Topic: RPC Miners (CPU/4way/CUDA/OpenCL) - page 4. (Read 548504 times)
Yeah... 11Gh/s from an Nvidia card... actually, from any card... is unheard of. You're likely seeing a bug. That's a little less than 10 times what a dual 5970 setup gets.
Only time I got numbers like that on the CUDA miner was when my graphics drivers crashed.
Hi all, I've just started mining and am getting what would appear to be some strange numbers given what others are posting, I just want to validate that it's not some bug or something is wrong and that it's actually working as desired.
A dump from the command line:
>rpcminer-cuda.
exe -user=xxxx -password=xxxx -gpu -aggression=14
Client will start 1 miner threads
Work will be refreshed every 4000 ms
1 CUDA GPU devices found
Setting CUDA device to first device found
Target = xxxx
Loading module bitcoinminercuda_11.cubin
CUDA initialized
Done allocating CUDA resources for (16,16)
Finding best configuration step end (16,16) 1813ms prev best=9223372036854775807ms
Done allocating CUDA resources for (16,32)
Finding best configuration step end (16,32) 914ms prev best=1813ms
Done allocating CUDA resources for (16,64)
Finding best configuration step end (16,64) 490ms prev best=914ms
Done allocating CUDA resources for (16,128)
Finding best configuration step end (16,128) 335ms prev best=490ms
Done allocating CUDA resources for (16,256)
Finding best configuration step end (16,256) 331ms prev best=335ms
Done allocating CUDA resources for (32,16)
Finding best configuration step end (32,16) 956ms prev best=331ms
Done allocating CUDA resources for (32,32)
Finding best configuration step end (32,32) 490ms prev best=331ms
Done allocating CUDA resources for (32,64)
Finding best configuration step end (32,64) 334ms prev best=331ms
Done allocating CUDA resources for (32,128)
Finding best configuration step end (32,128) 331ms prev best=331ms
Done allocating CUDA resources for (32,256)
Finding best configuration step end (32,256) 330ms prev best=331ms
Done allocating CUDA resources for (64,16)
Finding best configuration step end (64,16) 939ms prev best=330ms
Done allocating CUDA resources for (64,32)
Finding best configuration step end (64,32) 483ms prev best=330ms
Done allocating CUDA resources for (64,64)
Finding best configuration step end (64,64) 332ms prev best=330ms
Done allocating CUDA resources for (64,128)
Finding best configuration step end (64,128) 329ms prev best=330ms
Done allocating CUDA resources for (64,256)
Finding best configuration step end (64,256) 328ms prev best=329ms
Done allocating CUDA resources for (128,16)
Finding best configuration step end (128,16) 933ms prev best=328ms
Done allocating CUDA resources for (128,32)
Finding best configuration step end (128,32) 483ms prev best=328ms
Done allocating CUDA resources for (128,64)
Finding best configuration step end (128,64) 329ms prev best=328ms
Done allocating CUDA resources for (128,128)
Finding best configuration step end (128,128) 329ms prev best=328ms
Done allocating CUDA resources for (128,256)
Finding best configuration step end (128,256) 327ms prev best=328ms
Done allocating CUDA resources for (128,256)
6759605144 khash/s
11503132515 khash/s
12118605140 khash/s
11352202147 khash/s
10592490938 khash/s
11374405002 khash/s
12092100194 khash/s
10122935165 khash/s
12085704851 khash/s
10960440978 khash/s
12213746319 khash/s
12179113316 khash/s
11705168381 khash/s
12204549939 khash/s
12103577788 khash/s
10324611659 khash/s
10190482336 khash/s
11422309222 khash/s
I've got an Nvidia Quadro FX 1800, so figured it was safe to try the high aggression. Prior to adding the aggression I was only getting avg. 12000 khash/s, but now I'm getting those ridiculously high numbers.
It's only pegging 1 CPU core too, so I'm averaging 12% load there and the system is as responsive as if I were doing nothing.
Is this all for real? Are those numbers real?
And if it is for real, how long would it take to calculate a block?
A dump from the command line:
Code:
>rpcminer-cuda.
exe -user=xxxx -password=xxxx -gpu -aggression=14
Client will start 1 miner threads
Work will be refreshed every 4000 ms
1 CUDA GPU devices found
Setting CUDA device to first device found
Target = xxxx
Loading module bitcoinminercuda_11.cubin
CUDA initialized
Done allocating CUDA resources for (16,16)
Finding best configuration step end (16,16) 1813ms prev best=9223372036854775807ms
Done allocating CUDA resources for (16,32)
Finding best configuration step end (16,32) 914ms prev best=1813ms
Done allocating CUDA resources for (16,64)
Finding best configuration step end (16,64) 490ms prev best=914ms
Done allocating CUDA resources for (16,128)
Finding best configuration step end (16,128) 335ms prev best=490ms
Done allocating CUDA resources for (16,256)
Finding best configuration step end (16,256) 331ms prev best=335ms
Done allocating CUDA resources for (32,16)
Finding best configuration step end (32,16) 956ms prev best=331ms
Done allocating CUDA resources for (32,32)
Finding best configuration step end (32,32) 490ms prev best=331ms
Done allocating CUDA resources for (32,64)
Finding best configuration step end (32,64) 334ms prev best=331ms
Done allocating CUDA resources for (32,128)
Finding best configuration step end (32,128) 331ms prev best=331ms
Done allocating CUDA resources for (32,256)
Finding best configuration step end (32,256) 330ms prev best=331ms
Done allocating CUDA resources for (64,16)
Finding best configuration step end (64,16) 939ms prev best=330ms
Done allocating CUDA resources for (64,32)
Finding best configuration step end (64,32) 483ms prev best=330ms
Done allocating CUDA resources for (64,64)
Finding best configuration step end (64,64) 332ms prev best=330ms
Done allocating CUDA resources for (64,128)
Finding best configuration step end (64,128) 329ms prev best=330ms
Done allocating CUDA resources for (64,256)
Finding best configuration step end (64,256) 328ms prev best=329ms
Done allocating CUDA resources for (128,16)
Finding best configuration step end (128,16) 933ms prev best=328ms
Done allocating CUDA resources for (128,32)
Finding best configuration step end (128,32) 483ms prev best=328ms
Done allocating CUDA resources for (128,64)
Finding best configuration step end (128,64) 329ms prev best=328ms
Done allocating CUDA resources for (128,128)
Finding best configuration step end (128,128) 329ms prev best=328ms
Done allocating CUDA resources for (128,256)
Finding best configuration step end (128,256) 327ms prev best=328ms
Done allocating CUDA resources for (128,256)
6759605144 khash/s
11503132515 khash/s
12118605140 khash/s
11352202147 khash/s
10592490938 khash/s
11374405002 khash/s
12092100194 khash/s
10122935165 khash/s
12085704851 khash/s
10960440978 khash/s
12213746319 khash/s
12179113316 khash/s
11705168381 khash/s
12204549939 khash/s
12103577788 khash/s
10324611659 khash/s
10190482336 khash/s
11422309222 khash/s
I've got an Nvidia Quadro FX 1800, so figured it was safe to try the high aggression. Prior to adding the aggression I was only getting avg. 12000 khash/s, but now I'm getting those ridiculously high numbers.
It's only pegging 1 CPU core too, so I'm averaging 12% load there and the system is as responsive as if I were doing nothing.
Is this all for real? Are those numbers real?
And if it is for real, how long would it take to calculate a block?
I compiled a binary of rpcminer for Mac OS 10.6 ‘Snow Leopard’. This binary distribution is built from the source code distribution from 2011-02-27. There are three necessary platform modifications to the original source to make it compile and run on Mac OS.
Download rpcminer for Mac [binary] (1.244 MiB) at MegaUpload.
I hope someone will find it useful.
Donations: 14V7RiwD3joqDF2Jv6zst3ZnQsR686u9iH 
It’s been tested on a MacBook Air running Mac OS 10.6.5 and a Mac Mini running 10.6.7 against the mining.bitcoin.cz pool.
Download rpcminer for Mac [binary] (1.244 MiB) at MegaUpload.
I hope someone will find it useful.
Donations: 14V7RiwD3joqDF2Jv6zst3ZnQsR686u9iH It’s been tested on a MacBook Air running Mac OS 10.6.5 and a Mac Mini running 10.6.7 against the mining.bitcoin.cz pool.
Thanks for taking the time to compile a binary for other OSx users. I've tried it on 10.6.6 but it wouldn't run, "cannot execute binary file". Are there any prerequisites I might be missing or could I screw up something otherwise?
You and I have the exact same problem.
I compiled a binary of rpcminer for Mac OS 10.6 ‘Snow Leopard’. This binary distribution is built from the source code distribution from 2011-02-27. There are three necessary platform modifications to the original source to make it compile and run on Mac OS.
Download rpcminer for Mac [binary] (1.244 MiB) at MegaUpload.
Update: easy-run variant
I hope someone will find it useful. Donations: 14V7RiwD3joqDF2Jv6zst3ZnQsR686u9iH
It’s been tested on a MacBook Air running Mac OS 10.6.5 and a Mac Mini running 10.6.7 against the mining.bitcoin.cz pool.
Update: easy-run variant
I hope someone will find it useful.
Donations: 14V7RiwD3joqDF2Jv6zst3ZnQsR686u9iH It’s been tested on a MacBook Air running Mac OS 10.6.5 and a Mac Mini running 10.6.7 against the mining.bitcoin.cz pool.
Here's some RPC miners for all those mining via RPC. There's a CPU miner based on the Crypto++ code, a 4way miner using the 4way code included with Bitcoin, and a CUDA and OpenCL miner. The included readme.txt has instructions on how to supply arguments to the miners. Supplying a -help argument to the executable will show you the available options as well.
Want these miners as a screensaver? See this thread.
Downloads (Updated 2011-02-27)
RPC Miner Windows Binaries
RPC Miner Source
Use the CUDA miner for NVidia cards.
Use the OpenCL miner for ATI cards.
Use CMake to compile from the source.
Tweaking command line parameters for the CUDA and OpenCL miners will be necessary if you want to get optimum performance.
Want these miners as a screensaver? See this thread.
Downloads (Updated 2011-02-27)
RPC Miner Windows Binaries
RPC Miner Source
Use the CUDA miner for NVidia cards.
Use the OpenCL miner for ATI cards.
Use CMake to compile from the source.
Tweaking command line parameters for the CUDA and OpenCL miners will be necessary if you want to get optimum performance.
anyway to support sse2 on ubuntu ?
9100 is not CUDA or OPENCL supported. So u can't use it to mine.
http://en.wikipedia.org/wiki/CUDA
http://www.nvidia.com/object/cuda_gpus.html
Only 9100M supports & thats for notebook not desktop pc.
http://en.wikipedia.org/wiki/CUDA
http://www.nvidia.com/object/cuda_gpus.html
Only 9100M supports & thats for notebook not desktop pc.
Hello, I'm new at this.
I want to use the CPU, since my computer is shit.
However, when I run the CPU one, I get the following:
Clearly, I know I need some stuff to do pooled mining, but I haven't the foggiest clue what to do. I'm not a very technical person, I'm afraid. If someone could baby me through this, I'd be very grateful.
I want to use the CPU, since my computer is shit.
However, when I run the CPU one, I get the following:
Quote
Client will start 1 miner threads
Work will be refreshed every 4000 ms
Could not retrieve work from RPC server.
CURL return value = 7
Could not retrieve work from RPC server.
CURL return value = 7
Work will be refreshed every 4000 ms
Could not retrieve work from RPC server.
CURL return value = 7
Could not retrieve work from RPC server.
CURL return value = 7
Clearly, I know I need some stuff to do pooled mining, but I haven't the foggiest clue what to do. I'm not a very technical person, I'm afraid. If someone could baby me through this, I'd be very grateful.
What are the performance arguements with this miner? I loaded up the rpcminer-opencl on Radeon 6870's and its only giving 200Mh/s where as poclbm did ~260
Also, how do i dictate which GPU the device runs on so I can run multiple instantces. Right now it only runs on the first GPU out of four
Also, how do i dictate which GPU the device runs on so I can run multiple instantces. Right now it only runs on the first GPU out of four
What's wrong is you have an 8600GS. Sorry, you aren't going to get much out of that.
Ok, i have some more problems with an NVIDIA card (which i'm not physically at atm). At first, all the miners I tried (phoenix, poclbm-mod, rpcminer-cuda) would crash with something like "Import Error - LoadMemoryBlah - failed to import /_pyocl.py", basically OpenCL and CUDA didn't work.
Once I updated the NVIDIA drivers from nvidia.com (auto-detect function), the miners ran.
The miners detected
#1 - The CUDA platform
#2 - One cuda device. The NVIDIA 8600 GS card
I tried different settings with these 3 miners, but never got the speed above 1.5 Mhash ?!
What the heck is wrong?
Finally, in an act of desperation, i tried the rpcminer-cpu using 4 threads on the "Intel Core2 Quad" or something like that and got speeds of 4.4 Mhash.
How the heck can the CPU be faster than the GPU? Something must be seriously wrong, right?
Once I updated the NVIDIA drivers from nvidia.com (auto-detect function), the miners ran.
The miners detected
#1 - The CUDA platform
#2 - One cuda device. The NVIDIA 8600 GS card
I tried different settings with these 3 miners, but never got the speed above 1.5 Mhash ?!
What the heck is wrong?
Finally, in an act of desperation, i tried the rpcminer-cpu using 4 threads on the "Intel Core2 Quad" or something like that and got speeds of 4.4 Mhash.
How the heck can the CPU be faster than the GPU? Something must be seriously wrong, right?
Edit: Nevermind! I used incorrect arguments.
I use these settings for maximum speed -aggression=8 -gpugrid=256 -gputhreads=512
Unfortunately we will not see optimized CUDA miner as OpenCL
Unfortunately we will not see optimized CUDA miner as OpenCL
GTX460@800MHz
rpcminer-cuda 70Mhash/s
phoenix or poclbm gives 59Mhash/s
rpcminer-cuda 70Mhash/s
phoenix or poclbm gives 59Mhash/s
What settings do you use? CUDA is much worse on both Quadro FX 5800 and Geforce 9800GT for me. On the Quadro FX 5800 (still a rather high end card) I only get 50 Mh! - surely that's not right?
GTX460@800MHz
rpcminer-cuda 70Mhash/s
phoenix or poclbm gives 59Mhash/s
rpcminer-cuda 70Mhash/s
phoenix or poclbm gives 59Mhash/s
rpcminer-cuda.exe is for Nvidia cards
poclbm is for opencl, AMD cards.
GTX460@800MHz
rpcminer-cuda 70Mhash/s
phoenix or poclbm gives 59Mhash/s
rpcminer-cuda 70Mhash/s
phoenix or poclbm gives 59Mhash/s
Man its not even work using CUDA... I'm getting 30x more on an AMD Radeon GPU using the same mount of power on OpenCL
I have to say the same. With my Nvidia 9800GT, poclbm OpenCL maxes at 25 MH (30 if I use -f 0), whereas rpcminer-cuda gives 18 MH if I'm lucky - and I've played with the settings a lot.
[oh and rpcminer-opencl doesn't work for me at all. It never actually connects after initialising. In windows task manager I just see rpc-opencl use more and more RAM, until eventually it crashes(!)]
This to me doesn't make sense - CUDA is more low-level and optimized for Nvidia, is it not? Surely it should run faster than OpenCL?
P.S. hello all - first time poster, long time lurker
Need to follow this thread for update news. I was a month behind on the other clients.
Jump to: