mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Hardware > GPU Computing > GpuOwl

Reply
 
Thread Tools
Old 2020-10-27, 09:14   #2542
moebius
 
moebius's Avatar
 
Jul 2009
Germany

26·7 Posts
Default

Here again an updated version of the timing-list in .ots format for
OpenOffice.org Calc
LibreOffice Calc
StarOffice
IBM Lotus Symphony
NeoOffice (Mac OS)
Many thanks to the contributors!
Attached Files
File Type: zip TOP100-Grafic_Cards.zip (22.0 KB, 19 views)

Last fiddled with by moebius on 2020-10-27 at 09:41
moebius is offline   Reply With Quote
Old 2020-10-27, 22:08   #2543
firejuggler
 
firejuggler's Avatar
 
Apr 2010
Over the rainbow

9A416 Posts
Default

as requested,

Code:
D:\Download\prime95\gpuowl>gpuowl-win.exe -iters 200000 -prp 77936867
2020-10-27 22:39:48 gpuowl v6.11-380-g79ea0cc
2020-10-27 22:39:48 config: -maxAlloc 4500
2020-10-27 22:39:48 config: -device 0
2020-10-27 22:39:48 config: -iters 200000 -prp 77936867
2020-10-27 22:39:48 device 0, unique id ''
2020-10-27 22:39:48 GeForce GTX 1660 Ti-0 77936867 FFT: 4M 1K:8:256 (18.58 bpw)
2020-10-27 22:39:48 GeForce GTX 1660 Ti-0 Expected maximum carry32: 583B0000
2020-10-27 22:39:48 GeForce GTX 1660 Ti-0 OpenCL args "-DEXP=77936867u -DWIDTH=1024u -DSMALL_HEIGHT=256u -DMIDDLE=8u -DPM1=0 -DMM_CHAIN=1u -DMM2_CHAIN=2u -DMAX_ACCURACY=1 -DWEIGHT_STEP_MINUS_1=0xa.c42d0d7cec038p-5 -DIWEIGHT_STEP_MINUS_1=-0x8.0e50c8817ddf8p-5  -cl-unsafe-math-optimizations -cl-std=CL2.0 -cl-finite-math-only "
2020-10-27 22:39:51 GeForce GTX 1660 Ti-0

2020-10-27 22:39:51 GeForce GTX 1660 Ti-0 OpenCL compilation in 2.26 s
2020-10-27 22:39:54 GeForce GTX 1660 Ti-0 77936867 OK        0 loaded: blockSize 400, 0000000000000003
2020-10-27 22:39:54 GeForce GTX 1660 Ti-0 validating proof residues for power 8
2020-10-27 22:39:54 GeForce GTX 1660 Ti-0 Proof using power 8
2020-10-27 22:40:02 GeForce GTX 1660 Ti-0 77936867 OK      800   0.00%; 6467 us/it; ETA 5d 20:01; 1579c241dc63eca6 (check 2.68s)
2020-10-27 23:03:12 GeForce GTX 1660 Ti-0 Stopping, please wait..
2020-10-27 23:03:17 GeForce GTX 1660 Ti-0 77936867 OK   200000   0.26%; 6991 us/it; ETA 6d 06:58; f0b04b45b0855bd2 (check 2.82s)
2020-10-27 23:03:17 GeForce GTX 1660 Ti-0 Exiting because "stop requested"
2020-10-27 23:03:17 GeForce GTX 1660 Ti-0 Bye
Attached Thumbnails
Click image for larger version

Name:	Gpu_screen.gif
Views:	18
Size:	24.2 KB
ID:	23655  
firejuggler is online now   Reply With Quote
Old 2020-10-27, 22:22   #2544
moebius
 
moebius's Avatar
 
Jul 2009
Germany

26·7 Posts
Default

Thx a lot!
moebius is offline   Reply With Quote
Old 2020-10-28, 04:39   #2545
bverka86
 
May 2011

108 Posts
Default

Quote:
Originally Posted by preda View Post
For what task? P-1 is special, and is being reworked anyway.
Hey @preda, regarding respecting the "-log" argument, would changing (Gpu.cpp:1562) "... k % 10000 == 0 ..." to "... k % args.logStep == 0 ... " safely change the log output frequency without breaking another portion of the program?
bverka86 is offline   Reply With Quote
Old 2020-10-28, 07:52   #2546
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

101001011112 Posts
Default

Quote:
Originally Posted by bverka86 View Post
Hey @preda, regarding respecting the "-log" argument, would changing (Gpu.cpp:1562) "... k % 10000 == 0 ..." to "... k % args.logStep == 0 ... " safely change the log output frequency without breaking another portion of the program?
Well in fact not really. The behavior of the -log right now is to control how often to do the GEC check + save (by default 200k). That 10k display does not do much anything other than a bit of timing, no overhead, and does not have a config value (flag) yet.

But, what are you trying to achieve? (what is the reason for the change?)
preda is offline   Reply With Quote
Old 2020-10-28, 14:58   #2547
rebirther
 
rebirther's Avatar
 
Sep 2011
Germany

22·647 Posts
Default

Code:
2020-10-28 15:53:30 Note: not found 'config.txt'
2020-10-28 15:53:30 config: -iters 200000 -prp 77936867 
2020-10-28 15:53:30 device 0, unique id ''
2020-10-28 15:53:30 gfx1012-0 77936867 FFT: 4M 1K:8:256 (18.58 bpw)
2020-10-28 15:53:30 gfx1012-0 Expected maximum carry32: 583B0000
2020-10-28 15:53:30 gfx1012-0 OpenCL args "-DEXP=77936867u -DWIDTH=1024u -DSMALL_HEIGHT=256u -DMIDDLE=8u -DPM1=0 -DAMDGPU=1 -DMM_CHAIN=1u -DMM2_CHAIN=2u -DMAX_ACCURACY=1 -DWEIGHT_STEP_MINUS_1=0xa.c42d0d7cec038p-5 -DIWEIGHT_STEP_MINUS_1=-0x8.0e50c8817ddf8p-5  -cl-unsafe-math-optimizations -cl-std=CL2.0 -cl-finite-math-only "
2020-10-28 15:53:35 gfx1012-0 OpenCL compilation in 5.25 s
2020-10-28 15:53:37 gfx1012-0 77936867 OK        0 loaded: blockSize 400, 0000000000000003
2020-10-28 15:53:37 gfx1012-0 validating proof residues for power 8
2020-10-28 15:53:37 gfx1012-0 Proof using power 8
2020-10-28 15:53:41 gfx1012-0 77936867 OK      800   0.00%; 2971 us/it; ETA 2d 16:19; 1579c241dc63eca6 (check 1.24s)
2020-10-28 15:55:46 gfx1012-0 Stopping, please wait..
2020-10-28 15:55:48 gfx1012-0 77936867 OK    43200   0.06%; 2980 us/it; ETA 2d 16:29; b2fa36c486872d95 (check 1.23s)
2020-10-28 15:55:48 gfx1012-0 Exiting because "stop requested"
2020-10-28 15:55:48 gfx1012-0 Bye

Sapphire Pulse RX5500 XT 8GB
rebirther is offline   Reply With Quote
Old 2020-10-28, 17:11   #2548
moebius
 
moebius's Avatar
 
Jul 2009
Germany

44810 Posts
Default

The current timming list for graphics cards (always the latest version) can now be downloaded from google drive. Comment function is activated.
TOP100-Grafic_Cards.ots

Last fiddled with by moebius on 2020-10-28 at 17:40
moebius is offline   Reply With Quote
Old 2020-10-29, 01:53   #2549
bverka86
 
May 2011

23 Posts
Default

Ultimately I'm looking to just reduce the output, so I don't have to scroll for ages to find when errors are occurring (during the peak of the day for example, if the card is overheating in my box).

The RX470 is a new addition to this machine, so I'm still monitoring for stability issues for the compute task.

For anyone interested in cutting out the 10K iteration output, preda provided the following solution:



******

In the file "Gpu.cpp" :
Change this line (#1605)

if (k % 10000 == 0 && !doCheck) {

to:

if (false && k % 10000 == 0 && !doCheck) {

that should remove the extra log.

******


Quote:
Originally Posted by preda View Post
Well in fact not really. The behavior of the -log right now is to control how often to do the GEC check + save (by default 200k). That 10k display does not do much anything other than a bit of timing, no overhead, and does not have a config value (flag) yet.

But, what are you trying to achieve? (what is the reason for the change?)
bverka86 is offline   Reply With Quote
Old 2020-10-29, 12:07   #2550
Xyzzy
 
Xyzzy's Avatar
 
"Mike"
Aug 2002

171728 Posts
Default

Quote:
Originally Posted by moebius View Post
The current timming list for graphics cards (always the latest version) can now be downloaded from google drive.
Have you considered using Google "Sheets" to make the spreadsheet visible without needing local software?
Xyzzy is offline   Reply With Quote
Old 2020-10-29, 13:12   #2551
moebius
 
moebius's Avatar
 
Jul 2009
Germany

1C016 Posts
Default

Good idea, but now it's a new link. The table is now in ODF (.ods) format. I hope you can read it well in the browser regarding the resolution.
https://drive.google.com/file/d/1Tim...l57X_RWO0/view
moebius is offline   Reply With Quote
Old 2020-10-29, 13:17   #2552
Viliam Furik
 
Jul 2018
Martin, Slovakia

3708 Posts
Default

Quote:
Originally Posted by moebius View Post
Good idea, but now it's a new link. The table is now in ODF (.ods) format. I hope you can read it well in the browser regarding the resolution.
https://drive.google.com/file/d/1Tim...l57X_RWO0/view
I think Xyzzy meant you could make the table directly in Google Sheets, online. That means without using any offline program anywhere in the process.

But uploading it to Google Drive also works.
Viliam Furik is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
mfakto: an OpenCL program for Mersenne prefactoring Bdot GPU Computing 1657 2020-10-27 01:23
GPUOWL AMD Windows OpenCL issues xx005fs GpuOwl 0 2019-07-26 21:37
Testing an expression for primality 1260 Software 17 2015-08-28 01:35
Testing Mersenne cofactors for primality? CRGreathouse Computer Science & Computational Number Theory 18 2013-06-08 19:12
Primality-testing program with multiple types of moduli (PFGW-related) Unregistered Information & Answers 4 2006-10-04 22:38

All times are UTC. The time now is 04:06.

Tue Nov 24 04:06:56 UTC 2020 up 75 days, 1:17, 4 users, load averages: 2.74, 2.70, 2.53

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.