, sacrificing only a small amount of performance compared to what’s been shown here, we are seeing more and more cases of modern games where DDR5 can make a difference.
So for those building a new system, paying the premium might be worth it in the long run, but either way you have the option of going DDR4 or DDR5 with the 13600K, making it a more flexible product.
Then there’s productivity performance, where the 7600X gets steamrolled and this is a big problem for AMD. Even when limited to a package power of 150 watts, which reduces performance by around 7%, the 13600K outscored the 7600X by over 40% in Cinebench R23, and this power limit doesn’t affect lightly threaded performance, just core-heavy workloads.
So no matter which way you slice it the 7600X gets embarrassed by the 13600K in productivity benchmarks, it’s a complete reversal of what we saw just a few generations ago when AMD walloped Intel with their 16-core desktop parts up against 8 and 10 core models.
In our opinion, AMD is in a position where they have to discount the 7600X towards the $200 mark in order to compete in anything outside of gaming. Not only that, but AMD needs quality sub-$200 AM5 motherboards for the 7600X to make sense even if the CPU is discounted.
As enticing as the AM5 platform is, if we were building a new PC today, for either gaming or work, we’d be going with the Core i5-13600K, it’s the obvious choice, just as the 12600K was when we reviewed it back in late 2021.
Shopping Shortcuts:
- Intel Core i5-13600K on Amazon
- Intel Core i7-13700K on Amazon
- Intel Core i9-13900K on Amazon
- AMD Ryzen 7 7700X on Amazon
- AMD Ryzen 9 7900X on Amazon
- AMD Ryzen 9 7950X on Amazon
- AMD Ryzen 7 5800X3D on Amazon
Intel vs AMD — which stocks should you invest in now?
AMD and Intel are one of the largest processor manufacturers in the world, to determine which of these stocks is better to buy, it is worth looking at financial results, market valuation and development prospects.
With a market capitalization of $227 billion as of July 13, Intel (INTC) is still the world’s leading PC CPU vendor and is gearing up for a major expansion in manufacturing capacity. nine0003
Advanced Micro Devices (AMD) is a fast-growing competitor that benefits from Intel manufacturing process delays. AMD’s capitalization is estimated by the market at about $108 billion, and shares are growing against the backdrop of optimistic analysts’ estimates.
Intel versus AMD — whose financial performance is better?
According to Mercury Research, an international research company in the PC component industry, in the first quarter of 2021, Intel held a dominant x86 central processing unit (x86 CPU) market share of 79.3% versus 20.7% for AMD.
However, if we compare the data with last year, it turns out that Intel has lost part of the share of this market, as much as AMD’s share has grown. At the same time, in the first quarter of 2021, AMD lost some market share compared to the previous fourth quarter of 2020.
Comparing the sales growth rates of Intel and AMD, AMD performs the best with an average growth rate of 52.7% for the last six quarterly revenues, while Intel: 6.7%, while the revenues of its last three quarters were down. nine0003
Intel and AMD profit statistics are also opposite — AMD’s gross and operating profit figures for the past year are growing, while Intel’s are declining.
However, Intel outperforms AMD in such an important economic indicator as free cash flow — the company’s working capital available for investing in the development and expansion of production or increasing share buybacks or dividends.
Intel pays dividends and AMD does not, which is also a significant difference for many investors. nine0003
AMD is the favorite of the market
Analysis of Intel and AMD shares clearly shows a higher appreciation of AMD shares by analysts and investors.
Of all ratings of Intel stock by investment firm analysts, 40% give it a Buy equivalent rating, with an average target price of $66.27 vs. $56.87 at Tuesday’s close, representing 16% upside potential. 5% for the next 12 months.
AMD shares receive a higher percentage of analyst buy recommendations, namely 59%, with an average target price of $105.65 indicating a 17% increase from Tuesday’s closing price of $90.3 per share.
Since the beginning of 2021 and over the past 12 months, Intel shares are up 14% and down 2.9%, respectively, while AMD shares are down 1.6% and up 68.4%, respectively.
Intel vs. AMD: The Way Forward
Market experts have praised AMD for the past few years under CEO Lisa Su. AMD is looking to increase its share of both the CPU, GPU and data center market. However, its competitors are such giants as Intel and Nvidia (NVDA), and they have much greater financial opportunities for growth and development.
nine0003
The GPU market is dominated by Nvidia with 81% share, AMD is also quite large — 19% in the fourth quarter (data from Mercury Research).
Competition for data center GPUs is also expected to intensify as Intel starts shipping its own data center GPUs, codenamed Ponte Vecchio, in Q4 or early 2022.
A big argument in favor of Intel for investors can be the implementation of the strategic initiatives of its new gene. director Pat Gelsinger to build two new processor manufacturing plants in Arizona. In addition to increasing its own chip production, Intel plans to become a contract manufacturer for other chip developers, as does the world’s largest contract chip manufacturer, Taiwan Semiconductor Manufacturing (TSM). nine0003
Capitalization of NVDIA and TSM as of July 13, 2021 was $500.9 billion and $559.6 billion, respectively.
A major server manufacturer switched to processors «cooler than AMD and Intel»
Technique
|
Share
nine0003
Gigabyte started selling servers based on ARM processors from Ampere.
They were created by former Intel president Renee James. The new CPUs outperform Intel chips in performance and power consumption by more than twice. With AMD, the difference isn’t that big, but it’s there nonetheless.
Servers without Xeon and Epyc
Gigabyte, a well-known manufacturer of servers, computers and components, has announced a new line of servers with Ampere processors. It was founded in 2017 by the ex-president of Intel, and its main activity is the development of high-performance processors with ARM architecture. nine0003
Gigabyte’s new line includes eight server models with indexes E252-P31, G242-P33, G242-P34, R152-P31, R152-P32, R272-P31, R272-P32 and R272-P33. Ampere-designed chips power each one, with anywhere from 32 to 80 processing cores.
Altra-based Gigabyte flagship server with advanced video subsystem
According to Gigabyte’s official website, the new servers are built on the Ampere Altra line of processors, which debuted in the spring of 2020.
According to the developers, Altra is ahead of its closest competitors by AMD and Intel, respectively, by 1.04 and 2.23 times. nine0003
Silicon «heart» of new servers
All eight servers are designed for rack installation — there are no tower models in this series yet. Six of them are 2U, and only two (R152-P31 and R152-P32) fit in a thin 1U case.
Special Processor
Ampere is backed by former Intel president Renee James . She left it in the summer of 2015 as a result of a reorganization. The true reasons for her dismissal have not been established. Up until her departure, she was effectively number two at Intel after her ex-CEO Brian Krzanich , who left his post in the summer of 2018
The processors used in the new Gigabyte servers are manufactured in near-to-the-state 7nm process technology. Depending on the number of cores, their TDP (thermal power dissipation value) reaches 210 watts.
The processors are based on the ARMv8.
2+ architecture with a number of improvements borrowed from ARM 8.4 and 8.3. The chips can operate at frequencies up to 3 GHz. Each of them has 64 KB of cache in the first level (L1), and the cache in the second level (L2) is 1 MB per core. The cache of the third level reaches 32 MB. nine0003
Ampere processors meet modern requirements in terms of support for DDR4-3200 RAM (four to eight channels are available). The maximum volume is 4 TB. They also have up to 128 PCIe 4.0 lines in assemblies on one CPU, and if the server has space for two processors, then the total number of lines will increase to 192.
What is the advantage of Altra
During the announcement of their chips, representatives of Ampere did not fail to compare Altra with Intel and AMD server CPUs that were current at that time. AMD Epyc 7742 and Intel Xeon Platinum 8280 were chosen as competitors, both of them lose to Altra in terms of performance. nine0003
How to get ₽30 million to bring an artificial intelligence solution to new markets
IT industry support
To show that the Altra is better than the competition, Ampere also made a comparison in terms of energy efficiency.
Here, too, everything is obvious — ARM processors always consume less energy compared to x86.
The comparison here is clearly not in favor of x86 solutions
AMD Epyc 7702 and Intel Xeon Platinum 8276 participated in the test. They turned out to be 1.14 and 2.11 times more “gluttonous” than Altra. nine0003
Note that in March 2021, Ampere showed even more modern Altra Max server processors, also based on ARM. The timing of their appearance in ready-made solutions after more than six months from the moment of the announcement is still unknown.
Learn more about Gigabyte
servers
Gigabyte’s new line of servers with Ampere processors can be divided into three sub-series. The first, which includes models with the letter G in the name, is aimed at increased loads on the video subsystem. It can be installed from two to four Nvidia video accelerators, plus there will be three more slots for other expansion cards. They are powered by two 80 Plus Platinum power supplies of 1.