Ryzen and Threadripper

Message boards : Number crunching : Ryzen and Threadripper
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 24 · 25 · 26 · 27 · 28 · 29 · 30 . . . 38 · Next

AuthorMessage
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 10043
Credit: 967,756,956
RAC: 1,536,903
United States
Message 2005875 - Posted: 5 Aug 2019, 16:30:10 UTC - in response to Message 2005855.  

Or just a another couple of months for ASUS to work the bugs out of the BIOS code hopefully. But ASUS is absolutely the last vendor to offer updated BIOS with the latest AGESA code.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 2005875 · Report as offensive     Reply Quote
Profile Jord
Volunteer tester
Avatar

Send message
Joined: 9 Jun 99
Posts: 14889
Credit: 4,143,633
RAC: 1,253
Netherlands
Message 2005896 - Posted: 5 Aug 2019, 20:02:13 UTC

Memory considerations for Ryzen 3000 systems as tested by Linus Media Group: https://www.youtube.com/watch?v=iHJ16hD4ysk
ID: 2005896 · Report as offensive     Reply Quote
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 10043
Credit: 967,756,956
RAC: 1,536,903
United States
Message 2005910 - Posted: 5 Aug 2019, 21:07:53 UTC - in response to Message 2005896.  

Memory considerations for Ryzen 3000 systems as tested by Linus Media Group: https://www.youtube.com/watch?v=iHJ16hD4ysk

Good video. Shows that AMD has been able to engineer the maximum performance even out of base Jedec memory speeds. So you don't have to spend the premium bucks on the fastest memory/lowest latency kits that you had to do for best performance on Ryzen Gen. 1 and Ryzen Gen. 2.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 2005910 · Report as offensive     Reply Quote
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 10043
Credit: 967,756,956
RAC: 1,536,903
United States
Message 2005924 - Posted: 5 Aug 2019, 23:02:20 UTC

And the leaked benchmarks for the Epyc processors using the new chiplet design are out. How about 64 cores and 128 threads and 256MB of cache?

https://www.tomshardware.com/news/amd-epyc-7742-vs-intel-xeon-benchmarks,40089.html
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 2005924 · Report as offensive     Reply Quote
Ian&Steve C.
Avatar

Send message
Joined: 28 Sep 99
Posts: 1898
Credit: 837,601,725
RAC: 2,758,356
United States
Message 2005933 - Posted: 6 Aug 2019, 0:42:33 UTC - in response to Message 2005924.  

And the leaked benchmarks for the Epyc processors using the new chiplet design are out. How about 64 cores and 128 threads and 256MB of cache?

https://www.tomshardware.com/news/amd-epyc-7742-vs-intel-xeon-benchmarks,40089.html


Sounds like a CPU that SETI wouldnt be able to load 100% because of the CPU WU restrictions :P
Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

ID: 2005933 · Report as offensive     Reply Quote
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 10043
Credit: 967,756,956
RAC: 1,536,903
United States
Message 2005936 - Posted: 6 Aug 2019, 1:08:24 UTC

Nope you would still be restricted to only 100 cpu tasks running on it at any time. Might be enough reduction to not have the cpu exhibit overcommitment and can use the other 28 threads to feed 20 gpus with 8 threads left over to run the Desktop.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 2005936 · Report as offensive     Reply Quote
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 3699
Credit: 221,385,146
RAC: 390,888
United States
Message 2005956 - Posted: 6 Aug 2019, 4:43:55 UTC - in response to Message 2005936.  

Nope you would still be restricted to only 100 cpu tasks running on it at any time. Might be enough reduction to not have the cpu exhibit overcommitment and can use the other 28 threads to feed 20 gpus with 8 threads left over to run the Desktop.


Is the 100 cpu task a BOINC limit or a Seti limit?
If it is a Seti limit you could be running another cpu-based project with no loss of production.

Tom
A proud member of the OFA (Old Farts Association)
"Over the hill? WHAT Hill? I don't REMEMBER any hill...." (from a bumper sticker I bought at a truck stop).
"If its Tourist Season why can't we shoot them?" (another bumper sticker)
ID: 2005956 · Report as offensive     Reply Quote
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 10043
Credit: 967,756,956
RAC: 1,536,903
United States
Message 2005958 - Posted: 6 Aug 2019, 4:50:03 UTC - in response to Message 2005956.  

I believe it is only a Seti limit. But I have no experience with any other projects cpu applications. I wonder if the QC Chemistry app at GPUGrid has any limit.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 2005958 · Report as offensive     Reply Quote
jsm

Send message
Joined: 1 Oct 16
Posts: 83
Credit: 38,941,362
RAC: 80,099
Isle of Man
Message 2005961 - Posted: 6 Aug 2019, 6:17:13 UTC - in response to Message 2005958.  

Is it perhaps time for those in charge of SETI development to review all such limits in the light of improved or enhanced capability? Surely these will be constants easily modified. Of particular concern is the limit on downloading background tasks on my 2990 because every time SETI goes off line I run out of work very quickly.
Who is actually responsible for making this sort of decision?
jsm
ID: 2005961 · Report as offensive     Reply Quote
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 10043
Credit: 967,756,956
RAC: 1,536,903
United States
Message 2005964 - Posted: 6 Aug 2019, 6:44:03 UTC

It would have to be a consensus among all the BOINC and Seti developers to increase the task per cpu count. The controls are in the server side software. The lead developer would be David Anderson.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 2005964 · Report as offensive     Reply Quote
rob smith Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer moderator
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 17870
Credit: 408,000,970
RAC: 44,487
United Kingdom
Message 2005965 - Posted: 6 Aug 2019, 7:08:06 UTC

It is down to the day-to-day project team, which is lead by Eric Korpela.
These day DA is taking a back seat from day-to-day life of SETI@Home, spending a chunk of his time working on Nebula and Science United.
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 2005965 · Report as offensive     Reply Quote
rob smith Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer moderator
Volunteer tester

Send message
Joined: 7 Mar 03
Posts: 17870
Credit: 408,000,970
RAC: 44,487
United Kingdom
Message 2005966 - Posted: 6 Aug 2019, 7:09:08 UTC - in response to Message 2005958.  

SETI is one of very few projects that use the processor limits.
Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?
ID: 2005966 · Report as offensive     Reply Quote
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 11751
Credit: 177,533,650
RAC: 164,703
Australia
Message 2005967 - Posted: 6 Aug 2019, 7:18:51 UTC - in response to Message 2005956.  

Is the 100 cpu task a BOINC limit or a Seti limit?

It is a Seti@home limit because the Seti@home servers can't cope with the load a larger amount of Results-out-in-the-field produces.
Grant
Darwin NT
ID: 2005967 · Report as offensive     Reply Quote
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 10043
Credit: 967,756,956
RAC: 1,536,903
United States
Message 2005969 - Posted: 6 Aug 2019, 7:20:46 UTC - in response to Message 2005965.  

It is down to the day-to-day project team, which is lead by Eric Korpela.
These day DA is taking a back seat from day-to-day life of SETI@Home, spending a chunk of his time working on Nebula and Science United.

I'm not talking about project management. I am speaking of code development. Eric has done very little in the past. Look at the contributors list at github/boinc. DA is the major code writer.
https://github.com/BOINC/boinc/graphs/contributors

The server code commits have been predominately by Tristan Olive, Shawn Kwang, Christian Beer, Kevin Reed and Vitali Koshura. Those are the guys you need to petition to increase the cpu task limit.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 2005969 · Report as offensive     Reply Quote
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5214
Credit: 436,543,385
RAC: 109,354
United States
Message 2005998 - Posted: 6 Aug 2019, 13:13:27 UTC - in response to Message 2005958.  
Last modified: 6 Aug 2019, 13:14:52 UTC

To Keith’s question on GPUGrid. I don’t think there is. I restrict by how many days worth, usually half day with no additional. Since there hasn’t been any for a long time it’s hard to remember.
ID: 2005998 · Report as offensive     Reply Quote
jsm

Send message
Joined: 1 Oct 16
Posts: 83
Credit: 38,941,362
RAC: 80,099
Isle of Man
Message 2006206 - Posted: 7 Aug 2019, 20:21:36 UTC - in response to Message 2005969.  

Well whoever is responsible I hope they will consider that to hold back frm change can lead to stagnation. I would be disappointed if TR3 comes out with 64/128, I shell out a considerable sum and then find it cannot run SETI at better than a Ryzen 3 m/c.
jsm
ID: 2006206 · Report as offensive     Reply Quote
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 11751
Credit: 177,533,650
RAC: 164,703
Australia
Message 2006276 - Posted: 8 Aug 2019, 6:23:19 UTC

AMD Rome Second Generation EPYC Review: 2x 64-core Benchmarked.
AMD have finally done it- they didn't just catch up with Intel, they're way out in front (for most workloads).

Summary-
Sixty-four cores. Each core with an improved Zen 2 core, offering ~15% better IPC performance than Naples (as tested in our consumer CPU review), and doubled AVX2/FP performance. The chip has a total of 256 MB of L3 cache, and 128 PCIe 4.0 lanes. AMD's second generation EPYC, in this case the EPYC 7742, is a behemoth...

For those with little time: at the high end with socketed x86 CPUs, AMD offers you up to 50 to 100% higher performance while offering at a 40% lower price. Unless you go for the low end server CPUs, there is no contest: AMD offers much better performance for a much lower price than Intel, with more memory channels and over 2x the number of PCIe lanes. These are also PCIe 4.0 lanes. What if you want to more than 2 TB of RAM in your dual socket server? The discount in favor of AMD just became 50%.

Grant
Darwin NT
ID: 2006276 · Report as offensive     Reply Quote
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 11751
Credit: 177,533,650
RAC: 164,703
Australia
Message 2006281 - Posted: 8 Aug 2019, 8:13:24 UTC

Phoronix has also taken a look at the new Epyc Rome CPUs.

AMD EPYC 7502 + EPYC 7742 Linux Performance Benchmarks
Grant
Darwin NT
ID: 2006281 · Report as offensive     Reply Quote
Profile ML1
Volunteer moderator
Volunteer tester

Send message
Joined: 25 Nov 01
Posts: 9949
Credit: 7,398,999
RAC: 73
United Kingdom
Message 2006297 - Posted: 8 Aug 2019, 12:19:00 UTC - in response to Message 2006281.  
Last modified: 8 Aug 2019, 12:19:37 UTC

Phoronix has also taken a look at the new Epyc Rome CPUs.

AMD EPYC 7502 + EPYC 7742 Linux Performance Benchmarks

Tom's Hardware also reports on that from earlier 'leaked' results:

AMD EPYC Rome 7742 Battles Intel Xeon in Benchmarks Posted Online

An anonymous source has submitted initial benchmarks of AMD's rumored 64-core, 128-thread EPYC 7742 processor to the publicly available OpenBenchmarking database. The posting has since been removed, but we managed to grab the test results...

... The EPYC 7742's alleged dominance continued to be seen in rendering workloads. The 64-core beast purportedly outperformed two Xeon Platinum 8280s...




The return of AMD has been a very long time coming...

Here's hoping there's no repeat of the 'dirty tricks' to stymie AMD this time round!


Happy fast crunchin'!
Martin
See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
ID: 2006297 · Report as offensive     Reply Quote
Profile Bill Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 30 Nov 05
Posts: 257
Credit: 4,373,402
RAC: 15,032
United States
Message 2006300 - Posted: 8 Aug 2019, 12:30:49 UTC

This may be obvious, but if you have so many PCIe lanes, memory lanes, and cores, wouldn't a server chock full of GPUs be the better way if you wanted to just purely crunch Seti/Boinc? I'm sure the cost is what prohibits most people from going this route, but I'm wondering if there are other reasons besides this.
Seti@home classic: 1,456 results, 1.613 years CPU time
ID: 2006300 · Report as offensive     Reply Quote
Previous · 1 . . . 24 · 25 · 26 · 27 · 28 · 29 · 30 . . . 38 · Next

Message boards : Number crunching : Ryzen and Threadripper


 
©2019 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.