Info [Toms, Anand] AMD EPYC Benchmarks

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

BigDaveX

Senior member
Jun 12, 2014
440
216
116
Kudos to AMD for besting Intel in their own game. I just hope that corporate customers actually switch to AMD instead of waiting for Intel to counterpunch.
That's kind of the one fly in the ointment with this launch - while the enthusiast community are cheering the utter humiliation that Intel has just experienced, odds are most corporate buyers are still just blithely going to the Dell website and buying yet another Broadwell-based rack server, simply because they want something that has a reliable track record and will slot easily into their current set-up.
 
Reactions: Ranulf

Abwx

Lifer
Apr 2, 2011
11,057
3,715
136
And the FUD has started,

Is AMD Rome burning?
https://fudzilla.com/news/pc-hardware/49183-is-amd-rome-burning


I will post the entire article for those that dont want to click to read it,



Hot voltage..?.
What is this new "sense of physics"..?.

Whatever the voltage laws of physics tell me that 1.4V loading a 0.01R resistance will produce 200W, that s the amount of Rome at full tilt, on the other hand i can have 1.2V with the 56C core Xeon as 0.003R load that will produce 400W heat, those who did this FUD are not even informed about Ohm s law, hence the "hot" for a parameter that has no significance alone if not weighted by the current.

Edit : some slide



https://www.computerbase.de/2019-08/amd-epyc-7xx2-rome-zen-2-specs-release/
 
Last edited:
Reactions: lightmanek

Asterox

Golden Member
May 15, 2012
1,026
1,775
136
Whew. That is not pretty for Intel. And the hits just keep coming every year. Intel won't be able to defend themselves until 2021 (Sapphire Rapids), and that thing has to face off against Zen 4. I don't think Cooper Lake is going to help them much.

It is ok, a mighty rescuer or ace up in Intel sleeve arrives soon.

 
Reactions: lightmanek

scannall

Golden Member
Jan 1, 2012
1,946
1,638
136
That's kind of the one fly in the ointment with this launch - while the enthusiast community are cheering the utter humiliation that Intel has just experienced, odds are most corporate buyers are still just blithely going to the Dell website and buying yet another Broadwell-based rack server, simply because they want something that has a reliable track record and will slot easily into their current set-up.
I would imagine more than a few procurement officers will be taking nice paid for vacations, or perhaps a shiny new car in the near future.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,640
14,630
136
It is ok, a mighty rescuer or ace up in Intel sleeve arrives soon.

$50,000 for the new Intel 9200 series 400 watt processor ? That still has all the security flaws ? And they have no idea how they are going to cool it ?

As he says in that video, "Intels new Bulldozer moment" and things don't look good for Intel.
 
Reactions: Drazick

naukkis

Senior member
Jun 5, 2002
722
610
136
Dekstop Ryzen needs 1,4V to clock 4+ghz. Rome instead uses about 1V to reach those 3+ghz. If 64c Rome would use 1,4v voltages if would consume something like 800w of power.
 

zir_blazer

Golden Member
Jun 6, 2013
1,179
441
136
You initially stated the 7252 was a 16c and I assumed that was correct. I'm not sure what seems to be confusing you but the core counts and cache make sense at each level to me. (I had it all laid out and the post got eaten)
I edited because I was talking about the wrong model. Let me put the list:

https://www.servethehome.com/wp-con...C-7002-SKU-List-and-Value-Comparison-Full.jpg


I was talking about the 7282 (16C, 64 MiB Cache L3) and the 7302 (16C, 128 MiB Cache L3). I think I got them confused with the 7252 caused I was first looking at the Cores number then jumped to the Threads one.
 

beginner99

Diamond Member
Jun 2, 2009
5,211
1,582
136
That's impossible, we have our French ex-Intel that states:

With all the news, across the internet, that praise Epyc ... comments like those above, kinda stick out.

He has a point though about the output. In the athlon days AMD was simply supply limited. And given they have issue with supply here of 3900x I do question how big a dent they can make. Unless 3900x is limited due to clock speed (binning) are because high demand of Rome.
 

Asterox

Golden Member
May 15, 2012
1,026
1,775
136

Kenmitch

Diamond Member
Oct 10, 1999
8,505
2,249
136
He has a point though about the output. In the athlon days AMD was simply supply limited. And given they have issue with supply here of 3900x I do question how big a dent they can make. Unless 3900x is limited due to clock speed (binning) are because high demand of Rome.

Hard to figure out what's up with the 3900x. Could be nothing more than a bad sales prediction from some bean counters.

Image is from ShopBlt's 3900x pre order page as of this this typed message.



https://www.shopblt.com/cgi-bin/sho...6QP158P.shtml&order_id=214903134#Availability

Will be interesting to compare the above image in a couple of days to see if the past due and projected 8-10 delivery make a dent in the pre orders.

Are there any other vendors that supply similar data to compare?
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,640
14,630
136

Hitman928

Diamond Member
Apr 15, 2012
5,392
8,280
136
So he says that they could be as cheap as $25k. But what idiot would pay $25k for a CPU with 56 cores and that is slower than the 7k EPYC with 64 cores ? And the data center has to provide power and cooling for 400 watts instead of 225 watts ? Again the stupidity of Intel and the people that buy these hot power hungry monsters.

The 56 core intel CPUs should perform significantly faster if you're running a lot of AVX512 code (which is very, very niche right now). There are a couple of other very niche applications where the intel will win. Usually those very very niche markets are used to paying absurd prices for products catered to them so the chip will find a home or two, but really it's more of a PR move than anything.
 
Reactions: lightmanek

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,640
14,630
136
The 56 core intel CPUs should perform significantly faster if you're running a lot of AVX512 code (which is very, very niche right now). There are a couple of other very niche applications where the intel will win. Usually those very very niche markets are used to paying absurd prices for products catered to them so the chip will find a home or two, but really it's more of a PR move than anything.
I thought the benchmarks said that Rome does surprising good in AVX512 for a CPU without the special hardware ?
 

beginner99

Diamond Member
Jun 2, 2009
5,211
1,582
136

Topweasel

Diamond Member
Oct 19, 2000
5,436
1,655
136
I thought the benchmarks said that Rome does surprising good in AVX512 for a CPU without the special hardware ?
They do but only because of the massive lead in core count and decent clock advantage. A 56c Intel should be well more than competent against a 64c Rome in avx-512.
 

Hitman928

Diamond Member
Apr 15, 2012
5,392
8,280
136
I thought the benchmarks said that Rome does surprising good in AVX512 for a CPU without the special hardware ?

AMD is running unoptimized AVX2 code vs tuned AVX512 code on the intel CPUs and does do surprisingly well, but that is also against 28 core intel CPUs. Compared to 56 core intel CPUs, it won't be able to keep up on heavy AVX512 code, though perf/w will look better for Epyc than just pure perf.
 
Reactions: lightmanek

Abwx

Lifer
Apr 2, 2011
11,057
3,715
136
They do but only because of the massive lead in core count and decent clock advantage. A 56c Intel should be well more than competent against a 64c Rome in avx-512.

It wont, because Intel s 512bit width FPUs are so innefficient that AVX512 frequency offset become so big that it cant compensate for AMD s better throughput in AVX2, that s basically what Anandtech test displayed.
It s no more Epyc 1 with half the FPU width, wich was a technical choice as pointed by AMD, according to their engineer 256bit wide vectors and FPUs would had slaughtered FP perf/watt even when using only AVX1 and legacy SSE2-4.2.
 
Reactions: moinmoin

DrMrLordX

Lifer
Apr 27, 2000
21,709
10,983
136
And the FUD has started,

The stupid part here is that desktop parts don't require 1.4v for anything except maybe single-core boost scenarios where current/power draw is so low that it can blow the doors of any voltage "recommendation" without damaging the chip or doing anything "hot". Rome doesn't even have boost frequencies that high. Did any of these idiots actually think, "hey, instead of FUDing, why don't we just do a voltage histogram on a Rome review sample and see what it really uses?".
 

Kenmitch

Diamond Member
Oct 10, 1999
8,505
2,249
136
The stupid part here is that desktop parts don't require 1.4v for anything except maybe single-core boost scenarios where current/power draw is so low that it can blow the doors of any voltage "recommendation" without damaging the chip or doing anything "hot". Rome doesn't even have boost frequencies that high. Did any of these idiots actually think, "hey, instead of FUDing, why don't we just do a voltage histogram on a Rome review sample and see what it really uses?".

If these FUDers researched before hand then I doubt they'd have and FUD fuel left. Makes sense that the professional FUDers just go with the FUD 1st policy.
 

moinmoin

Diamond Member
Jun 1, 2017
4,975
7,736
136
If these FUDers researched before hand then I doubt they'd have and FUD fuel left. Makes sense that the professional FUDers just go with the FUD 1st policy.
They call themselves FUDzilla after all.
 

itsmydamnation

Platinum Member
Feb 6, 2011
2,804
3,267
136
stolen line for line form realworld tech:

https://www.sap.com/dmc/exp/2018-benchmark-directory/#/sd?id=5e123370-56ea-419a-bf99-af36d4622391

HPE ProLiant DL385 Gen10, 2xAMD EPYC 7702, 2.0 GHz, 2S/128C/256T, 1024 GB, WS2019, MS SQL Server 2017 = 45,600 Users

For comparison:

HPE ProLiant DL385 Gen10, 2xAMD EPYC 7601, 2.2 GHz, 2S/64C/128T, 1024 GB, WS2012, MS SQL Server 2012 = 23,820 Users

HPE Proliant DL380 Gen10, 2xIntel Xeon Pt 8280L, 2.7 GHz, 2S/56C/112T, 768 GB, WS2016DE, MS SQL Server 2012 = 29,231 Users


Other OS/Db:
Dell EMC PowerEdge R740xd, 2xIntel Xeon Pt 8280, 2.7 GHz, 2S/56C/112T, 768 GB, SLES 15, SAP ASE 16 = 35,505 Users


Very expensive:

FUJITSU Server PRIMERGY RX4770 M5, 4xXeon Pt 8280, 2.7 GHz, 4S/112C/224T, 11536 GB (??? - mistake?) , WS2016DE, MS SQL Server 2012 = 57,500 Users


IBM Power System E980, 16xPOWER9, 3.9 GHz, 16S/192Cs/1536T, 8192 GB, AIX 7.2, Db2 10.5 = 205,000 Users

Fujitsu SPARC M10-4S, 32xSPARC64 X+, 3.7 GHz, 32S/512C/1024T, 8192 GB, Solaris 11, Oracle 12c = 153,050 Users
 

krumme

Diamond Member
Oct 9, 2009
5,955
1,595
136
Yep. For some reason those numbers is not discussed much.

DB performance gets a huge uplift. What lies beneath is not only better latency but much more predictable latency and thereby performance. This cpu is a beast at reasonably pricing and it's more importantly safer to invest in. It's a robust solution to 95% server needs. That's quite a difference to Epyc. It's more versatile.

What i not discussed then is also the impact for sparc and IBM. If amd continues this progress just remotely a lot of players besides Intel is pressured heavily.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |