Reply With QuoteMy CUDA GPU dnetc client stopped working!
This beta release expired on Jan 19 11:39:23 UTC.
Please download a newer beta, or run a standard-release client.
Is there a newer release that works with nVidia/CUDA ???
Where is the link to the download?
I'm running about 585 Mkeys/s (...GX2), sure is fun breaking the speed limit....
you are s.o.l like the rest of us.... :-(
![[NGS]Cpt00Kirk's Avatar](image.php?u=3045&dateline=1221115504 "[NGS]Cpt00Kirk's Avatar")
well guys ..... in the good old days we turned back the system clock a couple of days!! to run a program.
so i tried it with the beta and guess what? it works like a charm!!
because of the time difference we had the same problem 1 day earlier, so take back you clock and it will run. blocks are counted @ dnet so happy crunching untill the new one comes out
Dutch Power Cows at your service
ps.. my 8800GT is doing 315+ Mkeys/s
Dutch Power Cows
I'm getting a steady 637K on my GX2 with the latest beta.
I'm having fun!!! I'm just not sure if it's net fun or gross fun.
For DC projects the GX2 is faster then the GTX280 and costs about half. It does use more power but you that is to be expected. I have a board that will hold three but I don't have another two cards and a nuclear reactor to power it.
I'm having fun!!! I'm just not sure if it's net fun or gross fun.
GTX295
[Jan 28 06:04:11 UTC] RC5-72: using core #7 (CUDA 4-pipe 128-thd).
[Jan 28 06:04:22 UTC] RC5-72: Benchmark for core #7 (CUDA 4-pipe 128-thd)
0.00:00:08.64 [502,692,486 keys/sec]
I'll be watching your performance closely
http://teamstats.macnn.com/rc572/sta...sort=StatsWeek
Yea but that's only one core. My score is for both cores.Originally Posted by IronBits
GTX295
[Jan 28 06:04:11 UTC] RC5-72: using core #7 (CUDA 4-pipe 128-thd).
[Jan 28 06:04:22 UTC] RC5-72: Benchmark for core #7 (CUDA 4-pipe 128-thd)
0.00:00:08.64 [502,692,486 keys/sec]
I'll be watching your performance closely
http://teamstats.macnn.com/rc572/sta...sort=StatsWeek
I'm having fun!!! I'm just not sure if it's net fun or gross fun.
dnetc.exe -multiok
afaik
wbr, Me. Dead J. Dona \
I just ran it normally and it detected two GPU's and started two crunchers.
I'm having fun!!! I'm just not sure if it's net fun or gross fun.
Guru, you BIG PHAT Gorilla!
I just tried dual GPUs and got it working with two inexpensive cards. Pretty simple with the MultiOK=1 switch.
9800GT ... [247.29 Mkeys/s]
9600GSO . [114,428,713 keys/s]
I stopped dnetc Cuda, shutdown, pulled the 9600GSO, rebooted.
Put the 9800GT back on F@H where it gets ~4,000 PPD .
A fellow team member from XS has kindly informed me that any CUDA clients prior to 2.2 compatible that return WU's will not be accepted. So that means we have to use the current CUDA client, which is CRAP slow (about half speed).
http://n0cgi.distributed.net/cgi/dne...gi?user=bovine
Also it hints that maybe ALL the CUDA work prior to this was a waste???Dear friends,
We have discovered our nVidia CUDA clients prior to v2.9105.512 had a
problem that would cause RC5-72 results to skip part of the
block. This issue turned out to be caused by a bug in the CUDA
compiler itself, which was fixed beginning in the CUDA 2.2 SDK. Going
forward we will only be releasing clients for CUDA version 2.2 and
higher.
The fixed behavior unfortunately reveals that new CUDA clients will be
about half the speed of the older buggy CUDA versions. We understand
that the apparent speed decrease will seem disappointing, but it's
important to note the earlier speeds were not measuring useful
work. Going forward, speed comparisons should only be made with CUDA
2.2 or higher speeds, as these are the "correct" speeds. Also, please
remember the CUDA clients are still much faster than traditional CPU
clients.
If you are still running a CUDA beta client, we encourage you to
update to the current versions available on our pre-release page:
http://www.distributed.net/download/prerelease.php Results returned by
any earlier clients will no longer be accepted by our keymaster. Users
with prior stats credit from affected clients will not be
retroactively removed.
Due to aspects of our network communication protocol, we are not able
to remotely shutdown only the older, buggy, CUDA clients so we will be
implementing a method to send large, dummy blocks to older CUDA
clients instead.
Since all dnetc CUDA versions released so far have only been "beta"
clients with built-in expiration dates, the impact should be
contained. The last round of beta CUDA clients would have expired at
approximately the end of August 2009.
Thanks again to all of our beta testers that have been helping us
validate this exciting new technology.
Last edited by riptide; 08-07-2009 at 06:38 PM.
That was posted on
:: 27-Jul-2009 03:02 GMT (Monday) ::
Latest pre-release version
[x86/CUDA-2.2] v2.9105.512 (beta8) 2009-07-26
I'm using
dnetc v2.9105-512-GTL-09072609-*dev* for CUDA 2.2 on Win32 (WindowsNT 6.0).
nvcuda.dll Version: 8.15.11.9038
[Aug 08 01:05:27 UTC] *** This BETA release expires in 37.08:20:23.00. ***
Seems to be working fine at 158.56 Mkeys/s with a GTX 275
I'm seeing 148Mkeys on a 8800GTX @ 1729 shaders (was>350 before thisThat was posted on
:: 27-Jul-2009 03:02 GMT (Monday) ::
Latest pre-release version
[x86/CUDA-2.2] v2.9105.512 (beta8) 2009-07-26
I'm using
dnetc v2.9105-512-GTL-09072609-*dev* for CUDA 2.2 on Win32 (WindowsNT 6.0).
nvcuda.dll Version: 8.15.11.9038
[Aug 08 01:05:27 UTC] *** This BETA release expires in 37.08:20:23.00. ***
Seems to be working fine at 158.56 Mkeys/s with a GTX 275
What drivers? I'm on latest 190.38 / 2.3 CUDA SDK etc + Server 2008 R2 64bit.
Your 275 should be destroying them blocks even at our reduced rates now. Obviously the only other thing really is other apps swallowing your CPU or your GPU OR wrong core selected.
Last edited by riptide; 08-08-2009 at 12:06 AM.
Smoking now, thanks for the tip!
300,717,520 keys/sec
PHP Code:[Aug 08 04:14:40 UTC] RC5-72: using core #0 (CUDA 1-pipe 64-thd).
[Aug 08 04:14:57 UTC] RC5-72: Benchmark for core #0 (CUDA 1-pipe 64-thd)
0.00:00:14.76 [300,717,520 keys/sec]
[Aug 08 04:14:57 UTC] RC5-72: using core #1 (CUDA 1-pipe 128-thd).
[Aug 08 04:15:17 UTC] RC5-72: Benchmark for core #1 (CUDA 1-pipe 128-thd)
0.00:00:16.32 [268,265,609 keys/sec]
[Aug 08 04:15:17 UTC] RC5-72: using core #2 (CUDA 1-pipe 256-thd).
[Aug 08 04:15:36 UTC] RC5-72: Benchmark for core #2 (CUDA 1-pipe 256-thd)
0.00:00:17.32 [181,761,737 keys/sec]
[Aug 08 04:15:36 UTC] RC5-72: using core #3 (CUDA 2-pipe 64-thd).
[Aug 08 04:15:56 UTC] RC5-72: Benchmark for core #3 (CUDA 2-pipe 64-thd)
0.00:00:16.32 [268,265,609 keys/sec]
[Aug 08 04:15:56 UTC] RC5-72: using core #4 (CUDA 2-pipe 128-thd).
[Aug 08 04:16:16 UTC] RC5-72: Benchmark for core #4 (CUDA 2-pipe 128-thd)
0.00:00:17.32 [181,914,864 keys/sec]
[Aug 08 04:16:16 UTC] RC5-72: using core #5 (CUDA 2-pipe 256-thd).
[Aug 08 04:16:36 UTC] RC5-72: Benchmark for core #5 (CUDA 2-pipe 256-thd)
0.00:00:17.01 [163,986,486 keys/sec]
[Aug 08 04:16:36 UTC] RC5-72: using core #6 (CUDA 4-pipe 64-thd).
[Aug 08 04:16:56 UTC] RC5-72: Benchmark for core #6 (CUDA 4-pipe 64-thd)
0.00:00:17.32 [181,761,737 keys/sec]
[Aug 08 04:16:56 UTC] RC5-72: using core #7 (CUDA 4-pipe 128-thd).
[Aug 08 04:17:16 UTC] RC5-72: Benchmark for core #7 (CUDA 4-pipe 128-thd)
0.00:00:17.25 [177,540,098 keys/sec]
[Aug 08 04:17:16 UTC] RC5-72: using core #8 (CUDA 4-pipe 256-thd).
[Aug 08 04:17:37 UTC] RC5-72: Benchmark for core #8 (CUDA 4-pipe 256-thd)
0.00:00:17.03 [179,131,744 keys/sec]
[Aug 08 04:17:37 UTC] RC5-72: using core #9 (CUDA 1-pipe 64-thd busy wait).
[Aug 08 04:17:55 UTC] RC5-72: Benchmark for core #9 (CUDA 1-pipe 64-thd busy wait)
0.00:00:14.71 [297,742,505 keys/sec]
[Aug 08 04:17:55 UTC] RC5-72: using core #10 (CUDA 1-pipe 64-thd sleep 100us).
[Aug 08 04:18:13 UTC] RC5-72: Benchmark for core #10 (CUDA 1-pipe 64-thd sleep 100us)
0.00:00:16.36 [267,927,345 keys/sec]
[Aug 08 04:18:13 UTC] RC5-72: using core #11 (CUDA 1-pipe 64-thd sleep dynamic).
[Aug 08 04:18:32 UTC] RC5-72: Benchmark for core #11 (CUDA 1-pipe 64-thd sleep dynamic)
0.00:00:16.45 [269,409,506 keys/sec]
[Aug 08 04:18:32 UTC] RC5-72 benchmark summary :
Default core : #0 (CUDA 1-pipe 64-thd)
Fastest core : #0 (CUDA 1-pipe 64-thd)
Ha! Classic mistake that was. Ssssh. I won't tell anyone.Smoking now, thanks for the tip!
300,717,520 keys/sec
PHP Code:[Aug 08 04:14:40 UTC] RC5-72: using core #0 (CUDA 1-pipe 64-thd).
[Aug 08 04:14:57 UTC] RC5-72: Benchmark for core #0 (CUDA 1-pipe 64-thd)
0.00:00:14.76 [300,717,520 keys/sec]
[Aug 08 04:14:57 UTC] RC5-72: using core #1 (CUDA 1-pipe 128-thd).
[Aug 08 04:15:17 UTC] RC5-72: Benchmark for core #1 (CUDA 1-pipe 128-thd)
0.00:00:16.32 [268,265,609 keys/sec]
[Aug 08 04:15:17 UTC] RC5-72: using core #2 (CUDA 1-pipe 256-thd).
[Aug 08 04:15:36 UTC] RC5-72: Benchmark for core #2 (CUDA 1-pipe 256-thd)
0.00:00:17.32 [181,761,737 keys/sec]
[Aug 08 04:15:36 UTC] RC5-72: using core #3 (CUDA 2-pipe 64-thd).
[Aug 08 04:15:56 UTC] RC5-72: Benchmark for core #3 (CUDA 2-pipe 64-thd)
0.00:00:16.32 [268,265,609 keys/sec]
[Aug 08 04:15:56 UTC] RC5-72: using core #4 (CUDA 2-pipe 128-thd).
[Aug 08 04:16:16 UTC] RC5-72: Benchmark for core #4 (CUDA 2-pipe 128-thd)
0.00:00:17.32 [181,914,864 keys/sec]
[Aug 08 04:16:16 UTC] RC5-72: using core #5 (CUDA 2-pipe 256-thd).
[Aug 08 04:16:36 UTC] RC5-72: Benchmark for core #5 (CUDA 2-pipe 256-thd)
0.00:00:17.01 [163,986,486 keys/sec]
[Aug 08 04:16:36 UTC] RC5-72: using core #6 (CUDA 4-pipe 64-thd).
[Aug 08 04:16:56 UTC] RC5-72: Benchmark for core #6 (CUDA 4-pipe 64-thd)
0.00:00:17.32 [181,761,737 keys/sec]
[Aug 08 04:16:56 UTC] RC5-72: using core #7 (CUDA 4-pipe 128-thd).
[Aug 08 04:17:16 UTC] RC5-72: Benchmark for core #7 (CUDA 4-pipe 128-thd)
0.00:00:17.25 [177,540,098 keys/sec]
[Aug 08 04:17:16 UTC] RC5-72: using core #8 (CUDA 4-pipe 256-thd).
[Aug 08 04:17:37 UTC] RC5-72: Benchmark for core #8 (CUDA 4-pipe 256-thd)
0.00:00:17.03 [179,131,744 keys/sec]
[Aug 08 04:17:37 UTC] RC5-72: using core #9 (CUDA 1-pipe 64-thd busy wait).
[Aug 08 04:17:55 UTC] RC5-72: Benchmark for core #9 (CUDA 1-pipe 64-thd busy wait)
0.00:00:14.71 [297,742,505 keys/sec]
[Aug 08 04:17:55 UTC] RC5-72: using core #10 (CUDA 1-pipe 64-thd sleep 100us).
[Aug 08 04:18:13 UTC] RC5-72: Benchmark for core #10 (CUDA 1-pipe 64-thd sleep 100us)
0.00:00:16.36 [267,927,345 keys/sec]
[Aug 08 04:18:13 UTC] RC5-72: using core #11 (CUDA 1-pipe 64-thd sleep dynamic).
[Aug 08 04:18:32 UTC] RC5-72: Benchmark for core #11 (CUDA 1-pipe 64-thd sleep dynamic)
0.00:00:16.45 [269,409,506 keys/sec]
[Aug 08 04:18:32 UTC] RC5-72 benchmark summary :
Default core : #0 (CUDA 1-pipe 64-thd)
Fastest core : #0 (CUDA 1-pipe 64-thd)
Yeah, that's the conclusion I came to after reading that update. What a shame.Also it hints that maybe ALL the CUDA work prior to this was a waste???
yup, pretty sad for the project really as there has been quite a bit of work done with the cuda clients. :-( Luckily they caught it, and luckily they brand the results so they can go back and remove all the bogus stuff from the results database. Really a shame though.Yeah, that's the conclusion I came to after reading that update. What a shame.
So has anyone done any RC5 crunching with one of the Nvida Quadro cards? If not, does anyone have one that they wouldn't mind running some benchmarks on it? Just really curious how they stack up benchmark wise with the GeForce cards.
