mersenneforum.org  

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

Reply
 
Thread Tools
Old 2020-09-05, 15:58   #2421
moebius
 
moebius's Avatar
 
Jul 2009
Germany

38110 Posts
Smile

Quote:
Originally Posted by Viliam Furik View Post
I checked the link. It seems that they only ship to some European countries.
Yes to the 16 most important European countries, but only to entrepreneurs. Private individuals can only order within Germany.

Last fiddled with by moebius on 2020-09-05 at 16:32
moebius is online now   Reply With Quote
Old 2020-09-05, 16:08   #2422
Uncwilly
6809 > 6502
 
Uncwilly's Avatar
 
"""""""""""""""""""
Aug 2003
101×103 Posts

25·3·7·13 Posts
Default

Quote:
Originally Posted by moebius View Post
Yes to the 16 most important European countries, but only to entrepreneurs. Private individuals can only order within Germany.
There are people on this forum with poor vision that have to use screen magnifiers or use other tools to be able to read the forum. Others read the forum on their phone. Please don't post small text as your default. Do this in consideration for others.
Uncwilly is online now   Reply With Quote
Old 2020-09-05, 17:07   #2423
storm5510
Random Account
 
storm5510's Avatar
 
Aug 2009
U.S.A.

2×809 Posts
Default

Quote:
Originally Posted by Uncwilly View Post
There are people on this forum with poor vision that have to use screen magnifiers or use other tools to be able to read the forum. Others read the forum on their phone. Please don't post small text as your default. Do this in consideration for others.
My other tool is a 4-inch diameter magnifying glass. Thankfully, Firefox has a zoom feature which I keep set to 120%.

I virtually had my head handed to me for mentioning this several months back by a few individuals who I will not name. I will simply say that I have a very long memory, and leave it go at that. Karma exists!
storm5510 is offline   Reply With Quote
Old 2020-09-05, 17:30   #2424
ATH
Einyen
 
ATH's Avatar
 
Dec 2003
Denmark

23×7×53 Posts
Default

Ok, I guess it was the old version v6.11-278-ga39cc1a I was still using, because it is slightly faster than the newest one (311 µs vs 322 µs at 3M FFT on V100). It
seems this old version have a different FFT crossover point.

But that version did do 350+ successful TC in 50M-53M, and it still gives correct results if I force 3M FFT. But I guess I better switch to the new version.

Code:
2020-09-05 15:51:04 config: -use ORIG_SLOWTRIG,CARRY32,NO_T2_SHUFFLE,OUT_WG=64,OUT_SIZEX=8,OUT_SPACING=4,IN_WG=128,IN_SIZEX=16,IN_SPACING=4
2020-09-05 15:51:04 config: -use STATS
2020-09-05 15:51:04 device 0, unique id ''
2020-09-05 15:51:04 Tesla V100-SXM2-16GB-0 54101381 FFT: 2.75M 256:11:512 (18.76 bpw)
2020-09-05 15:51:04 Tesla V100-SXM2-16GB-0 Expected maximum carry32: 511C0000
2020-09-05 15:51:04 Tesla V100-SXM2-16GB-0 NO_T2_SHUFFLE not used
2020-09-05 15:51:04 Tesla V100-SXM2-16GB-0 OpenCL args "-DEXP=54101381u -DWIDTH=256u -DSMALL_HEIGHT=512u -DMIDDLE=11u -DWEIGHT_STEP=0x1.2df227890f2f2p+0 -DIWEIGHT_STEP=0x1.b2174acb8edb3p-1 -DWEIGHT_BIGSTEP=0x1.ae89f995ad3adp+0 -DIWEIGHT_BIGSTEP=0x1.306fe0a31b715p-1 -DPM1=0 -DMM_CHAIN=2u -DMM2_CHAIN=3u -DCARRY32=1 -DIN_SIZEX=16 -DIN_SPACING=4 -DIN_WG=128 -DNO_T2_SHUFFLE=1 -DORIG_SLOWTRIG=1 -DOUT_SIZEX=8 -DOUT_SPACING=4 -DOUT_WG=64 -DSTATS=1  -cl-fast-relaxed-math -cl-std=CL2.0 "
2020-09-05 15:51:08 Tesla V100-SXM2-16GB-0

2020-09-05 15:51:08 Tesla V100-SXM2-16GB-0 OpenCL compilation in 3.85 s
2020-09-05 15:51:08 Tesla V100-SXM2-16GB-0 54101381 LL        0 loaded: 0000000000000004
2020-09-05 15:51:38 Tesla V100-SXM2-16GB-0 Roundoff: N=100000, mean 0.290692, SD 0.018230, CV 0.062711, max 0.437387, z 11.5 (pErr 99.999508%)
2020-09-05 15:51:38 Tesla V100-SXM2-16GB-0 54101381 LL   100000   0.18%;  296 us/it; ETA 0d 04:26; 98857a05e8c1d7ca
2020-09-05 15:52:07 Tesla V100-SXM2-16GB-0 Roundoff: N=100000, mean 0.290730, SD 0.017680, CV 0.060812, max 0.433674, z 11.8 (pErr 99.957106%)
2020-09-05 15:52:07 Tesla V100-SXM2-16GB-0 54101381 LL   200000   0.37%;  296 us/it; ETA 0d 04:26; 92591a2844762ceb
2020-09-05 15:52:37 Tesla V100-SXM2-16GB-0 Roundoff: N=100000, mean 0.290681, SD 0.017714, CV 0.060941, max 0.448847, z 11.8 (pErr 99.965018%)
2020-09-05 15:52:37 Tesla V100-SXM2-16GB-0 54101381 LL   300000   0.55%;  296 us/it; ETA 0d 04:26; 02f62a87c5bde161
2020-09-05 15:53:06 Tesla V100-SXM2-16GB-0 Roundoff: N=100000, mean 0.290703, SD 0.017701, CV 0.060890, max 0.442235, z 11.8 (pErr 99.962138%)
2020-09-05 15:53:06 Tesla V100-SXM2-16GB-0 54101381 LL   400000   0.74%;  296 us/it; ETA 0d 04:25; de0885175f96345b
ATH is online now   Reply With Quote
Old 2020-09-07, 00:33   #2425
mathwiz
 
Mar 2019

101 Posts
Default upload.py stuck

I'm seeing ./tools/upload.py consistently getting stuck around 97%-98% on multiple machines. Is this a known issue?
mathwiz is offline   Reply With Quote
Old 2020-09-07, 01:10   #2426
ATH
Einyen
 
ATH's Avatar
 
Dec 2003
Denmark

23·7·53 Posts
Default

Mine has often been stuck at 96%. When I try later to upload it again, it says it was already uploaded, and the certificate has already been approved.
ATH is online now   Reply With Quote
Old 2020-09-07, 01:58   #2427
Prime95
P90 years forever!
 
Prime95's Avatar
 
Aug 2002
Yeehaw, FL

2×43×83 Posts
Default

Quote:
Originally Posted by mathwiz View Post
I'm seeing ./tools/upload.py consistently getting stuck around 97%-98% on multiple machines. Is this a known issue?
I may have found the problem on the server side. We were getting timeouts at 300 seconds due to a locking issue.
Prime95 is offline   Reply With Quote
Old 2020-09-08, 19:52   #2428
moebius
 
moebius's Avatar
 
Jul 2009
Germany

3·127 Posts
Default

Quote:
Originally Posted by moebius View Post
now no longer available, you have to be quick nowadays
moebius is online now   Reply With Quote
Old 2020-09-09, 16:00   #2429
Viliam Furik
 
Jul 2018
Martin, Slovakia

22·5·11 Posts
Default Error on load

Code:
2020-09-09 17:57:48 config: -device 1
2020-09-09 17:57:48 config: -proof 8
2020-09-09 17:57:48 config: -nospin
2020-09-09 17:57:48 config: -jacobi 500000
2020-09-09 17:57:48 device 1, unique id ''
2020-09-09 17:57:48 gfx906-1 107826457 FFT: 6M 1K:12:256 (17.14 bpw)
2020-09-09 17:57:48 gfx906-1 Expected maximum carry32: 28B00000
2020-09-09 17:57:49 gfx906-1 OpenCL args "-DEXP=107826457u -DWIDTH=1024u -DSMALL_HEIGHT=256u -DMIDDLE=12u -DPM1=0 -DAMDGPU=1 -DWEIGHT_STEP_MINUS_1=0xd.11da6d175f558p-4 -DIWEIGHT_STEP_MINUS_1=-0xe.631e6aa29fc1p-5  -cl-unsafe-math-optimizations -cl-std=CL2.0 -cl-finite-math-only "
2020-09-09 17:57:49 gfx906-1 ASM compilation failed, retrying compilation using NO_ASM
2020-09-09 17:57:53 gfx906-1 OpenCL compilation in 4.76 s
2020-09-09 17:57:54 gfx906-1 107826457 EE 101144000 loaded: blockSize 400, e2ec8057148fe5cb (expected a63990e20177b9c6)
2020-09-09 17:57:54 gfx906-1 Exiting because "error on load"
2020-09-09 17:57:54 gfx906-1 Bye
Is there something I can do about this? Other than restart the test?
Viliam Furik is offline   Reply With Quote
Old 2020-09-09, 16:11   #2430
moebius
 
moebius's Avatar
 
Jul 2009
Germany

5758 Posts
Default

Quote:
Originally Posted by Viliam Furik View Post
Is there something I can do about this? Other than restart the test?
If it is a PRP test, there should be a 107826457-old.owl file in the directory of the exponent, For LL I don't know.

Last fiddled with by moebius on 2020-09-09 at 16:13
moebius is online now   Reply With Quote
Old 2020-09-09, 16:17   #2431
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

3·2,957 Posts
Default

Quote:
Originally Posted by Viliam Furik View Post
Code:
2020-09-09 17:57:54 gfx906-1 Exiting because "error on load"
2020-09-09 17:57:54 gfx906-1 Bye
Is there something I can do about this? Other than restart the test?
No. This is a long standing bug in gpuOwl. Mihai was kinda busy with switching to the new methods (certs) or he doesn't care
This is bothering, but harmless. I mean, not totally harmless, in the sense that if you don't spot it in time, your card(s) can stay idle for ages. For me it happens with the card that drives the display, mostly, regardless of which card I use for that, or in which slot I put it, while the other card(s) won't pop such errors, or would pop them more seldom. The solution I use to avoid continuously watching if the owl crashed or not, is along starting the owl from a batch, which also has the advantage that I don't need to remember the parameters and options all the time, something along the lines:
Code:
:again
gpuowl -device 0 -nospin -log blahblahblah-other-parameters
goto again
pause
This way, as long as there is work to do in the file, any "reload" will be handles by the endless loop in the batch.

Last fiddled with by LaurV on 2020-09-09 at 16:22
LaurV is offline   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 1656 2020-10-13 14:21
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 21:46.

Mon Oct 26 21:46:18 UTC 2020 up 46 days, 18:57, 0 users, load averages: 1.64, 1.78, 1.81

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.