mersenneforum.org  

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

Reply
 
Thread Tools
Old 2020-10-07, 20:37   #67
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

24608 Posts
Default

Quote:
Originally Posted by retina View Post
How come a mutex or semaphore can't work here? That is the primary use case those primitives were made for.
Those are traditionally used for exclusion between the threads of a single process. In our situation we needed locking between processes. One way to lock across-processes is to use a form of shared-memory, but that's not very portable unix/windows. Given the very low performance requirements of our locking, a filesystem solution is fine.
preda is offline   Reply With Quote
Old 2020-10-07, 20:40   #68
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

24·83 Posts
Default

Quote:
Originally Posted by Aramis Wyler View Post
I find myself again on the same page as M344587487 and again confused by the response.

The P-1 portion of the PRP run is in the begining, right? It does FS up to the bound, stops PRP'ing, runs the SS, and then picks back up again with the PRP. Wouldn't it release the memory at the end of the SS and spend the rest of the time with small memory footprint, at which point we could start the second PRP job with large memory?
I found M344587487's explanation (below) perfectly spot-on, that's exacly what is happening. (also what you ask, above, is correct yes)

Quote:
Assuming the large memory requirement is within a single time interval memlock is a simple way to get the processes out of phase and should do the job fine. Assuming the normal case of the queued exponents being close and slowly increasing there should be barely any stalls after the initial one, barring the occasional small exponent from a previously expired allocation knocking the processes back in phase. It might be wise to let the processes get a bit more out of phase than immediately needed to account for most of the variability but that's only if micro-stalls are considered a problem.

Last fiddled with by preda on 2020-10-07 at 20:41
preda is offline   Reply With Quote
Old 2020-10-07, 21:07   #69
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

24·83 Posts
Default

For anybody trying out the new P-1, please don't forget to start with a few known-factors and verify they are detected correctly. If any factor should be found and isn't please report as it's a very serious bug.
preda is offline   Reply With Quote
Old 2020-10-08, 09:11   #70
firejuggler
 
firejuggler's Avatar
 
Apr 2010
Over the rainbow

2×17×73 Posts
Default

Hi preda. no bug found so far, but something strange.
Code:
D:\Download\prime95\gpuowl\v7>gpuowl-win.exe -maxAlloc 4500 -B1 200000  -prp 13820713
2020-10-08 10:36:51 gpuowl v7.0-18-g69c2b85
2020-10-08 10:36:51 Note: not found 'config.txt'
2020-10-08 10:36:51 config: -maxAlloc 4500 -B1 200000 -prp 13820713
2020-10-08 10:36:51 device 0, unique id ''
2020-10-08 10:36:51 GeForce GTX 1660 Ti-0 13820713 FFT: 768K 256:6:256 (17.57 bpw)
2020-10-08 10:36:51 GeForce GTX 1660 Ti-0 13820713 OpenCL args "-DEXP=13820713u -DWIDTH=256u -DSMALL_HEIGHT=256u -DMIDDLE=6u -DCARRY64=1 -DCARRYM64=1 -DWEIGHT_STEP_MINUS_1=0xa.fe659a0540518p-5 -DIWEIGHT_STEP_MINUS_1=-0x8.2ebd6d5c39b58p-5  -cl-unsafe-math-optimizations -cl-std=CL2.0 -cl-finite-math-only "
2020-10-08 10:36:53 GeForce GTX 1660 Ti-0 13820713

2020-10-08 10:36:53 GeForce GTX 1660 Ti-0 13820713 OpenCL compilation in 1.89 s
2020-10-08 10:36:53 GeForce GTX 1660 Ti-0 13820713 maxAlloc: 4.4 GB
2020-10-08 10:36:53 GeForce GTX 1660 Ti-0 13820713 Space for 1484 B1 buffers (available mem 4453.5 MB, buf size 3.0 MB)
2020-10-08 10:36:53 GeForce GTX 1660 Ti-0 13820713 B1=200000 (288610 bits)
2020-10-08 10:36:53 GeForce GTX 1660 Ti-0 13820713 PRP starting from beginning
2020-10-08 10:36:53 GeForce GTX 1660 Ti-0 13820713 powerSmooth(200000), 288610 bits, took 0.00s (CPU)
2020-10-08 10:36:53 GeForce GTX 1660 Ti-0 13820713 Aquired memory lock 'memlock-0'
2020-10-08 10:36:53 GeForce GTX 1660 Ti-0 13820713 B1: allocating 1479 buffers
2020-10-08 10:36:53 GeForce GTX 1660 Ti-0 13820713 Starting B1=200000, first bit 25
2020-10-08 10:36:54 GeForce GTX 1660 Ti-0 13820713 OK        0 loaded: blockSize 500, 0000000000000003
2020-10-08 10:36:54 GeForce GTX 1660 Ti-0 13820713 validating proof residues for power 8
2020-10-08 10:36:54 GeForce GTX 1660 Ti-0 13820713 Proof using power 8
2020-10-08 10:37:00 GeForce GTX 1660 Ti-0 13820713 B1 fold(1479) (78 set) took 3.18s
2020-10-08 10:37:00 GeForce GTX 1660 Ti-0 13820713 OK     1000   0.01%; 2216 us/it; ETA 0d 08:30; dcd805da65a00649
2020-10-08 10:42:59 GeForce GTX 1660 Ti-0 13820713 B1 fold(1479) (1479 set) took 8.90s
2020-10-08 10:42:59 GeForce GTX 1660 Ti-0 13820713 OK   200000   1.45%; 1757 us/it; ETA 0d 06:39; 3d376960be2ed1dc
2020-10-08 10:45:47 GeForce GTX 1660 Ti-0 13820713 B1 fold(1479) (1458 set) took 9.57s
2020-10-08 10:45:47 GeForce GTX 1660 Ti-0 13820713 B1 200000: releasing 1479 buffers
2020-10-08 10:45:48 GeForce GTX 1660 Ti-0 13820713 Released memory lock 'memlock-0'
2020-10-08 10:45:48 GeForce GTX 1660 Ti-0 13820713 OK   289000   2.09%; 1782 us/it; ETA 0d 06:42; 3e21edc60d08e110
2020-10-08 10:45:48 GeForce GTX 1660 Ti-0 13820713 B1 completed. Starting GCD
2020-10-08 10:45:48 GeForce GTX 1660 Ti-0 13820713 P2 (200000,4000000) will continue from B2=0
2020-10-08 10:45:48 GeForce GTX 1660 Ti-0 13820713 P2 B1=200000, B2=4000000, D=2310: 265319 primes in [199815, 4002075], selected 220904 (83.3%) (44415 doubles + 176489 singles)
2020-10-08 10:45:48 GeForce GTX 1660 Ti-0 13820713 P2 B1=200000, B2=4000000, D=2310 from B2=0 : 1646 blocks starting at 87
2020-10-08 10:45:48 GeForce GTX 1660 Ti-0 13820713 P2 Aquired memory lock 'memlock-0'
2020-10-08 10:45:48 GeForce GTX 1660 Ti-0 13820713 P2 Allocated 240 P2 buffers
2020-10-08 10:45:50 GeForce GTX 1660 Ti-0 13820713 P2 Setup 240 P2 buffers in 2173.0 ms
2020-10-08 10:45:51 GeForce GTX 1660 Ti-0 13820713 P2   202125/4000000 (  0%); 152 muls, 1838 us/mul
2020-10-08 10:45:51 GeForce GTX 1660 Ti-0 13820713 P2 GCD : no factor
2020-10-08 10:46:37 GeForce GTX 1660 Ti-0 13820713 P2   664125/4000000 ( 12%); 29116 muls, 1605 us/mul
2020-10-08 10:47:23 GeForce GTX 1660 Ti-0 13820713 P2  1126125/4000000 ( 24%); 27881 muls, 1626 us/mul
2020-10-08 10:48:09 GeForce GTX 1660 Ti-0 13820713 P2  1588125/4000000 ( 37%); 27217 muls, 1712 us/mul
2020-10-08 10:48:53 GeForce GTX 1660 Ti-0 13820713 P2  2050125/4000000 ( 49%); 26781 muls, 1646 us/mul
2020-10-08 10:49:37 GeForce GTX 1660 Ti-0 13820713 P2  2512125/4000000 ( 61%); 26355 muls, 1652 us/mul
2020-10-08 10:50:20 GeForce GTX 1660 Ti-0 13820713 P2  2974125/4000000 ( 73%); 26071 muls, 1640 us/mul
2020-10-08 10:50:51 GeForce GTX 1660 Ti-0 13820713 P2 Starting GCD
2020-10-08 10:50:54 GeForce GTX 1660 Ti-0 13820713 P2 GCD : no factor
2020-10-08 10:51:03 GeForce GTX 1660 Ti-0 13820713 P2  3436125/4000000 ( 85%); 25943 muls, 1660 us/mul
2020-10-08 10:51:44 GeForce GTX 1660 Ti-0 13820713 P2  3898125/4000000 ( 97%); 25680 muls, 1624 us/mul
2020-10-08 10:51:54 GeForce GTX 1660 Ti-0 13820713 P2  4002075/4000000 (100%); 5708 muls, 1621 us/mul
2020-10-08 10:51:54 GeForce GTX 1660 Ti-0 13820713 P2 Starting GCD
2020-10-08 10:51:54 GeForce GTX 1660 Ti-0 13820713 P2 Released memory lock 'memlock-0'
2020-10-08 10:51:57 GeForce GTX 1660 Ti-0 13820713 GCD: no factor
2020-10-08 10:51:57 GeForce GTX 1660 Ti-0 13820713 {"status":"NF", "exponent":"13820713", "worktype":"PM1", "B1":"200000", "B2":"4000000", "fft-length":"786432", "program":{"name":"gpuowl", "version":"v7.0-18-g69c2b85"}, "computer":"GeForce GTX 1660 Ti-0", "timestamp":"2020-10-08 08:51:57 UTC"}
As you can see, there seem to have 3 GCD, and the first one is stated as P2.
Windows 10, and compiled by kriesel.

Last fiddled with by firejuggler on 2020-10-08 at 09:12
firejuggler is offline   Reply With Quote
Old 2020-10-08, 10:05   #71
retina
Undefined
 
retina's Avatar
 
"The unspeakable one"
Jun 2006
My evil lair

2·17·173 Posts
Default

Quote:
Originally Posted by M344587487 View Post
Creating a file lock is implementing a mutex at the process level ....
Quote:
Originally Posted by preda View Post
Those are traditionally used for exclusion between the threads of a single process.
https://docs.microsoft.com/en-us/win...-mutex-objects
Quote:
You can use a mutex object to protect a shared resource from simultaneous access by multiple threads or processes.
It really is their primary purpose.

?
retina is online now   Reply With Quote
Old 2020-10-08, 10:38   #72
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

24×83 Posts
Default

Quote:
Originally Posted by retina View Post
https://docs.microsoft.com/en-us/win...-mutex-objects It really is their primary purpose.

?
That's a windows only API, not applicable. I was refering to mutex in the C++ sense (which is platform agnostic being part of the C++ language).
preda is offline   Reply With Quote
Old 2020-10-08, 10:44   #73
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

24·83 Posts
Default

Quote:
Originally Posted by firejuggler View Post
Hi preda. no bug found so far, but something strange.
[..]
As you can see, there seem to have 3 GCD, and the first one is stated as P2.
It's not possible to run "intermediary" GCD on P1 (i.e. first stage), only at the end. Thus when P1 ends, a first GCD is started. P2 is also started imediatelly after P1 (thus, in parallel with the P1 GCD). At some point during P2 that first GCD completes and is reported.

During P2 (second stage), it is possible to run GCD as often as desired. Right now a GCD is started every 5minutes after the completion of the previous GCD. This is likely overkill (too often) and I plan to change the interval to 10m or 15m. But that's the explanations for the middle GCD you see. One more GCD at the end of P2.
preda is offline   Reply With Quote
Old 2020-10-08, 10:48   #74
M344587487
 
M344587487's Avatar
 
"Composite as Heck"
Oct 2017

12728 Posts
Default

You really need to tag MS API links as NSFW, I gagged ;)


You're right, it looks like it's trivial to have completely separate (as in no child/fork) processes share a posix named semaphore so that should work just as well: https://github.com/yvoinov/interproc...er/named_sem.h
M344587487 is online now   Reply With Quote
Old 2020-10-08, 17:11   #75
kriesel
 
kriesel's Avatar
 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

10010100010002 Posts
Default

Quote:
Originally Posted by preda View Post
During P2 (second stage), it is possible to run GCD as often as desired. Right now a GCD is started every 5minutes after the completion of the previous GCD. This is likely overkill (too often) and I plan to change the interval to 10m or 15m. But that's the explanations for the middle GCD you see. One more GCD at the end of P2.
Optimizing the GCD frequency in P2 for both gpu usage and cpu usage versus odds of saving further effort by detecting an already passed factor would be good. Every 5 minutes is rather too frequent, especially as exponent increases, and on slower hardware. The cpu usage spent on many intermediate GCDs has a cast, as it would otherwise be useful in PRP on mprime running on the same system.
Or add a -use option GCDinterval n where n is minutes between P2 GCDs, or n% of B2 or something that scales well.

There is seemingly a way to change B1; drop the number of tests saved from 1 or 2 to zero, and a continuation shows B1=0.

Gpuowl-win v7.0-18 worked to duplicate a factor found recently with prime95. https://www.mersenne.ca/exponent/100342469
But there was an issue with v7.0 renaming worktodo.txt-tmp to worktodo.txt afterward, that caused the run to terminate.
If the iteration times look a bit long, it's because there was an instance of v6.11-380 running LLDC of 178476583 on the same 8GB RX480 also.

rx480, Windows 7 Pro x64
v6.11-380 alone, 178M LL: 6733us/it = 148.52 it/sec

V7.0-18 alone, 100M PRP&P-1: 4358us/it = 229.46 it/sec

Combined:
V7.0 9528 us/it 104.95 it/sec 45.74% of single throughput
V6.11 12232 us/it 81.75 it/sec 55.04% of single
100.78%

Code:
>gpuowl-win
2020-10-08 09:42:14 gpuowl v7.0-18-g69c2b85
2020-10-08 09:42:14 config: -device 0 -user kriesel -cpu condorella/rx480 -yield -maxAlloc 7000 -proof 9 -use NO_ASM
2020-10-08 09:42:14 device 0, unique id ''
2020-10-08 09:42:14 condorella/rx480 100342469 FFT: 5.50M 1K:11:256 (17.40 bpw)
2020-10-08 09:42:17 condorella/rx480 100342469 OpenCL args "-DEXP=100342469u -DWIDTH=1024u -DSMALL_HEIGHT=256u -DMIDDLE=11u -DAMDGPU=1 -DCARRY64=1 -DCARRYM64=1 -DWEIGHT_STEP_MINUS_1=0x8.450b256a8e68p-4 -DIWEIGHT_STEP_MINUS_1=-0xa.e75b5052df808p-5 -DNO_ASM=1  -cl-unsafe-math-optimizations -cl-std=CL2.0 -cl-finite-math-only "
2020-10-08 09:42:22 condorella/rx480 100342469 OpenCL compilation in 4.95 s
2020-10-08 09:42:22 condorella/rx480 100342469 maxAlloc: 6.8 GB
2020-10-08 09:42:22 condorella/rx480 100342469 Space for 304 B1 buffers (available mem 6684.4 MB, buf size 22.0 MB)
2020-10-08 09:42:22 condorella/rx480 100342469 B1=12000 (17314 bits)
2020-10-08 09:42:22 condorella/rx480 100342469 PRP starting from beginning
2020-10-08 09:42:22 condorella/rx480 100342469 powerSmooth(12000), 17314 bits, took 0.00s (CPU)
2020-10-08 09:42:22 condorella/rx480 100342469 Aquired memory lock 'memlock-0'
2020-10-08 09:42:23 condorella/rx480 100342469 B1: allocating 299 buffers
2020-10-08 09:42:23 condorella/rx480 100342469 Starting B1=12000, first bit 21
2020-10-08 09:42:25 condorella/rx480 100342469 OK        0 loaded: blockSize 500, 0000000000000003
2020-10-08 09:42:25 condorella/rx480 100342469 validating proof residues for power 9
2020-10-08 09:42:25 condorella/rx480 100342469 Proof using power 9
2020-10-08 09:42:35 condorella/rx480 100342469 B1 fold(299) (83 set) took 2.95s
2020-10-08 09:42:35 condorella/rx480 100342469 OK     1000   0.00%; 6385 us/it; ETA 7d 09:58; 1676d0797efebcee
2020-10-08 09:43:54 condorella/rx480 100342469 B1 fold(299) (293 set) took 4.60s
2020-10-08 09:43:54 condorella/rx480 100342469 B1 12000: releasing 299 buffers
2020-10-08 09:43:54 condorella/rx480 100342469 Released memory lock 'memlock-0'
2020-10-08 09:43:54 condorella/rx480 100342469 OK    17500   0.02%; 4482 us/it; ETA 5d 04:54; 44fb78d0e7bf8ca0
2020-10-08 09:43:54 condorella/rx480 100342469 B1 completed. Starting GCD
2020-10-08 09:43:54 condorella/rx480 100342469 P2 (12000,10000000) will continue from B2=0
2020-10-08 09:43:55 condorella/rx480 100342469 P2 B1=12000, B2=10000000, D=210: 663159 primes in [11865, 10000095], selected 566879 (85.5%) (96280 doubles + 470599 singles)
2020-10-08 09:43:55 condorella/rx480 100342469 P2 B1=12000, B2=10000000, D=210 from B2=0 : 47563 blocks starting at 57
2020-10-08 09:43:55 condorella/rx480 100342469 P2 Aquired memory lock 'memlock-0'
2020-10-08 09:43:55 condorella/rx480 100342469 P2 Allocated 24 P2 buffers
2020-10-08 09:43:55 condorella/rx480 100342469 P2 Setup 24 P2 buffers in 757.6 ms
2020-10-08 09:43:56 condorella/rx480 100342469 P2    12075/10000000 (  0%); 17 muls, 7303 us/mul
2020-10-08 09:45:34 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 09:48:00 condorella/rx480 100342469 P2   474075/10000000 (  5%); 31200 muls, 7837 us/mul
2020-10-08 09:50:35 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 09:52:17 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 09:53:37 condorella/rx480 100342469 P2   936075/10000000 (  9%); 28809 muls, 11720 us/mul
2020-10-08 09:57:18 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 09:59:00 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 09:59:07 condorella/rx480 100342469 P2  1398075/10000000 ( 14%); 27935 muls, 11790 us/mul
2020-10-08 10:04:00 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 10:04:30 condorella/rx480 100342469 P2  1860075/10000000 ( 19%); 27440 muls, 11799 us/mul
2020-10-08 10:05:42 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 10:09:52 condorella/rx480 100342469 P2  2322075/10000000 ( 23%); 26965 muls, 11941 us/mul
2020-10-08 10:10:43 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 10:12:25 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 10:15:11 condorella/rx480 100342469 P2  2784075/10000000 ( 28%); 26655 muls, 11965 us/mul
2020-10-08 10:17:25 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 10:19:07 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 10:20:25 condorella/rx480 100342469 P2  3246075/10000000 ( 32%); 26509 muls, 11881 us/mul
2020-10-08 10:24:08 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 10:25:36 condorella/rx480 100342469 P2  3708075/10000000 ( 37%); 26202 muls, 11861 us/mul
2020-10-08 10:25:49 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 10:30:49 condorella/rx480 100342469 P2  4170075/10000000 ( 42%); 26038 muls, 12026 us/mul
2020-10-08 10:30:51 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 10:32:33 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 10:36:04 condorella/rx480 100342469 P2  4632075/10000000 ( 46%); 25904 muls, 12173 us/mul
2020-10-08 10:37:34 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 10:39:15 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 10:41:14 condorella/rx480 100342469 P2  5094075/10000000 ( 51%); 25791 muls, 12009 us/mul
2020-10-08 10:44:16 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 10:45:58 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 10:46:24 condorella/rx480 100342469 P2  5556075/10000000 ( 56%); 25567 muls, 12156 us/mul
2020-10-08 10:50:58 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 10:51:31 condorella/rx480 100342469 P2  6018075/10000000 ( 60%); 25506 muls, 12016 us/mul
2020-10-08 10:52:40 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 10:56:36 condorella/rx480 100342469 P2  6480075/10000000 ( 65%); 25377 muls, 12037 us/mul
2020-10-08 10:57:40 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 10:59:21 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 11:01:43 condorella/rx480 100342469 P2  6942075/10000000 ( 69%); 25398 muls, 12098 us/mul
2020-10-08 11:04:21 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 11:06:03 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 11:06:50 condorella/rx480 100342469 P2  7404075/10000000 ( 74%); 25239 muls, 12161 us/mul
2020-10-08 11:11:03 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 11:11:57 condorella/rx480 100342469 P2  7866075/10000000 ( 79%); 25133 muls, 12228 us/mul
2020-10-08 11:12:45 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 11:17:02 condorella/rx480 100342469 P2  8328075/10000000 ( 83%); 25060 muls, 12192 us/mul
2020-10-08 11:17:47 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 11:19:28 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 11:22:07 condorella/rx480 100342469 P2  8790075/10000000 ( 88%); 25078 muls, 12142 us/mul
2020-10-08 11:24:29 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 11:26:11 condorella/rx480 100342469 P2 GCD : no factor
2020-10-08 11:27:09 condorella/rx480 100342469 P2  9252075/10000000 ( 93%); 24943 muls, 12142 us/mul
2020-10-08 11:31:11 condorella/rx480 100342469 P2 Starting GCD
2020-10-08 11:32:12 condorella/rx480 100342469 P2  9714075/10000000 ( 97%); 24818 muls, 12208 us/mul
2020-10-08 11:32:53 condorella/rx480 100342469 P2 GCD : 55788405910316024926498537
2020-10-08 11:32:53 condorella/rx480 100342469 P2 Released memory lock 'memlock-0'
2020-10-08 11:32:57 condorella/rx480 100342469 GCD: 55788405910316024926498537
2020-10-08 11:32:57 condorella/rx480 100342469 {"status":"F", "exponent":"100342469", "worktype":"PM1", "B1":"12000", "fft-length":"5767168", "factors":["55788405910316024926498537"], "program":{"name":"gpuowl", "version":"v7.0-18-g69c2b85"}, "user":"kriesel", "computer":"condorella/rx480", "timesta
mp":"2020-10-08 16:32:57 UTC"}
2020-10-08 11:32:57 condorella/rx480 Exception NSt10filesystem7__cxx1116filesystem_errorE: filesystem error: cannot rename: Permission denied [worktodo.txt-tmp] [worktodo.txt]
2020-10-08 11:32:57 condorella/rx480 Bye
kriesel is online now   Reply With Quote
Old 2020-10-08, 18:16   #76
firejuggler
 
firejuggler's Avatar
 
Apr 2010
Over the rainbow

2·17·73 Posts
Default

Another good news, the early P2 GCD work. https://www.mersenne.ca/exponent/37425961
B1 set to 31k B2 set to 7M (while 80k would have been sufficient)

Code:
 

D:\Download\prime95\gpuowl\v7>gpuowl-win.exe -maxAlloc 4000 -B1 31000 -B2 7000000 -prp 37425961
2020-10-08 20:00:47 gpuowl v7.0-18-g69c2b85
2020-10-08 20:00:47 Note: not found 'config.txt'
2020-10-08 20:00:47 config: -maxAlloc 4000 -B1 31000 -B2 7000000 -prp 37425961
2020-10-08 20:00:47 device 0, unique id ''
2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961 FFT: 2M 1K:4:256 (17.85 bpw)
2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961 OpenCL args "-DEXP=37425961u -DWIDTH=1024u -DSMALL_HEIGHT=256u -DMIDDLE=4u -DCARRY64=1 -DCARRYM64=1 -DWEIGHT_STEP_MINUS_1=0xe.6912aa9d921f8p-7 -DIWEIGHT_STEP_MINUS_1=-0xc.f3c708d54b84p-7  -cl-unsafe-math-optimizations -cl-std=CL2.0 -cl-finite-math-only "
2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961

2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961 OpenCL compilation in 0.01 s
2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961 maxAlloc: 3.9 GB
2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961 Space for 485 B1 buffers (available mem 3882.7 MB, buf size 8.0 MB)
2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961 B1=31000 (44674 bits)
2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961 PRP starting from beginning
2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961 powerSmooth(31000), 44674 bits, took 0.00s (CPU)
2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961 Aquired memory lock 'memlock-0'
2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961 B1: allocating 480 buffers
2020-10-08 20:00:47 GeForce GTX 1660 Ti-0 37425961 Starting B1=31000, first bit 22
2020-10-08 20:00:50 GeForce GTX 1660 Ti-0 37425961 OK        0 loaded: blockSize 500, 0000000000000003
2020-10-08 20:00:50 GeForce GTX 1660 Ti-0 37425961 validating proof residues for power 8
2020-10-08 20:00:50 GeForce GTX 1660 Ti-0 37425961 Proof using power 8
2020-10-08 20:01:01 GeForce GTX 1660 Ti-0 37425961 B1 fold(480) (81 set) took 3.64s
2020-10-08 20:01:01 GeForce GTX 1660 Ti-0 37425961 OK     1000   0.00%; 6959 us/it; ETA 3d 00:21; dc6ced5ee551c3c1
2020-10-08 20:04:46 GeForce GTX 1660 Ti-0 37425961 B1 fold(480) (479 set) took 6.04s
2020-10-08 20:04:46 GeForce GTX 1660 Ti-0 37425961 B1 31000: releasing 480 buffers
2020-10-08 20:04:46 GeForce GTX 1660 Ti-0 37425961 Released memory lock 'memlock-0'
2020-10-08 20:04:46 GeForce GTX 1660 Ti-0 37425961 OK    45000   0.12%; 4961 us/it; ETA 2d 03:31; c072adee7a8d1d8f
2020-10-08 20:04:46 GeForce GTX 1660 Ti-0 37425961 B1 completed. Starting GCD
2020-10-08 20:04:46 GeForce GTX 1660 Ti-0 37425961 P2 (31000,7000000) will continue from B2=0
2020-10-08 20:04:46 GeForce GTX 1660 Ti-0 37425961 P2 B1=31000, B2=7000000, D=210: 473313 primes in [30975, 7000035], selected 403004 (85.1%) (70309 doubles + 332695 singles)
2020-10-08 20:04:46 GeForce GTX 1660 Ti-0 37425961 P2 B1=31000, B2=7000000, D=210 from B2=0 : 33186 blocks starting at 148
2020-10-08 20:04:46 GeForce GTX 1660 Ti-0 37425961 P2 Aquired memory lock 'memlock-0'
2020-10-08 20:04:46 GeForce GTX 1660 Ti-0 37425961 P2 Allocated 24 P2 buffers
2020-10-08 20:04:47 GeForce GTX 1660 Ti-0 37425961 P2 Setup 24 P2 buffers in 705.9 ms
2020-10-08 20:04:47 GeForce GTX 1660 Ti-0 37425961 P2    31185/7000000 (  0%); 16 muls, 4941 us/mul
2020-10-08 20:04:57 GeForce GTX 1660 Ti-0 37425961 P2 GCD : no factor
2020-10-08 20:07:22 GeForce GTX 1660 Ti-0 37425961 P2   493185/7000000 (  7%); 30917 muls, 5011 us/mul
2020-10-08 20:09:48 GeForce GTX 1660 Ti-0 37425961 P2   955185/7000000 ( 13%); 28765 muls, 5055 us/mul
2020-10-08 20:09:58 GeForce GTX 1660 Ti-0 37425961 P2 Starting GCD
2020-10-08 20:10:09 GeForce GTX 1660 Ti-0 37425961 P2 GCD : 24152172843149824433
2020-10-08 20:10:09 GeForce GTX 1660 Ti-0 37425961 P2 Released memory lock 'memlock-0'
2020-10-08 20:10:11 GeForce GTX 1660 Ti-0 37425961 GCD: 24152172843149824433
2020-10-08 20:10:11 GeForce GTX 1660 Ti-0 37425961 {"status":"F", "exponent":"37425961", "worktype":"PM1", "B1":"31000", "fft-length":"2097152", "factors":["24152172843149824433"], "program":{"name":"gpuowl", "version":"v7.0-18-g69c2b85"}, "computer":"GeForce GTX 1660 Ti-0", "timestamp":"2020-10-08 18:10:11 UTC"}
2020-10-08 20:10:11 GeForce GTX 1660 Ti-0 Bye
firejuggler is offline   Reply With Quote
Old 2020-10-08, 19:07   #77
preda
 
preda's Avatar
 
"Mihai Preda"
Apr 2015

24608 Posts
Default

Quote:
Originally Posted by kriesel View Post
2020-10-08 11:32:57 condorella/rx480 Exception NSt10filesystem7__cxx1116filesystem_errorE: filesystem error: cannot rename: Permission denied [worktodo.txt-tmp] [worktodo.txt]
Hi Ken, could you please check, is there a genuine permission issue preventing the file rename?

PS: there may have been an issue with the worktodo.txt rename affecting some windows. I commited an attempted fix (Ken to confirm).

Last fiddled with by preda on 2020-10-08 at 21:40
preda is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
GpuOwl PRP-Proof changes preda GpuOwl 20 2020-10-17 06:51
gpuowl: runtime error SELROC GpuOwl 59 2020-10-02 03:56
gpuOWL for Wagstaff GP2 GpuOwl 22 2020-06-13 16:57
gpuowl tuning M344587487 GpuOwl 14 2018-12-29 08:11
How to interface gpuOwl with PrimeNet preda PrimeNet 2 2017-10-07 21:32

All times are UTC. The time now is 15:21.

Tue Dec 1 15:21:17 UTC 2020 up 82 days, 12:32, 3 users, load averages: 1.81, 1.90, 1.85

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.