mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Factoring (https://www.mersenneforum.org/forumdisplay.php?f=19)
-   -   Faster GPU-ECM with CGBN (https://www.mersenneforum.org/showthread.php?t=27103)

wombatman 2022-04-03 17:07

[QUOTE=SethTro;603168]This is expected. I'm balancing binary size and compile time vs range of numbers that can be tested.

If you want to run ECM on numbers > 1020 bits look around line 670 in cgbn_stage1.cu[/QUOTE]

Good deal. Thanks!

EdH 2022-07-12 00:30

[QUOTE=SethTro;601241]IFAIK you don't need to compile CGBN (or setup the googletest), you just need the folder downloaded ("cloned") from Github.[/QUOTE]I've brought the folder down (via git) without installing CGBN for a second machine and although cgbn.h is present, ECM says it is not:[code]configure: Using CGBN from /home/math56/Math/CGBN/include/cgbn
checking if CGBN is present... no
configure: error: cgbn.h not found (check if /cgbn needed after <PATH>/include)[/code]I've tried with and without /cgbn after /include, per the readme, with no success. I've also tried copying cgbn.h into different places.[code]$ ls /home/math56/Math/CGBN/include/cgbn
arith cgbn.cu [B]cgbn.h[/B] core impl_mpz.cc
cgbn_cpu.h cgbn_cuda.h cgbn_mpz.h impl_cuda.cu[/code]

Rubiksmath 2022-11-13 03:26

Yes, actually I was wondering about being able to run GPU CGBN ECM on larger inputs, I've had a look at line ~670 in cgbn_stage1.cu where I do see stuff pertaining to increasing this limit. I tried using the pair (64,65536) on line 674, but that most definitely doesn't work, (probably isn't listed in the file for a reason). I've dropped that down to what is shown as being possible in the file (32,32768), and it works just fine. Is there any way for me to run ECM on CGBN for inputs 32768 bits < x < 65536 bits? it's not a huge deal if I can't, but I just want to make sure if there is a genuine limitation on that front or not.

frmky 2022-11-13 04:28

[QUOTE=Rubiksmath;617646] Is there any way for me to run ECM on CGBN for inputs 32768 bits < x < 65536 bits?[/QUOTE]
The mp library this uses, [URL="https://github.com/NVlabs/CGBN"]CGBN[/URL], only supports up to 32K bits.

Rubiksmath 2022-11-13 05:36

Yeah its fine, but I realised it's close to the crossover point for prime95 being faster on my setup anyways, but thanks.


All times are UTC. The time now is 18:06.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2022, Jelsoft Enterprises Ltd.