Core Combat: Intel Xeon W-3175X vs. AMD Threadripper 2990WX

[redgarl]

Just because you are not going to buy a 28 or 32 core processor for playing games does not in and of itself mean that gaming performance is not important.
If I were going to dump $8000 to $12000 on a computer for rendering for example, I would not want to have to turn around and drop another $3000 to $4000 on a gaming computer.

-------------

You need to understand that the amount of content to be "rendered" that an individual can create.. would easily be handled by a Ryzen 7 or Core i7.

Don't just throw the word "rendering" around senselessly.

Do you know how much time would it take to create content that probably would take 1 hour to render on a Ryzen 7 2700X? At the very very minimum - 1 full day!

So no individual would throw away $8000 to just save maybe 30 mins on a project thats anyway going to take days to complete!!!

The gaming benchmarks are extremely stupid and there is no sensible justification for it. (Benchmarks for the sake of benchmarks implies you are stupid - which is whats happening here.)

Glad to see that I am not the only one with a critical sense.

 

[PaulAlcorn]

Paul, for real, how many of these benchmarks can you run on Linux in order to mitigate the Windows scheduler issue, please?

While I am a Windows user and all of my & my team's current use cases all fall on Windows software except for two which can run on Linux, but don't (yet?) there are many people who are not in that situation.

Also, seriously, gaming benchmarks??

Running Linux isn't an issue, but spinning up benchmarks for multiple chips is a time consuming endeavor. Windows is the predominant OS for workstations, whereas we use Linux for server testing.

Yes, we run games. We do not deviate from our standard test regimen on any chip. This is because some might think we are hiding poor performance, etc, for the device under test. However, we will add tests to reflect specialized use cases, such as the workstation tests above.

Nobody buying those CONSIDER gaming in their purchase... so yes, it is misleading.

<Quote>Do we really need a gaming performance segment for such a product review? Nobody is going to play on a 28-32 core CPU... especially since AMD mentioned WX were NOT aimed at gamers.

As far as features go... how Intel is winning when the only advantage it has is six-channel memory over AMD socket? 64 PCIe lanes, 64mb of L3 cache, 2933 MHz memory, 32 cores, lower temps and better power...</Quote>

Please take this review with a bucket of salt! )

ROFL, my comment was deleted. because it was asking the right questions.

It is not really the products that pissed me off, it is the biased process that are totally arbitrary.

In the end, you are comparing a product that was not design for some aspects, by using benchmarks not proving anything because the OS is gimping the result on one of the platform, and that cost 3 times less than the other.

I mean sure, if you want, I am pissed about the fact that Intel tied when it should have lost... in a biased comparison with questionable choices of tests... with questionable review process with an author who is not able to take the heat by confronting everyone not agreeing with his methodology and his conclusion.

Basically, as an electrical engineer having to review process in projects and products, if someone was doing something like that in my department, I can assure you that a chat would be in order.

Sure, the Windows scheduler exacerbates the 2990WX's architectural eccentricities, but that's what people who buy this will encounter in real use.

 

[PaulAlcorn]

<Quote>Do we really need a gaming performance segment for such a product review? Nobody is going to play on a 28-32 core CPU... especially since AMD mentioned WX were NOT aimed at gamers.

As far as features go... how Intel is winning when the only advantage it has is six-channel memory over AMD socket? 64 PCIe lanes, 64mb of L3 cache, 2933 MHz memory, 32 cores, lower temps and better power...</Quote>

Please take this review with a bucket of salt! )

The 64MB of cache is split across four dies, so it is not centralized. Cache is designed to feed the processor with data. Does the 2990WX win in cache bandwidth and latency? No.

64 PCIe lanes is a win, sure, but again, these are spread among the dies, so data requests to remote PCIe lanes incur higher latency, which has an impact on performance.

2933MHz isn't as impressive if you can't beat the competition on latency and bandwidth, which is impossible if two of the four die aren't connected to memory controllers. The benchmarks tell the story there.

Not all cores are created equal. 32 AMD cores don't beat Intel's 28 in the majority of workloads.

Lower temps is a win.

Lower power is a win, if the performance-to-power ratio supports that. If you draw twice the power, yet produce three times the performance, the extra power is worth it.

Features don't equate to a win if they don't deliver real-world performance advantages. As we said, Threadripper wins in some areas, but not in others.

 

[cryoburner]

I agree that weighting is the best approach, but unfortunately deriving a weighting metric would involve opinion. How much weight do we assign to each category?
...
We eschew weighting in favor of presenting raw data, which in turn allows the reader to select areas of interest for them, and then make decisions accordingly. Simply put, if a certain category isn't important to you, remove that checkmark from your calculation.

Like Emerald, I'm also not much a fan of the way these "versus" articles are done. And of course, weighting wouldn't improve anything, as different categories bear different relevance depending on one's needs. As for avoiding opinion though, even in the current format, opinion seems to be figured into declaring winners of categories, and different parts within a category may bear different weight to different users, so looking at the table of check marks at the end and a "winner" based on them doesn't really tell us much. For example, even if one piece of hardware performs best overall in a particular category of software, it might not necessary perform better in the piece of software that would be most important to a particular user.

Honestly, I kind of think declaring winners and losers in an article like this and adding them up in a chart at the end is kind of silly, and seems to serve little usefulness, since it's oversimplifying the results. It also ignores the fact that there are more than just two products available on the market in a given segment. Does it really make much sense to compare products head to head at entirely different price points while ignoring other available products in more similar price segments? Opinion is already playing a role deciding which products to line up against one another. I'd rather see an article comparing the strengths and weaknesses of various products within a segment, such as all current HEDT processors, for example. Though that's probably not too different from what reviews already do.

 

[anant.cool.007]

https://www.pugetsystems.com/labs/hpc/AMD-Threadripper-and-1-4-NVIDIA-2080Ti-and-2070-for-NAMD-Molecular-Dynamics-1321/

^ THAT is a real world use case for TR's and Xeons, ofcourse this one is missing from the review.

Also, for features - TR2990 has greater Cache than Xeon. It has more PCIe lanes than the Xeon and it runs on faster memory as well.

It has higher latency but that doesn't change the fact that AMD has more lanes and cache

Also, all cores are not created equal, seriously I can't argue with you here because you think gaming is a real world workload for these CPU's and then softwares like NAMD are missing from your benchmarks..

As RedGarl said, this is misleading. btw you banned her also, for disagreeing with you? SAD.

 

[jgraham11]

Comparing a $2000 CPU to a $3000 CPU? This is the biggest load of shit I think Tom's has done in a while (except for the Nvidia just buy it "Why would you want to live in a world that wasn't raytraced")

A more appropriate comparison would actually be with the AMD 2950x as it is more closer core count, better suited to non-rendering type work loads and typically outperforms the CPU the 2990WX in again non-rendering work loads. BTW only costs 1/3 the cost of the Intel chip. Way to go Tom's: losing credibility every article published like this.

 

[anant.cool.007]

jgraham, did you read the one where they paired a $60 Athlon with a ~$800 1080 and then reviewed the CPU, that one was epic! xD

 

[PaulAlcorn]

https://www.pugetsystems.com/labs/hpc/AMD-Threadripper-and-1-4-NVIDIA-2080Ti-and-2070-for-NAMD-Molecular-Dynamics-1321/

^ THAT is a real world use case for TR's and Xeons, ofcourse this one is missing from the review.

Also, for features - TR2990 has greater Cache than Xeon. It has more PCIe lanes than the Xeon and it runs on faster memory as well.

It has higher latency but that doesn't change the fact that AMD has more lanes and cache

Also, all cores are not created equal, seriously I can't argue with you here because you think gaming is a real world workload for these CPU's and then softwares like NAMD are missing from your benchmarks..

As RedGarl said, this is misleading. btw you banned her also, for disagreeing with you? SAD.

That article focuses solely on NAMD. We also tested NAMD with SPECworkstation 3 in our review, and it is presented in this article as well.

However, we also tested against competing devices, whereas that article does not (still interesting info, though).

We also include testing for financial forecasting models, seismic processing, three-dimensional structural computations, molecular dynamics, and discrete energy minimization.

 

[shabbo]

Another BS faceoff by Tom's. Intel cost Over $1K more, which can easily be put towards a Radeon Pro. Also Intel will need a more expensive cooler. On top of all that the AMD platform can be easily updated to 7nm. Intel is a big day loser and doesn't deserve to even be in this league.

 

[Rob1C]

Want the uneven test?, OK let's see it!

Intel Xeon W-3175X vs. AMD 2x 7601

 

[jgraham11]

anant.cool.007: Yes I did, against the Pentium G5400, completely unrealistic for many reasons.
Did anyone notice in this article that they claim Intel won at Cinebench meanwhile they only compared the single threaded results.

Also that Toms buried decompression in their test suite , but certainly made sure to showcase compression. This again is quite a misrepresentation as most users will decompress way more than they compress.

Some simple questions: why is only the Intel chip shown overclocked, but the AMD chip in only auto-overclocked? Why are they calling the Intel chip a tie for overclocking when you need "to spend copious amounts of money on premium components to unlock the best the silicon has to offer." (Quote from the article). My guess is something similar to what they used at the last Computex back in June, remember it was Toms who first said they "forgot" to mention it was overclocked and chilled by a multi-horsepower chiller. (don't we all use one of those?? lol)
Why is the win given for gaming when the 99th percentile (which is the real indicator of gaming smoothness) within 2 FPS? Did anyone look at the Gaming price efficiency, the Intel CPU is quite literally off the charts! lol
Why is Intel given the features win when they both support ECC memory, the AMD platform supports faster memory and they AMD platform has 12 more PCIE lanes available. Their only justification seems to be because the Intel platform has server management capabilities (vPro, etc).

Honest hardware review sites are becoming harder and harder to find, I just want this behaviour of attempting to misguide buyers to stop.

 

[shmoochie]

How can you not see that by not assigning weights to each category, you ARE assigning a weight? You just assigned a 1 which implies that they are all equal. You are being disingenuous by saying you just present facts when you clearly have a weighted system at the bottom of the review. That table is absolutely useless, and your defense for including it/leaving it is bonkers at best. Stop leaving it up to the reader to sift through your mess. Just remove it.

 

[Olle P]

If I were going to dump $8000 to $12000 on a computer for rendering for example, I would not want to have to turn around and drop another $3000 to $4000 on a gaming computer.

If you spend $10k on a computer for rendering you will most likely use it professionally. Otherwise you're either extremely devoted to rendering or just plain stupid.

With that in mind, there should be no problems for you to get another $1,000 for a gaming computer.

 

[Blackbird77]

Incorrect comparison if you ask me. Should have been a pro vs. pro comparison. Threadrippers are awesome of course, but DON'T fall into the PRO category from AMD. They aren't the EPYC counterparts to INTEL Xeons. For the price of ONE Xeon, you can almost buy TWO Threadrippers, so if you could dump TWO treadrippers into ONE motherboard and pay the price of one Xeon, AMD would make Intel bite the dust in EVERY SINGLE benchmark. Kind of misleading comparison. This is like comparing apples to potatoes. Looks like an INTEL AD, saqying that Intel is usually better, but at the end saying, OK AMD is the underdog and almost is the same. For almost HALF the price and almost the same? You are taking Ferrari performance for Mustang money. Almost the same? Gotta be kidding me.

 

[chalabam]

Yes, we run games. We do not deviate from our standard test regimen on any chip. This is because some might think we are hiding poor performance, etc, for the device under test. However, we will add tests to reflect specialized use cases, such as the workstation tests above.

Thanks for testing games.

It is important to actually having the benchmarks rather than just an opinion.

To say that those processors aren't good choices for gaming is not the same as actually having the experimental data to back it up.

You should also had tested with cheap memory, to make better scatter plots of price vs performance. That memory should cost a lot, (I cannot find how much memory you used), because to feed many cores is also necessary to give enough memory to each core. So the difference in cost is important.

 

[chalabam]

The gaming benchmarks are extremely stupid and there is no sensible justification for it. (Benchmarks for the sake of benchmarks implies you are stupid - which is whats happening here.)

Your argument is stupid, because having the data is not the same as speculating.

A lot of people has interest on the gaming side, and they need to know the numerical facts rather than just opinions.

 

[Flemishdragon]

Why is the overclock number only next to Xeon cpu's and not the Threadripper? Is the amd max overclock on water only 4.2? That's to low for a workstation.

 

[Olle P]

Why is the overclock number only next to Xeon cpu's and not the Threadripper? Is the amd max overclock on water only 4.2? That's to low for a workstation.

If you read the "Overclocking" section you get the answer:
The TR use PBO that does an excellent job of maximizing performance based on the number of active threads. With manual overclocking you either have to settle for a lower all core speed or disable some/most cores to go faster on the remaining one(s).

 

[littleleo]

Paul, for real, how many of these benchmarks can you run on Linux in order to mitigate the Windows scheduler issue, please?

While I am a Windows user and all of my & my team's current use cases all fall on Windows software except for two which can run on Linux, but don't (yet?) there are many people who are not in that situation.

Also, seriously, gaming benchmarks??

Hey he is a writer it is what he does. Just read it.