Topic: [ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX] - page 937. (Read 3426921 times)
February 06, 2014, 10:44:41 AM
:O
Is this legit?
What do you reckon for 2x670s?
February 06, 2014, 10:44:19 AM
Can some one help me with gt640 2gb?
I try with
cudaminer --algo=keccak -d gt640 -l K2x32 --benchmark
K2x32 is from autotune without --algo=keccak
When i put --algo=keccak and the miner says "try cudaminer --help"
I try with
cudaminer --algo=keccak -d gt640 -l K2x32 --benchmark
K2x32 is from autotune without --algo=keccak
When i put --algo=keccak and the miner says "try cudaminer --help"
February 06, 2014, 10:42:06 AM
oke cbuchner1 thanks for updating, do you have perhaps a good setting for my poor Evga GT610 2 GB card ? i cant get it higher then 7.5 khash/s !
I do not think this device can be much faster. 48 CUDA cores on what I believe to be a single Compute 2.1 multiprocessor.
As you see it's already putting out more heat than your fan (or passive cooling) can transport away.
yep correct i did know that nut perhaps could it be better , saving for other card i think , thanks m8
February 06, 2014, 10:31:07 AM
./cudaminer --algo=keccak -d gtx780 -L 16 -l K192x22 --benchmark
115Mhash/s
and thats why i posted the screenshot :-)
EDIT:
so i just ran tests from
/cudaminer --algo=keccak -d gtx780 -L 92 -l K192x32 --benchmark
down to
/cudaminer --algo=keccak -d gtx780 -L 92 -l K192x24 --benchmark
changing the x** number down 1 each time. anything higher then x24 gives me stupid numbers, 24 gives me 126Mhash/s, and i lost about 5Mhash/s for every x** i take off after that
115Mhash/s
and thats why i posted the screenshot :-)
EDIT:
so i just ran tests from
/cudaminer --algo=keccak -d gtx780 -L 92 -l K192x32 --benchmark
down to
/cudaminer --algo=keccak -d gtx780 -L 92 -l K192x24 --benchmark
changing the x** number down 1 each time. anything higher then x24 gives me stupid numbers, 24 gives me 126Mhash/s, and i lost about 5Mhash/s for every x** i take off after that
February 06, 2014, 10:30:40 AM
you have one line at 0 and the next 1.6G
yeah don't trust numbers that are so obviously suspicious.
I am not even sure if I can trust the 450 kHash/s that someone was reporting on Windows.
February 06, 2014, 10:29:43 AM
you have one line at 0 and the next 1.6G
February 06, 2014, 10:28:32 AM
using ./cudaminer --algo=keccak -d gtx780 -L 16 -l K192x32 --benchmark
i am still getting 4.6million khash/s
i am still getting 4.6million khash/s
try x16 x20 x21...
x22 might already exceed the register file's 65536 limit because 96 registers per thread x 22 warps x 32 threads/warp > 65536.
February 06, 2014, 10:26:41 AM
using ./cudaminer --algo=keccak -d gtx780 -L 16 -l K192x32 --benchmark
i am still getting 4.6million khash/s
that being said, every 3 attempts it does nothing,
so its saying
4.6million khash/s
4.6million khash/s
0 khash/s
i am still getting 4.6million khash/s
that being said, every 3 attempts it does nothing,
so its saying
4.6million khash/s
4.6million khash/s
0 khash/s
February 06, 2014, 10:25:44 AM
ok downloaded directly from github, compiled as normal. Results;
https://i.imgur.com/e5LY5wN.png
yehhhhh
im getting results as high as 67million khash/s
https://i.imgur.com/e5LY5wN.png
yehhhhh
im getting results as high as 67million khash/s
improbable. probably broken launch config.
February 06, 2014, 10:23:39 AM
Actually, nvidia cards are are already ahead for script-jane Nfactor=9 (or at least pretty close)
in term of Hashrate / price ? I don't think so...
February 06, 2014, 10:22:27 AM
ok downloaded directly from github, compiled as normal. Results;
https://i.imgur.com/e5LY5wN.png
yehhhhh
im getting results as high as 67million khash/s
using ./cudaminer --algo=keccak -d gtx780 -L 16 -l K192x32 --benchmark
i am still getting 4.6million khash/s
https://i.imgur.com/e5LY5wN.png
yehhhhh
im getting results as high as 67million khash/s
using ./cudaminer --algo=keccak -d gtx780 -L 16 -l K192x32 --benchmark
i am still getting 4.6million khash/s
February 06, 2014, 10:21:00 AM
I settled for maxrregcount=96 in keccak.cu. Using __launch_bounds__ directives I make sure the other keccak512 kernels (for scrypt-jane) remain at 64 registers so they keep working.
Maybe some guys will want to test the Linux based Makefile.am (run ./autogen.sh) and run it on ther GTX 780's, Titans, 780 Ti's and GT 640 with Compute 3.5 ? You can go up to x24 launch configurations with this register count on Compute 3.5 devices. And hopefully we'll get into the >400 kHash/s range
Maybe some guys will want to test the Linux based Makefile.am (run ./autogen.sh) and run it on ther GTX 780's, Titans, 780 Ti's and GT 640 with Compute 3.5 ? You can go up to x24 launch configurations with this register count on Compute 3.5 devices. And hopefully we'll get into the >400 kHash/s range
how do we test this christian? sorry i am a little slow today, compile as normal?
yes, right from github. Don't forget the ./autogen.sh
I haven't modified the Visual Studio 2010 solution's maxrregcount yet!
February 06, 2014, 10:20:16 AM
Me want! Christian I think something must be throttling back the 780 to the 780ti. There isnt that much performance difference and right now its over a 300Mhash/s difference
it's the maxrregcount option during compilation.
February 06, 2014, 10:19:58 AM
Ok guys I have a couple of gts 450 1gb, what cuddaminer should I use for maxcoin? the default download from page one or a special ed (link if possible)?
Could you advise me on abat file configuration?
I am on windows!
Could you advise me on abat file configuration?
I am on windows!
February 06, 2014, 10:19:17 AM
I settled for maxrregcount=96 in keccak.cu. Using __launch_bounds__ directives I make sure the other keccak512 kernels (for scrypt-jane) remain at 64 registers so they keep working.
Maybe some guys will want to test the Linux based Makefile.am (run ./autogen.sh) and run it on ther GTX 780's, Titans, 780 Ti's and GT 640 with Compute 3.5 ? You can go up to x24 launch configurations with this register count on Compute 3.5 devices. And hopefully we'll get into the >400 kHash/s range
Maybe some guys will want to test the Linux based Makefile.am (run ./autogen.sh) and run it on ther GTX 780's, Titans, 780 Ti's and GT 640 with Compute 3.5 ? You can go up to x24 launch configurations with this register count on Compute 3.5 devices. And hopefully we'll get into the >400 kHash/s range
how do we test this christian? sorry i am a little slow today, compile as normal?
February 06, 2014, 10:19:06 AM
I wonder, if it is really good for the coin all that...
I think it's good for all crypto if @ least one coin make nvidia cards owners join the game... And I say that as an ait farm owner...
February 06, 2014, 10:15:45 AM
ok here we go:
COuld you please share the download with the rest of us windows users? Thanks
+1
February 06, 2014, 10:15:40 AM
I wonder, if it is really good for the coin all that...
I think it's good for all crypto if @ least one coin make nvidia cards owners join the game... And I say that as an Ati farm owner...
February 06, 2014, 10:15:38 AM
this is what I get on my 660Ti under Linux.
Code:
buchner@test02:~/CudaMiner-master$ ./cudaminer --algo=keccak -L 64 -l K1024x32 --benchmark
*** CudaMiner for nVidia GPUs by Christian Buchner ***
This is version 2014-02-04 (beta)
based on pooler-cpuminer 2.3.2 (c) 2010 Jeff Garzik, 2012 pooler
Cuda additions Copyright 2013,2014 Christian Buchner
LTC donation address: LKS1WDKGED647msBQfLBHV3Ls8sveGncnm
BTC donation address: 16hJF5mceSojnTD3ZTUDqdRhDyPJzoRakM
YAC donation address: Y87sptDEcpLkLeAuex6qZioDbvy1qXZEj4
[2014-02-06 15:13:53] 1 miner threads started, using 'keccak' algorithm.
[2014-02-06 15:14:09] GPU #0: GeForce GTX 660 Ti with compute capability 3.0
[2014-02-06 15:14:09] GPU #0: interactive: 1, tex-cache: 0 , single-alloc: 0
[2014-02-06 15:14:09] GPU #0: 32 hashes / 0.1 MB per warp.
[2014-02-06 15:14:09] GPU #0: using launch configuration K1024x32
[2014-02-06 15:14:13] GPU #0: GeForce GTX 660 Ti, 98313 khash/s
[2014-02-06 15:14:13] Total: 98313 khash/s
[2014-02-06 15:14:18] GPU #0: GeForce GTX 660 Ti, 98940 khash/s
[2014-02-06 15:14:18] Total: 98940 khash/s
February 06, 2014, 10:12:58 AM
I wonder, if it is really good for the coin all that...
Jump to: