[ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX]

[tachyon_john]

I'm sure there must be various ways to beat screen into submission and get it to do what people want, but on any machine where there are multiple CUDA toolkit versions laying around (I have about 6) it's best to have binaries compiled such that they don't depend on environment variables. 

Without dealing with this at linkage time, as soon as you have different binaries compiled against different CUDA toolkit versions, environment variables either have to be embedded in per-app startup scripts to ensure they pick up the right library, or you end up having various problems.  I think that starting with CUDA 5.5 and going forward, the easy way out is simply to use libcudart_static.a and avoid this problem altogether, but I realize that not everyone has an up-to-date CUDA toolkit installed, and you probably want make the cudaminer build work out-of-the-box on the principle of least user emails :-)

Cheers,
  John

Can't screen run a shell script that sets LD_LIBRARY_PATH so libcudart before starting cudaminer? Somehow I don't see how this incompatibility is a a real issue with cudaminer itself...

[1713986694]

1713986694

[1713986694]

1713986694

[_Wolverine_]

Can't screen run a shell script that sets LD_LIBRARY_PATH so libcudart before starting cudaminer? Somehow I don't see how this incompatibility is a a real issue with cudaminer itself...

Don't have to

this is how I do

Code:

cat miner.sh
LD_LIBRARY_PATH=/usr/local/cuda/lib64 ./CudaMiner/cudaminer.......

cat screen_miner.sh
screen -dmS cudaminer miner.sh

[cvax]

Just got a 9th GTX 750 Ti. Decided to try a different brand this time and I regret it. Maybe I just got a dud, but this new PNY GTX 750 Ti I got uses slightly more watts for less hashes than my Zotacs.

Max stable overclock Zotacs on Yacoin: 3.41kh/s
Max stable overclock PNY on Yacoin: 3.27kh/s

[gfgrimm]

Just got a 9th GTX 750 Ti. Decided to try a different brand this time and I regret it. Maybe I just got a dud, but this new PNY GTX 750 Ti I got uses slightly more watts for less hashes than my Zotacs.

Max stable overclock Zotacs on Yacoin: 3.41kh/s
Max stable overclock PNY on Yacoin: 3.27kh/s

I think the Zotac cards were one of the best ones to use...so you may be right.

[ithd]

My 5x 750 ti rig is giving me huge headaches. I'm running all 5 on powered risers and 6 pin connectors on a 750W RM corsair PSU. Asrock h81 btc pro mobo. 8gb RAM. Latest Nvidia drivers.

Is anyone getting huge lag/computer freezes when running multiple 750 Tis (5+) in one instance? I'm running windows 8.1 64 bit, and starting up all 5 at once freezes the PC (sometimes it's huge lag, sometimes I need to reboot, sometimes cudaminer crashes). Seems to only work with regular scrypt without freezing/crashing 100% of the time. I get 250 khs/card mining LTC at stock. I'm about to switch to Windows 7 to see if that changes anything.

If i don't use autotune, TDP usage seems to be pretty low (~850 khs for 5 cards), config: -l T5x24 -C 1 -H 0 -i 0
W/ autotune, I get low 80% TDP. 99% GPU usage.

Also any bit of OC to my cards (EVGA 750 ti ftw) seems to crash cudaminer about 10+ mins later (even "low" ones of +20 core, +200 mem). I don't know how other people seem to get huge clocks with these cards..

Any help would be greatly appreciated. Will throw 10 VTC anyone who can get me up and running stable at ~300+khs/card... Feel free to PM. Thanks!   

[jk_14]

No offense to those desiring screen to work with cudaminer, but I have to say, it seems like there are more pressing "bugs" and features for Christian to work out first. But then again, Christian is the one who is ranking the request priorities here.

BTW, Christian, if you need any additional testing against riser related items, let us know.

I have almost done 64bit linux bootable image for Nvidia Ti 750.
(similar to BAMT - any person familiar with BAMT should switch very easy to this new OS)

The only thing which block the image upload - is this problem regarding detached screen with cudaminer.
Imho a significant regress - would be lack of automated cudaminer startup, in case of power failure for example or after manual reboot.
And there are plenty of BAMT based miners...

aren't there alternatives to screen, like the bg / fg commands? What if you started a shell script with screen that runs cudaminer a bit later, say after a 1 seconds sleep command?

screen is de facto "standard", I would keep screen if possible
I have tried before the order: screen -> shell -> cudaminer  - no success, but I may look again a bit deeper...

[jk_14]

Hi,
  The problem with launching cudaminer from within screen is an issue with shared library linkage against libcudart.so.

There are at least two ways of solving this.  One way is to change the cudaminer build system so that it encodes the shared library location within the executable.  To do this, I changed the Makefile so that the linkage looks like this in the two places where -lcudart shows up:
  -lcudart -Wl,-rpath,/usr/local/cuda/lib64

The second way is to hack the Makefile and change cudaminer to link against libcudart_static.a (which is an alternative you can use if you compile with CUDA 5.5).

Using either solution, the problems with screen go away completely and it works as it should:

sphere - johns:/home/johns/cudaminer/mine/src.20140228/CudaMiner-2014-02-28 % screen -dm ./cudaminer --algo=scrypt -C 2 -d "Quadro 6000" -l F28x16 -D --benchmark
sphere - johns:/home/johns/cudaminer/mine/src.20140228/CudaMiner-2014-02-28 % screen -ls
There is a screen on:
        5002..sphere    (Detached)
1 Socket in /var/run/screen/S-johns.

I think Christian should be able to change his autoconf stuff to allow either of those approaches without too much difficulty, but for the time being people that care can simply hack the generated Makefile to cure this problem.

Cheers,
  John Stone

Great,

Thank you, I will check this asap.

[jstabb]

My 5x 750 ti rig is giving me huge headaches. I'm running all 5 on powered risers and 6 pin connectors on a 750W RM corsair PSU. Asrock h81 btc pro mobo. 8gb RAM. Latest Nvidia drivers.

Is anyone getting huge lag/computer freezes when running multiple 750 Tis (5+) in one instance? I'm running windows 8.1 64 bit, and starting up all 5 at once freezes the PC (sometimes it's huge lag, sometimes I need to reboot, sometimes cudaminer crashes). Seems to only work with regular scrypt without freezing/crashing 100% of the time. I get 250 khs/card mining LTC at stock. I'm about to switch to Windows 7 to see if that changes anything.

If i don't use autotune, TDP usage seems to be pretty low (~850 khs for 5 cards), config: -l T5x24 -C 1 -H 0 -i 0
W/ autotune, I get low 80% TDP. 99% GPU usage.

Also any bit of OC to my cards (EVGA 750 ti ftw) seems to crash cudaminer about 10+ mins later (even "low" ones of +20 core, +200 mem). I don't know how other people seem to get huge clocks with these cards..

Any help would be greatly appreciated. Will throw 10 VTC anyone who can get me up and running stable at ~300+khs/card... Feel free to PM. Thanks!  

First, try -H 2 instead of -H 0. Here's what I use to get ~300 kh/s per card (7 of the same card you have) with a +600 mem overclock (+0 GPU, GPU doesn't make as much difference on these cards and crashes quickly for me with any kind of OC):

cudaminer.exe -H 2 -i 0 -m 1 -b 4096 -l T5x24

Keep in mind when comparing overclock numbers with others on this thread that your card is already at +169 GPU and +183 mem over a stock 750ti. Any overclocks you do are on top of that.

Edit: Oh, and use the x86 version of cudaminer, it's about 5% faster than the x64 for these cards running scrypt.

[gfgrimm]

My 5x 750 ti rig is giving me huge headaches. I'm running all 5 on powered risers and 6 pin connectors on a 750W RM corsair PSU. Asrock h81 btc pro mobo. 8gb RAM. Latest Nvidia drivers.

Is anyone getting huge lag/computer freezes when running multiple 750 Tis (5+) in one instance? I'm running windows 8.1 64 bit, and starting up all 5 at once freezes the PC (sometimes it's huge lag, sometimes I need to reboot, sometimes cudaminer crashes). Seems to only work with regular scrypt without freezing/crashing 100% of the time. I get 250 khs/card mining LTC at stock. I'm about to switch to Windows 7 to see if that changes anything.

If i don't use autotune, TDP usage seems to be pretty low (~850 khs for 5 cards), config: -l T5x24 -C 1 -H 0 -i 0
W/ autotune, I get low 80% TDP. 99% GPU usage.

Also any bit of OC to my cards (EVGA 750 ti ftw) seems to crash cudaminer about 10+ mins later (even "low" ones of +20 core, +200 mem). I don't know how other people seem to get huge clocks with these cards..

Any help would be greatly appreciated. Will throw 10 VTC anyone who can get me up and running stable at ~300+khs/card... Feel free to PM. Thanks!   

I suspect that the Nvidia driver for W8.1x64 has a bug because I have 6x750 Ti's and I ran into the exact same thing.

Swap over to W7x64 and use settings: -l T5x24 -H2 -i 0 -C 1 -m 1
...I haven't really seen much difference between -m1 and no flag at all.

Use the x86 version of cudaminer when doing this on W7x64.

Finally, for whatever reason, Chrome running in the background with some Flash or HTML5 app (e.g. YouTube) gives a small boost.

Now you're good to go.

I really wanted this to work on W8.1x64 as well, but I think between MSFT/Nvidia/Cudaminer status quo, we're stuck with W7x64 for more than 2 cards running simultaneously.

[jstabb]

My 5x 750 ti rig is giving me huge headaches. I'm running all 5 on powered risers and 6 pin connectors on a 750W RM corsair PSU. Asrock h81 btc pro mobo. 8gb RAM. Latest Nvidia drivers.

Is anyone getting huge lag/computer freezes when running multiple 750 Tis (5+) in one instance? I'm running windows 8.1 64 bit, and starting up all 5 at once freezes the PC (sometimes it's huge lag, sometimes I need to reboot, sometimes cudaminer crashes). Seems to only work with regular scrypt without freezing/crashing 100% of the time. I get 250 khs/card mining LTC at stock. I'm about to switch to Windows 7 to see if that changes anything.

If i don't use autotune, TDP usage seems to be pretty low (~850 khs for 5 cards), config: -l T5x24 -C 1 -H 0 -i 0
W/ autotune, I get low 80% TDP. 99% GPU usage.

Also any bit of OC to my cards (EVGA 750 ti ftw) seems to crash cudaminer about 10+ mins later (even "low" ones of +20 core, +200 mem). I don't know how other people seem to get huge clocks with these cards..

Any help would be greatly appreciated. Will throw 10 VTC anyone who can get me up and running stable at ~300+khs/card... Feel free to PM. Thanks!   

I suspect that the Nvidia driver for W8.1x64 has a bug because I have 6x750 Ti's and I ran into the exact same thing.

Swap over to W7x64 and use settings: -l T5x24 -H2 -i 0 -C 1 -m 1
...I haven't really seen much difference between -m1 and no flag at all.

Use the x86 version of cudaminer when doing this on W7x64.

Finally, for whatever reason, Chrome running in the background with some Flash or HTML5 app (e.g. YouTube) gives a small boost.

Now you're good to go.

I really wanted this to work on W8.1x64 as well, but I think between MSFT/Nvidia/Cudaminer status quo, we're stuck with W7x64 for more than 2 cards running simultaneously.

Interesting. I have (3) GTX 750Ti and (1) GTX 780Ti on W8.1x64 and have not seen any issues. What motherboard are you running? Maybe it's a chipset driver issue rather than video card (my Win8.1 machine is on an old Asus P6T Deluxe for the record).

[gfgrimm]

My 5x 750 ti rig is giving me huge headaches. I'm running all 5 on powered risers and 6 pin connectors on a 750W RM corsair PSU. Asrock h81 btc pro mobo. 8gb RAM. Latest Nvidia drivers.

Is anyone getting huge lag/computer freezes when running multiple 750 Tis (5+) in one instance? I'm running windows 8.1 64 bit, and starting up all 5 at once freezes the PC (sometimes it's huge lag, sometimes I need to reboot, sometimes cudaminer crashes). Seems to only work with regular scrypt without freezing/crashing 100% of the time. I get 250 khs/card mining LTC at stock. I'm about to switch to Windows 7 to see if that changes anything.

If i don't use autotune, TDP usage seems to be pretty low (~850 khs for 5 cards), config: -l T5x24 -C 1 -H 0 -i 0
W/ autotune, I get low 80% TDP. 99% GPU usage.

Also any bit of OC to my cards (EVGA 750 ti ftw) seems to crash cudaminer about 10+ mins later (even "low" ones of +20 core, +200 mem). I don't know how other people seem to get huge clocks with these cards..

Any help would be greatly appreciated. Will throw 10 VTC anyone who can get me up and running stable at ~300+khs/card... Feel free to PM. Thanks!   

I suspect that the Nvidia driver for W8.1x64 has a bug because I have 6x750 Ti's and I ran into the exact same thing.

Swap over to W7x64 and use settings: -l T5x24 -H2 -i 0 -C 1 -m 1
...I haven't really seen much difference between -m1 and no flag at all.

Use the x86 version of cudaminer when doing this on W7x64.

Finally, for whatever reason, Chrome running in the background with some Flash or HTML5 app (e.g. YouTube) gives a small boost.

Now you're good to go.

I really wanted this to work on W8.1x64 as well, but I think between MSFT/Nvidia/Cudaminer status quo, we're stuck with W7x64 for more than 2 cards running simultaneously.

Interesting. I have (3) GTX 750Ti and (1) GTX 780Ti on W8.1x64 and have not seen any issues. What motherboard are you running? Maybe it's a chipset driver issue rather than video card (my Win8.1 machine is on an old Asus P6T Deluxe for the record).

For me, I did not see much stability issues until I pushed it beyond four cards. But there was definitely overhead noticed on the aggregate hash rate I was getting with four cards. I'm on Z87-Pro w/ Intel Core-i5 2670k. Latest drivers, etc.

[ManIkWeet]

Please people, try to keep your quote's smaller than/equal to your own text...

[Dotcommie]

First, try -H 2 instead of -H 0. Here's what I use to get ~300 kh/s per card (7 of the same card you have) with a +600 mem overclock (+0 GPU, GPU doesn't make as much difference on these cards and crashes quickly for me with any kind of OC): cudaminer.exe -H 2 -i 0 -m 1 -b 4096 -l T5x24 Keep in mind when comparing overclock numbers with others on this thread that your card is already at +169 GPU and +183 mem over a stock 750ti. Any overclocks you do are on top of that.

So you're doing +600 in addition to the EVGA overclock in bios or including? Mine seems to throw cuda 30 errors anytime I approach +415 in the precision x app. Still have only been able to get 315 on one of my evga ftw cards. The MSI gamer cards seem to get to 298 before giving the same error in cudaminer. Seem like totally stable overclocks besides those errors happening eventually.

[ivanlabrie]

...

So you're doing +600 in addition to the EVGA overclock in bios or including? Mine seems to throw cuda 30 errors anytime I approach +415 in the precision x app. Still have only been able to get 315 on one of my evga ftw cards. The MSI gamer cards seem to get to 298 before giving the same error in cudaminer. Seem like totally stable overclocks besides those errors happening eventually.

It's luck of the draw, oc wise, but it also depends on the types of memory chips used by the manufacturer.
We can say on average 300kh/s per 750ti after ocing it, which isn't half bad for 70-80w.

Now, is that smarter than mining yacoin with them? Probably not the case, keeping in mind all of them seem to do over 3.2kh/s and draw 18-20w whilst doing so.

[patoberli]

Isn't the card actually using under 40 W if not overclocked? So probably 45-50 overclocked?

[Napping]

Isn't the card actually using under 40 W if not overclocked? So probably 45-50 overclocked?

yeah, but it seem ppl usually include rig usage as well (cpu/psu losses). if yo usubtract 70-80W from total, it usually ends up about 50W per card with slight oc

[martinium]

Hi guys, I have a EVGA 750 Ti SC 2GB card and did the bios power mod and noticed something quite interesting.  I run 24/7 stable with no errors on +25 GPU and +500 MEM using precision x on win 7 x64.

My previous config I was using for my batch file based on reading forum posts was "cudaminer -a scrypt -o server -u user -p pass -l T5x24 -m 1 -C 1 -H 2 and was getting 295-297 kh/s stable. This got me full gpu being utilized and memory on card.

I did some more reading last night and switched to launch config of -l T10x24 and I now get between 310-312 kh/s 24/7 stable with same exact settings.  The craziest thing is that when I check my precision x software it only shows GPU utilization at 540 mhz compared to the full 1344 mhz of the prior launch config.  It's running faster at less GPU load and getting proper accepts!

I think this GPU has more untapped mining potential when we get actual maxwell launch configs and cuda 6 optimizations. Nvidia's mining future is looking brighter and brighter.

[cdogster]

I have a EVGA 750 Ti SC 2GB card and did the bios power mod

Can you please share a link to the bios power mod you used?

[lagur]

I have to wait for few minutes for the 2nd GPU (114kH/s) to match the hashrate of 1st GPU (145kH/s).

Is this normal?

[Benchman]

What is the maximum of GTX750 Ti in Maxcoin (keccak SHA3).

Can hit 200 mh/s max overclocked?