Topic: [ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX] - page 389. (Read 3426938 times)
June 03, 2014, 04:32:56 PM
Is cudaminer going to support blake multi mining? mine 5 coins at the same time!
June 03, 2014, 03:11:48 PM
MSI afterburner 3.0 is out http://www.guru3d.com/files_details/msi_afterburner_beta_download.html , anyone know how to unlock the core frequency with the 750ti like the older version?
Ahh nevermind, I've just noticed that msi updated nvidia driver to 337.50 witch have the limited core offset...
June 03, 2014, 03:09:09 PM
MSI afterburner 3.0 is out http://www.guru3d.com/files_details/msi_afterburner_beta_download.html , anyone know how to unlock the core frequency with the 750ti like the older version?
Ahh nevermind, I've just noticed that msi updated nvidia driver to 337.50 witch have the limited core offset...
June 03, 2014, 03:02:11 PM
MSI afterburner 3.0 is out http://www.guru3d.com/files_details/msi_afterburner_beta_download.html , anyone know how to unlock the core frequency with the 750ti like the older version?
June 03, 2014, 02:43:12 PM
CGwather new + 280x = 3,5 mx for x11
once more in English please?
But "CGWatcher, a GUI/monitor for CGMiner and BFGMiner" from http://manotechnology.blogspot.com/p/cgwatcher.html
June 03, 2014, 02:12:12 PM
CGwather new + 280x = 3,5 mx for x11
once more in English please?
June 03, 2014, 02:04:14 PM
Anyone knows if it's possible to compile blake for 2.1 compute? I downloaded 6th March commit from github that says
I excluded Titan and Kepler kernels from build, as I'm interested only in 2.1 compute (as I have nothing better ATM
) I edited source files to remove references to those kernels. I verified that all *.cu files are being compiled for compute_20,sm_21 (including solution itself), excluding nv_kernel.cu and nv_kernel2.cu, as compiling for that arch gives a lot of compiler errors.
Miner compiles and runs fine, but I'm getting only booo's Also is that hashrate normal or it's so big because miner does sth wrong? Any help appreciated, thanks.
Quote from: github
First working blake implementation in CUDA. Generic compute 1.0 kernel only - but already 50 MHash/s on GTX 750Ti. Optimization pending...
I excluded Titan and Kepler kernels from build, as I'm interested only in 2.1 compute (as I have nothing better ATM
) I edited source files to remove references to those kernels. I verified that all *.cu files are being compiled for compute_20,sm_21 (including solution itself), excluding nv_kernel.cu and nv_kernel2.cu, as compiling for that arch gives a lot of compiler errors.Miner compiles and runs fine, but I'm getting only booo's
Also is that hashrate normal or it's so big because miner does sth wrong? Any help appreciated, thanks.Code:
[2014-06-03 21:52:08] 1 miner threads started, using 'blake' algorithm.
[2014-06-03 21:52:08] Starting Stratum on stratum+tcp://eu3.blakecoin.com:3334
[2014-06-03 21:52:09] GPU #0: GeForce GT 620M with compute capability 2.1
[2014-06-03 21:52:09] GPU #0: interactive: 1, tex-cache: 0 , single-alloc: 0
[2014-06-03 21:52:09] GPU #0: 32 hashes / 4.0 MB per warp.
[2014-06-03 21:52:10] GPU #0: Performing auto-tuning (Patience...)
[2014-06-03 21:52:10] GPU #0: maximum total warps (BxW): 212
[2014-06-03 21:52:18] GPU #0: 1361454.55 hash/s with configuration F12x13
[2014-06-03 21:52:18] GPU #0: using launch configuration F12x13
[2014-06-03 21:52:20] GPU #0: GeForce GT 620M, 1245 khash/s
[2014-06-03 21:53:18] GPU #0: GeForce GT 620M, 1268 khash/s
[2014-06-03 21:54:19] GPU #0: GeForce GT 620M, 1259 khash/s
[2014-06-03 21:54:33] Stratum detected new block
[2014-06-03 21:54:33] GPU #0: GeForce GT 620M, 1254 khash/s
[2014-06-03 21:54:33] Stratum detected new block
[2014-06-03 21:54:33] GPU #0: GeForce GT 620M, 1178 khash/s
[2014-06-03 21:55:30] GPU #0: GeForce GT 620M, 1254 khash/s
[2014-06-03 21:56:10] GPU #0: GeForce GT 620M, 1252 khash/s
[2014-06-03 21:56:10] accepted: 0/1 (0.00%), 1252 khash/s (booooo)
[2014-06-03 21:57:10] GPU #0: GeForce GT 620M, 1256 khash/s
[2014-06-03 21:58:09] GPU #0: GeForce GT 620M, 1264 khash/s
[2014-06-03 21:59:10] GPU #0: GeForce GT 620M, 1258 khash/s
[2014-06-03 22:00:10] GPU #0: GeForce GT 620M, 1254 khash/s
[2014-06-03 22:01:09] GPU #0: GeForce GT 620M, 1263 khash/s
June 03, 2014, 01:59:50 PM
CGWatcher new + 280x = 3,5 mx for x11
June 03, 2014, 01:42:37 PM
Code:
while (!isDevsPublishedNewAlgo)
{
if isProgrammer()
contributeNewAlgo();
else
mineWithWhateverAlgoIsThere();
}
June 03, 2014, 01:41:10 PM
Code:
while (!isDevsPublishedNewAlgo)
{
if (isProgrammer())
contributeNewAlgo();
else
mineWithWhateverAlgoIsThere();
}
June 03, 2014, 12:48:25 PM
Do we have time frame on when we will have X13 for nvidia?
it's not that more profitable in comparison to x11 anyway
seems that amd whales are all there now...
June 03, 2014, 12:32:09 PM
Do we have time frame on when we will have X13 for nvidia?
Code:
while (!isDevsPublishedNewAlgo)
{
wait();
}
June 03, 2014, 12:30:08 PM
Do we have time frame on when we will have X13 for nvidia?
Code:
while (!isDevsPublishedNewAlgo)
{
wait();
}
June 03, 2014, 12:18:13 PM
Do we have time frame on when we will have X13 for nvidia?
June 03, 2014, 11:04:37 AM
Hi,
MRC and CACH use the same setting. I use --launch-config t5x24 -L 2 -b 4096 and got 9.5khs per card. Maybe you could squeeze out a little more (1khs max) if you have time to finetune but think that 280x only get 16-17 khs, this number is fair. From N=13 is really where nVidia shines. The distance with AMD is always shorter.
Some number for you to compare:
N12: 280x 56khs - 750ti 23khs
N13: 280x 17khs - 750ti 9.5khs
N14: 280x 4khs - 750ti 3khs
Regards,
MRC and CACH use the same setting. I use --launch-config t5x24 -L 2 -b 4096 and got 9.5khs per card. Maybe you could squeeze out a little more (1khs max) if you have time to finetune but think that 280x only get 16-17 khs, this number is fair. From N=13 is really where nVidia shines. The distance with AMD is always shorter.
Some number for you to compare:
N12: 280x 56khs - 750ti 23khs
N13: 280x 17khs - 750ti 9.5khs
N14: 280x 4khs - 750ti 3khs
Regards,
Many thanks AizenSou. I tried your settings, both with uppercase and lowercase T, as well as with various lookup gap sizes, no luck. Your settings as posted, produce the dreaded "result does not validate on CPU" message (or something along those lines), while variants got me crashing the driver right after launching.
As I check my performance over the last 24 hours, a 3x 750ti rig is averaging at about 27.7Kh/s with my settings, so I suppose very very close to what you're reaching as well. Here are my settings in case they help anyone else (and might be worth you checking them out also just in case):
Code:
-i 0 -L 4 -l T5x12
That's Win8 x64, latest official cudaminer release (28-02), nvidia drivers 337.88.
Happy Mining,
~ Myagui
June 03, 2014, 08:39:40 AM
I had 9 copies of Watch Dogs for buying the Evga Super Clock 750ti from Fry's.
That must be Fry's own offer then since I can't see it on any other retailer. Did you also get the 3 x in game credits as well?
If so, that's a nice big saving for 9 cards
I got Watch Dogs and DayLight. But Daylight was only going for around $10 on Ebay.
I believe it might be Fry's also. Since when I sell the code it does have FRYS as part of the code.
June 03, 2014, 08:37:26 AM
N12: 280x 56khs - 750ti 23khs
N13: 280x 17khs - 750ti 9.5khs
N14: 280x 4khs - 750ti 3khs
Regards,
...anybody tried N15 (yac) at the moment?
June 03, 2014, 08:28:54 AM
MRC/USD market is open
https://alcurex.com/index.php/crypto/index
The only thing which could rescue MRC right now. MRC is profitable for nVidia users since it becomes N-factor = 13 a few days ago.
https://alcurex.com/index.php/crypto/index
The only thing which could rescue MRC right now. MRC is profitable for nVidia users since it becomes N-factor = 13 a few days ago.
AizenSou, would you have good settings for the 750ti on MRC at nfactor 13? I've been asking here, but no luck yet.
The best I've managed so far is ~9Kh/s per card... (I'm not on MRC, but CACH, identical mining specs at the moment)
Thanks
~ Myagui
Hi,
MRC and CACH use the same setting. I use --launch-config t5x24 -L 2 -b 4096 and got 9.5khs per card. Maybe you could squeeze out a little more (1khs max) if you have time to finetune but think that 280x only get 16-17 khs, this number is fair. From N=13 is really where nVidia shines. The distance with AMD is always shorter.
Some number for you to compare:
N12: 280x 56khs - 750ti 23khs
N13: 280x 17khs - 750ti 9.5khs
N14: 280x 4khs - 750ti 3khs
Regards,
June 03, 2014, 07:56:35 AM
I had 9 copies of Watch Dogs for buying the Evga Super Clock 750ti from Fry's.
That must be Fry's own offer then since I can't see it on any other retailer. Did you also get the 3 x in game credits as well?
If so, that's a nice big saving for 9 cards
June 03, 2014, 07:49:13 AM
To top it off you can get up to $40 for the Watch Dog code also on ebay.
At least that is the game code I have been getting. the DayLight expired.
According to Nvidia, the 750 Ti and non-Ti version are not included in the free Watch_Doges (
) promotion...http://www.geforce.co.uk/whats-new/articles/watch-dogs-nvidia-geforce-gtx-bundle
I had 9 copies of Watch Dogs for buying the Evga Super Clock 750ti from Fry's.
Jump to: