20201213, 04:31  #188 
"Mihai Preda"
Apr 2015
2×677 Posts 
I'm personally not using 6.x myself, thus I don't have a lot of motivation to improve it. From my POV, 7.x is now better in a couple of ways than 6.x, and I prefer to focus my (limited) resources on 7.x. OTOH in the open spirit I'm not against others having different oppinions, and using different versions in different ways etc.

20201213, 21:45  #189 
Jul 2003
So Cal
2,083 Posts 
P1 factors seem to always be reported as found by stage 1. This leads to confusing reports like https://www.mersenne.ca/exponent/104592713. Can this be fixed easily or is this behavior intentional?

20201219, 15:00  #191  
Random Account
Aug 2009
19·101 Posts 
Quote:


20201219, 21:14  #192 
"Oliver"
Sep 2017
Porta Westfalica, DE
2^{2}·3^{2}·13 Posts 
Yes, that would exclude everybody who is doing P1 for exponent with known residues.
Last fiddled with by kruoli on 20201219 at 21:15 Reason: Conjunctive. 
20201220, 18:56  #193 
Random Account
Aug 2009
19·101 Posts 
I run firsttime P1's from Primenet by manual reservation. The only thing done to them prior is TF to a specific bit level, Currently, it is 2^76. I can also get these from GPUto72's website. A P1 with a "known residue" would indicate to me that is has a factor, meaning it is composite and requires no further effort.

20210117, 05:21  #194 
May 2020
11001_{2} Posts 
Does anyone have any benchmarks running gpuowl on the new gtx 30 series or the new AMD 6800x series gpus?

20210117, 05:43  #195 
1976 Toyota Corona years forever!
"Wayne"
Nov 2006
Saskatchewan, Canada
43·107 Posts 

20210117, 12:52  #196 
"Viliam FurÃk"
Jul 2018
Martin, Slovakia
446_{10} Posts 

20210307, 19:14  #197 
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest
47×107 Posts 
Gpuowl speed regression
Among the v7.2 builds I've made and run on Win 10, 53 is fastest. But for straight PRP, v6.11364 or 380 are faster still. There are some caveats to that result summary; tested on one system & one OS, with multiple fast gpus supported by a slow cpu. But some cases were substantial enough to more than erase the ostensible advantage of doing PRP & P1 stage 1 simultaneously with many of the same squarings, compared to running separate V6.11 P1 to normal bounds followed by V6.11 PRP. That's comparing time per iteration after initialization time, bringing gpu up to temperature for stable ~steady state iteration times, etc. Additional initialization in V7.2 puts it at an additional disadvantage (trig table setup for example).
In all V7.2 tests mentioned here, PRP iterations were comparable to V6.11 iterations; PRP was continuing in iterations after P1 had already completed, or P1 was suppressed by B1=0,B2=0; in the worktodo line or testssaved=0 in the worktodo line for V7.2x timing tests. Part one PRP iteration speed regression for gpuowlwin v7.2x on Radeon VII, Win 10 Pro x64 Variation ~37% 623M 7312/7087 = 1.032 927M 12263/11458 = 1.070 Code:
gpuowl us/it us/it version 623M 927M V7.221 7119 11960 V7.269 7312 12257 V7.263 7311 12263 V7.253 x x system restart after these attempts v7.253 7088 11459 V7.239 fatal cl compile errors, so no timings possible V7.221 7106 11955 * good repeatability after restart and other versions v7.213 7141 12365 v7.253 7087 11458 * fastest and very good repeatability of timing, saves ~8 days on 927M fft size 36M 52M split 4k:9:512 4K:13:512 gpu # 2 3 This multigpu slowcpu system sometimes exhibits loss of communication with a gpu, which can bog down other gpus also while system interrupts occupy 1 of the 2 cpu cores for ~1 hour or until a restart clears the condition. Note, gpus on this system are configured to run at reduced electrical power. Part two 100M, 300M, 900M, v6.11 & 7.253 PRP iteration speed regression for gpuowlwin v6.11x on Radeon VII, Win 10 Pro x64 Variation 1.2  5.2% 100M 869/826 = 1.052, v6.11380 fastest, > ~1.025 cost of usualbounds (P1+PRP) / PRP 300M 3360/3310 = 1.015, v6.11364 fastest, > ~1.01 cost of usualbounds (P1 stage 1 + PRP) / PRP 900M 10895/10763 = 1.012, V6.11364 380 tie for fastest, ~1.01 cost of usualbounds (P1 stage 1 + PRP) / PRP Differences between v6.11364 and v6.11380 were +/ 1 count, not significant. Code:
gpuowl us/it us/it us/it version 300M 103M 900M 6.11380 3311 826 10763 6.11364 3310 827 10763 7.253 3360 869 10895 6.11318 3327 836 10815 fft size 16M 5.5M 52M split 1k:8:1k 1K:11:256 4k:13:512 gpu # 4 4 4 Note, gpus on this system are configured to run at reduced electrical power. Last fiddled with by kriesel on 20210307 at 20:10 
20210307, 20:06  #198  
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest
47×107 Posts 
Quote:
Code:
iters <N> : run next PRP test for <N> iterations and exit. Multiple of 10000. It may be faster though to stick with v6.11380 for P1, if the system is reliable enough and the exponent P1 quick enough not to need the better error detection in P1 of V7.2. Last fiddled with by kriesel on 20210307 at 20:32 

Thread Tools  
Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
GpuOwl PRPProof changes  preda  GpuOwl  20  20201017 06:51 
gpuowl: runtime error  SELROC  GpuOwl  59  20201002 03:56 
gpuOWL for Wagstaff  GP2  GpuOwl  22  20200613 16:57 
gpuowl tuning  M344587487  GpuOwl  14  20181229 08:11 
How to interface gpuOwl with PrimeNet  preda  PrimeNet  2  20171007 21:32 