• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD THREADRIPPER VS INTEL SKYLAKE X

Doobedoo said:
Ouch that Xeon SKylaleSP 8180 cost's almost £9300 each, damn that is one expensive cpu!

That's over £18,500 to reach that score and about £37,000 for the quad which is a little over twice the performance of the Epyc 7601.

Epyc 7601 cost's less than half of what the 8180 cost so you could buy almost 9 7601's to every quad socket 8180's.
Click to expand...

Yeah, Intel reps have their work cut out for them over the next few years.
 
Doobedoo said:
Thinking about it, some of those cinebench scores seem a bit off. I've seen 16c 32t 1950x's get about 3000, so 5500 for 64c 128t doesn't seem right.

Perhaps cinebench is not the correct tool to measure the performance of these server chips.
Click to expand...

It is not right, it must be a single 7601. My best score on my 1950X (running 3.9 ish) is 3335. Considering Epyc clocks lower I think 6k looks about spot on for 7601, just shy of double the oc performance of threadripper.
 
Blackbadger said:
You are correct server 2012 is not a ideal OS to use, and from my own testing 2016 does give better performance for things such as cinebench.
However it is reporting correctly. The screen shot clearly shows 2x 32 core processors and cinebench clearly shows 64c/128t thats how windows and cinebench display things.
The scores could be off and I expect them to be better once it is released, however the hardware is being correctly detected and displayed.

Here is a video that is actually relevant, showing the dual epyc running on server 2016 showing a similar score that is posted. It's a very pro AMD video, showing the price performance ratio against old xeon's.



C'mon be better humbug, less strawman nonsense. You are smarter than that. :)

In a high end build Xeon CPU prices although to us look astronomically expensive (which they are), they are not as significant as the whole platform cost. After all the rest of the parts are in place even at £8k per cpu, that is usually only a fraction of the actual system cost.
Click to expand...

£20,000 for 4 CPU system with the same performance as the Dual £8,000 CPU isn't tho ;)
 
people look at systems that are insane priced to build and look at the cost but not what they can earn ;)

just stepped in the new intel chips faster than all amd people trying justify the price or comparing to a slower amd chip.just dont.people who need them buy them they are the quickest they earn money for what they needed for.not playing minecraft or csgo.
 
~>Dg<~ said:
people look at systems that are insane priced to build and look at the cost but not what they can earn ;)

just stepped in the new intel chips faster than all amd people trying justify the price or comparing to a slower amd chip.just dont.people who need them buy them they are the quickest they earn money for what they needed for.not playing minecraft or csgo.
Click to expand...

People look at systems required to do a task. Most people with an IQ of over 100 will buy the most economical configuration for the job. Intel and Nvidia people will say buy the non AMD chip.just don't.something.something minecraft.csgo.
 
~>Dg<~ said:
people look at systems that are insane priced to build and look at the cost but not what they can earn ;)

just stepped in the new intel chips faster than all amd people trying justify the price or comparing to a slower amd chip.just dont.people who need them buy them they are the quickest they earn money for what they needed for.not playing minecraft or csgo.
Click to expand...

Your posts are not conducive to intelligible debate given most of it reads as incoherent rambling but from what i gather out of what you are trying top say is price is no object when performance is higher, but what about when performance is the same? When the highest end dual EPYC system is the same performance as the quad Intel system, with the same performance Intel costing £12,000 more you would not still suggest that system, given it cost you £12,000 and earns you the money at the same rate, given the power efficiency of AMD Ryzen architecture vs Intel's the Intel system also has higher running costs so you would actually earn more with the AMD system.

Here it is, Blackbadgers own video. The £8,000 Dual CPU EPYC 7601 scores almost exactly the same as the £19,000 Quad CPU Intel 8867v3.

 
It's comparing the EPYC vs 2 year and 6 year old Xeons. I posted it as you claimed I was making up things about how windows displays cpu's. And the benchmark was not running dual epyc's as it was server 2012.
You can see from the video it is running server 2016 running dual EPYC 7601's thats 64 cores 128 threads. And scores 5546

I already posted the performance of the current Xeon chips, yet you choose to ignore them.

It's obvious that the creator of the video is failing to make a sensible comparison by comparing against old tech instead of the current lineup.

I could not find a Skylake Xeon vs EPYC comparison vid. I guess it was a troll video and I fed the trolls by mistake... My bad and sorry for any confusion.
 
Last edited:
Blackbadger said:
It's comparing the EPYC vs 2 year and 6 year old Xeons. I posted it as you claimed I was making up things about how windows displays cpu's.
You can see from the video it is running server 2016 running dual EPYC 7601's thats 64 cores 128 threads

I already posted the performance of the current Xeon chips, yet you choose to pretend they don't exist.

The current Xeon lineup as you know is significantly faster even at a lower core count than EPYC. It's obvious that the creator of the video is failing to make a sensible comparison.
I could not find a Skylake Xeon vs EPYC comparison vid. I guess it was a troll video and I fed the trolls by mistake... My bad.

To clear things up to make it easier to understand. Take the base system without cpus at conservative 10k. Ram, HDD's software licenses etc.

Dual EPYC 7601 2x32core
Approx total system cost $18000
score 5546

Dual SkylakeSP 8168 2x24 core
Approx total system cost $22000
score 7188

That's a 29% speed increase for 22% more cost, and as you can add more cpu's with the intel system the price/performance ratio increases significantly.

Quad SkylakeSP 8168 4x24 core

If people who need this power for speed of getting tasks done. 25% improvement can save 1 week a month, 3 months a year of extra productivity. A hundred in extra power spent annually, or even £10,000 or £20,000 more in hardware makes sense.
Click to expand...

That looks to me like AMD might want to cut about $1,500 off their 7601 to bring them down to about $2000 to $2500. they may have to because with Intel offering the same or better price to performance AMD aren't going to sell any of them.

Its all good, maybe Intel will be forced to cut about $2000 of their chips in return. and so on..... bring those prices down to rock bottom. the pricing grown by insane amounts over the past decade, competition is wonderful. set new low pricing standards :)

Intel's HEDT chips are still overpriced tho, Intel are offering +10% for not far from twice the cost of Threadripper, its what $800 more? well over priced.
 
Last edited:
:) Agree the Intel HEDT are too expensive. I think the whole xeon line up is.

The EPYC systems do offer things like a far greater amount of PCI lanes, for those situations where they are needed it can be worth a lot.
Even if total speed per core is greater on intel, there are situations such as virt environments where more cores = more vm's and if that's what's needed the epyc range offer a compelling choice.

Even though I posted those scores for epyc, I as others do question them, it could be pre-release hardware or poor windows scheduling issues or similar, however similar scores are replicated by many sources.
I am sure there are workloads where it does excel and it's performance will beat a comparitive more expensive intel system. Having the ability to throw more chips in a intel system however will be a alternative, if you can pallet the cost.

I think cinebench in it's current form as suggested cannot cut the mustard anymore for tests such as these and starts to have issues with so many cores/speed. Maybe something with a bit more teeth like Corona or something else visual which we can compare to consumer parts maybe better.

Both platforms are immature so all results from both sides are likely to change over time.
 
When huge companies like Baidu are talking about Epyc,it shows you the cost savings must be very real:

https://www.youtube.com/watch?v=lIXTm63e974


Its to be expected when those high end Intel server CPUs have huge dies with the yields to match and AMD is using a much smaller chip. The basic Ryzen building block is an SOC of around 189MM2. Even the Intel 10C server CPU which is not an SOC(so needs a dedicated chipset) is something like 320MM2 in size plus a largish chipset chip.
 
CAT-THE-FIFTH said:
When huge companies like Baidu are talking about Epyc,it shows you the cost savings must be very real:

https://www.youtube.com/watch?v=lIXTm63e974


Its to be expected when those high end Intel server CPUs have huge dies with the yields to match and AMD is using a much smaller chip. The basic Ryzen building block is an SOC of around 189MM2. Even the Intel 10C server CPU which is not an SOC(so needs a dedicated chipset) is something like 320MM2 in size plus a largish chipset chip.
Click to expand...

I think AMD should make Intel practically give them away, its pretty clear AMD is having a huge effect on Intel both on Xeon's and HEDT, they have been forced to release 18 Core HEDT CPU's at a price previously occupied by a 10 core while that is now in the relative cheap bin well under $1000.

Some of Intel's previous Xeon's cost in excess of $10,000, the price of those have also halved.

Intel did this to try and stop AMD gaining any traction, thats already failed but i think AMD with much much lower running costs and any revenue from servers and HEDT is revenue they have not had in a decade should kick Intel where it hurts, AMD should pull down the cost of their chips even more to force Intel to sell what has always been a cash cow for them at unsustainably low prices.
 
Blackbadger said:
:) Agree the Intel HEDT are too expensive. I think the whole xeon line up is.

The EPYC systems do offer things like a far greater amount of PCI lanes, for those situations where they are needed it can be worth a lot.
Even if total speed per core is greater on intel, there are situations such as virt environments where more cores = more vm's and if that's what's needed the epyc range offer a compelling choice.

Even though I posted those scores for epyc, I as others do question them, it could be pre-release hardware or poor windows scheduling issues or similar, however similar scores are replicated by many sources.
I am sure there are workloads where it does excel and it's performance will beat a comparitive more expensive intel system. Having the ability to throw more chips in a intel system however will be a alternative, if you can pallet the cost.

I think cinebench in it's current form as suggested cannot cut the mustard anymore for tests such as these and starts to have issues with so many cores/speed. Maybe something with a bit more teeth like Corona or something else visual which we can compare to consumer parts maybe better.

Both platforms are immature so all results from both sides are likely to change over time.
Click to expand...

Yeah, something is off with those EPYC scores, 4 times the number of cores (64) running at more than half the speed of the 16 core Threadripper and it only scores 40% higher, that just doesn't add up which is why i was sceptical about how many CPU's that actually was.

Last night it was very late but i have now dug down into the numbers.

Here is what i did, i divided the score by the number of cores to give me a per core score, i did this for both the EPYC 7601 and the Threadripper 1950X, the 1950X is clocked 55% higher so i added that to the 7601 per core score.

7601 x2 @2.2Ghz: 5546 / 64 = 86.7
1950X @ 3.4Ghz: 3151 / 16 = 197

Difference between 2.2Ghz and 3.4Ghz = +55%

86.7 + 55% = 134.4 < now, they are the same chip and with the 55% clock speed deficit added on to the per core score it should score the same 197 points, it doesn't, its way down.

So there is something very wrong with those EPYC results: A little more digging, 2.2Ghz is 65% of 3.4Ghz so 65% of 197 = 128. 128 points is what should be EPYC's per core performance, so multiply 128 by 64 cores and you get 8,192 <that is around what the 7601 should score in Cinebench. R15.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom