PrimeGrid
Please visit donation page to help the project cover running costs for this month

Toggle Menu

Join PrimeGrid

Returning Participants

Community

Leader Boards

Results

Other

drummers-lowrise
1) Message boards : Wieferich and Wall-Sun-Sun Prime Search : Slow speed on RX 5700XT (Message 146482)
Posted 838 days ago by SAKAGE@AMD@jisakuProject donor
I confirmed that GPU usage has been improved in the WW1.03 application.
2) Message boards : Wieferich and Wall-Sun-Sun Prime Search : Slow speed on RX 5700XT (Message 145922)
Posted 848 days ago by SAKAGE@AMD@jisakuProject donor
There seems to be such a tendency in RADEON.
I don't know what caused it.
3) Message boards : Number crunching : Évariste Galois's Birthday Challenge (Message 144450)
Posted 889 days ago by SAKAGE@AMD@jisakuProject donor
Certainly the scores are strange.
4) Message boards : Number crunching : LLR2 installed on all big LLR projects (Message 143341)
Posted 923 days ago by SAKAGE@AMD@jisakuProject donor
This is great!
I will donate to this achievement. but sorry for small amount.
5) Message boards : Number crunching : International Bacon Day Challenge (Message 142976)
Posted 936 days ago by SAKAGE@AMD@jisakuProject donor
It seems that the task will not come again.
6) Message boards : Number crunching : International Bacon Day Challenge (Message 142973)
Posted 936 days ago by SAKAGE@AMD@jisakuProject donor
fixed.
It came.
7) Message boards : Number crunching : International Bacon Day Challenge (Message 142972)
Posted 936 days ago by SAKAGE@AMD@jisakuProject donor
There are no new tasks coming.
8) Message boards : Number crunching : Year of the Rat Challenge (Message 138986)
Posted 1107 days ago by SAKAGE@AMD@jisakuProject donor
I have two 3970X machines. FWIW, I find that MT tasks run ~2x faster under linux than Windows. My guess is that it has something to do with the way the OS keeps MT tasks assigned to the chiplets. I think linux is just better at it.


On my 2990WX, Linux is at least 4x faster than Windows on these SR5 tasks. The big TR's really need Linux to perform at their potential.


hmm..
Thank you for the helpful information.
9) Message boards : Number crunching : Year of the Rat Challenge (Message 138970)
Posted 1107 days ago by SAKAGE@AMD@jisakuProject donor
Hi SAKAGE@AMD@jisaku and Chooka,

I'm still interested in the performance you report and only wish I had the hardware to try it myself. In theory the Zen 2 CPUs should be really great in this use case, but I only bought up to 3700X (8 cores) so I can't test if something else is happening at higher core counts.

Could I request you do a Prime95 benchmark as I describe in the post linked?
http://www.primegrid.com/forum_thread.php?id=8240&nowrap=true#138938

SAKAGE@AMD@jisaku: please try 1, 2, 4, 8, 16, 32 workers on 3970X
Chook: please try 1, 2, 4, 8, 16 workers on both 1950X and 3950X

Obviously this should be done when nothing else is using CPU. It should be ok to suspend computation in BOINC temporarily for this. Results are also written to a text file results.bench.txt in the prime95 folder. Suggest you copy/paste the results in a private message to me to avoid filling this thread up.


I tested prime95 on 3970x and 3900x.
From this result, it seems that prime95 has a throughput corresponding to the number of cores.

CPUModel="AMD Ryzen Threadripper 3970X 32-Core Processor "
FFTlen=768K all-complex, (32 cores, 1 worker): 0.81 ms. Throughput: 1230.75 iter/sec.
FFTlen=768K all-complex, (32 cores, 2 workers): 0.63, 0.59 ms. Throughput: 3265.77 iter/sec.
FFTlen=768K all-complex, (32 cores, 4 workers): 0.58, 0.57, 0.60, 0.54 ms. Throughput: 6991.00 iter/sec.
FFTlen=768K all-complex, (32 cores, 8 workers): 0.73, 0.69, 0.75, 0.71, 0.78, 0.69, 0.69, 0.69 ms. Throughput: 11184.88 iter/sec.
FFTlen=768K all-complex, (32 cores, 16 workers): 1.47, 1.36, 1.38, 1.37, 1.43, 1.48, 1.35, 1.38, 1.49, 1.44, 1.35, 1.35, 1.36, 1.39, 1.37, 1.35 ms. Throughput: 11484.03 iter/sec.
FFTlen=768K all-complex, (32 cores, 32 workers): 11.87, 10.07, 8.86, 8.67, 9.76, 9.27, 8.96, 8.96, 10.31, 9.34, 9.81, 9.28, 9.46, 9.26, 9.03, 9.06, 10.95, 9.23, 8.53, 8.32, 11.08, 9.54, 9.00, 8.65, 9.56, 9.33, 9.41, 9.39, 11.29, 8.84, 9.05, 8.62 ms. Throughput: 3405.75 iter/sec.

CPUModel="AMD Ryzen 9 3900X 12-Core Processor "
FFTlen=768K all-complex, (12 cores, 1 worker): 0.49 ms. Throughput: 2056.66 iter/sec.
FFTlen=768K all-complex, (12 cores, 2 workers): 0.62, 0.59 ms. Throughput: 3295.53 iter/sec.
FFTlen=768K all-complex, (12 cores, 4 workers): 0.88, 0.88, 0.88, 0.88 ms. Throughput: 4558.04 iter/sec.
FFTlen=768K all-complex, (12 cores, 6 workers): 2.50, 1.27, 2.38, 1.28, 0.86, 0.87 ms. Throughput: 4696.47 iter/sec.
FFTlen=768K all-complex, (12 cores, 12 workers): 5.83, 5.46, 4.31, 4.41, 5.75, 5.02, 4.55, 4.52, 6.02, 4.28, 4.78, 5.94 ms. Throughput: 2405.59 iter/sec.
10) Message boards : Number crunching : Year of the Rat Challenge (Message 138960)
Posted 1108 days ago by SAKAGE@AMD@jisakuProject donor
I tried the setting of 1 thread and 8 threads, but the situation has not changed.
Only half the throughput (per core) compared to AM4 Ryzen (Zen2).


Next 10 posts
[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2023 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 1.86, 1.52, 1.56
Generated 29 Mar 2023 | 3:40:23 UTC