new to using cudaminer ..
as an example, if i have 3 cards running at 300khash, does the total of 900khash get reported to a mining pool?
or does a pool/worker read it as 300khash per card, as the share is accepted by the pool?
Topic: [ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX] - page 719. (Read 3426922 times)
March 08, 2014, 03:04:02 PM
March 08, 2014, 02:36:04 PM
Is it possible to set up two pools in one config with CUDAMiner? I'm looking to rent out my rig at Betarigs, and to do that I'd need to use two pool addresses and logins in my config.
Unfortunately there is no failover pool function for CUDAMiner yet. Would be very interested in this too. I am keen to give Betarigs a go wondering if it's possible to write a script that will switch off the backup miner when the machine is rented, and switch it back on once Betarig rental is done.
March 08, 2014, 12:06:18 PM
Anyone mining a scrypt-jane coin on their 750ti with Nfactor of 11 or higher? Keep getting bsods on mine with cmd line of: "cudaminer --algo=scrypt-jane:CACH -d 2 -o stratum+tcp:...." Using release 2-18-14. If you are can you share what your launch config is? Thank you.
Are you running the x86 or x64 version?
I get bsod when running the x86 version. It only does it on my msi 750ti. It does it if I running regular scrypt. I get the Nvidia driver error for the bsod.
Just got it working I had to choose a specific launch config kernel for it to not bsod. Looks like the autotune feature for Nfactor 11+ caused my pc to bsod. Getting about 55kh/s on my 750ti, not too bad getting about 60 kh/s on my 660ti on nfactor 11 coins. Anyone able to get more kh/s than 55 with their 750ti on nfactor 11 coins? Running kernel t128x2 -L 1.
March 08, 2014, 11:43:26 AM
You probably can run two cudaminers on a single card, but the performance is likely terrible...
I used to do this a lot, even run 3 instances of cudaminer on my GTX 680. No problem with anything, the hashrate was split evenly to each instance. 
March 08, 2014, 11:21:37 AM
Is it possible to set up two pools in one config with CUDAMiner? I'm looking to rent out my rig at Betarigs, and to do that I'd need to use two pool addresses and logins in my config.
How does work this renting business ? Does the guy who rent a rig, access the machine (through ssh or vpn) and control it or you just have to point the machine where he ask ?
March 08, 2014, 11:08:43 AM
Is it possible to set up two pools in one config with CUDAMiner? I'm looking to rent out my rig at Betarigs, and to do that I'd need to use two pool addresses and logins in my config.
What? No!Best you can do is terminate cudaminer after a period of time and start another one.
You probably can run two cudaminers on a single card, but the performance is likely terrible...
March 08, 2014, 11:02:16 AM
Is it possible to set up two pools in one config with CUDAMiner? I'm looking to rent out my rig at Betarigs, and to do that I'd need to use two pool addresses and logins in my config.
March 08, 2014, 09:26:20 AM
Try Win 7 but 64 bits and look if there is a bios update for the mobo. Disable in the bios everything not needed , com ports ..... etc.
March 08, 2014, 09:14:44 AM
Change to Windows 7 64bit maybe? Win8 is reported to have scalability issues.
oops I just read you're using Win 7 32 bit...
try Linux as a free alternative. OC support is forthcoming, but I have no idea about exact dates...
oops I just read you're using Win 7 32 bit...
try Linux as a free alternative. OC support is forthcoming, but I have no idea about exact dates...
March 08, 2014, 09:09:50 AM
Other then that make sure you pass -H 2 in your launch config
Yep the -H 2 parameter is active an already working
Just make some other test: teamviewer take indeed some cpu but is not the main issue here...I think that is something related to the 1X PCIE (bandwith? interrupt?)
With 5 cards still there is no wait to have a decent hash, taskmanager was not very accurate about witch process use most cpu, but monitor resource seems to be fine.
Look at the screenshot, the orange line on the graph is the "system interrupt" (there's a glitch on this screen so it's hard to read) is the process that take most of the cpu
Then with 3 cards (2 cards directly on 16x PCIE and the third one 16X slot to usb riser) the hashrate is finally ok and the "system interrupt" process don't take all the cpu
Another thing that I've noticed and baffle me: there are 2x2GB DDR onboard but with 5 cards only 1GB is available, with 4 cards only 2GB and finally all 4GB are usable with 3 cards
I really don't know what to do apart thinking to change motherboard...any suggestion?
March 08, 2014, 09:03:01 AM
I am getting with the -L2 -t30x16 around 121khash/sec was previously at -L2 -t30x32 at 277khash/s.
May-be dividing the second number by two will work (autotune is certainly a better option)
I was being dense...I hadn't rebooted since the cudaminer got a bunch of "does not validate" from mining when the nfactor changed, now that I have done so the hash is looking much more healthy!May-be dividing the second number by two will work (autotune is certainly a better option)
The best T autotune (T12x20 - 106kh) still marginally beats the best t one (t120x6 - 104kh) though (both with L2). My best speed is with exactly the same kernel settings as before, but with a lookup gap of 2.
I've still lost 65% of my 780 hash rate with the nfactor change, but that's much better than losing 80%.
March 08, 2014, 07:56:06 AM
Any news on a new cudaminer for Maxwell Nvidia 750, or the new Nvidia drivers, or the new drivers or best cuda 5.5? .
I would have news about the drivers, but if I told you they'd have to kill me.
Christian
March 08, 2014, 07:52:52 AM
Any news on a new cudaminer for Maxwell Nvidia 750, or the new Nvidia drivers, or the new drivers or best cuda 5.5? .
I would have news about the drivers, but if I told you they'd have to kill me.
Christian
March 08, 2014, 07:42:57 AM
That's strange I can't get passed 275 with my 780ti and using L2 give me a little more (that's why I was suggesting it)
Is your card plugged into a monitor (and/or using the google chrome trick... that's the only thing I haven't tried so far...)
It's kinda moot now that the nfactor has increased, but FWIW I have two 780s; the one that drives my monitors was getting 290-295kh, whilst the other one (no monitor connected) was getting 305-310kh. This is on my main PC rather than a mining rig, and Chrome is my browser of choice so is running just about all the time (gotta forum whore you know!). I've never specifically tried mining without Chrome running.Is your card plugged into a monitor (and/or using the google chrome trick... that's the only thing I haven't tried so far...)
With the new nfactor I'm struggling to get a decent speed though...best autotune I've found so far was for the same T12x20 but with a lookup gap of 2. This only yields a little over 50kh (over 80% down from before), whereas my R9 290s have only decreased by 50% (from ~340kh to ~170kh).
My 750ti is even worse - previously it was running at 150kh, now I can't get it over 10kh!
May-be dividing the second number by two will work (autotune is certainly a better option)
March 08, 2014, 07:38:19 AM
Thanks!
And you running the latest Nvidia drivers? 334.89? and your memory usage would be helpful...
2317MB memory usage (I know, it's not the max, not sure why)And you running the latest Nvidia drivers? 334.89? and your memory usage would be helpful...
Driver 331.39 appearantly
March 08, 2014, 07:32:58 AM
I just ran the blake algo on the blakecoin pool
The gtx780ti seems to run at 971Mhash/s, however all shares are rejected.
Debug gives this reject reason: H-not-zero
The gtx780ti seems to run at 971Mhash/s, however all shares are rejected.
Debug gives this reject reason: H-not-zero
H-not-zero means the hash of the header isn't zero for the first 32-bits.
Hmm, how's that possible considering this time I used the same sph_blake hashing
library as the cpuminer. I will have to focus a bit more on the scanhash function
then... I took it from maxcoin without much adjusting..
I guess I will have to join one of the non EU blakecoin pools, as the EU ones
don't send me the confirmation emails... EDIT: ah, successfully signed up for
the eu1 pool - and I can repro the problem now.
Christian
March 08, 2014, 07:30:51 AM
That's strange I can't get passed 275 with my 780ti and using L2 give me a little more (that's why I was suggesting it)
Is your card plugged into a monitor (and/or using the google chrome trick... that's the only thing I haven't tried so far...)
It's kinda moot now that the nfactor has increased, but FWIW I have two 780s; the one that drives my monitors was getting 290-295kh, whilst the other one (no monitor connected) was getting 305-310kh. This is on my main PC rather than a mining rig, and Chrome is my browser of choice so is running just about all the time (gotta forum whore you know!). I've never specifically tried mining without Chrome running.Is your card plugged into a monitor (and/or using the google chrome trick... that's the only thing I haven't tried so far...)
With the new nfactor I'm struggling to get a decent speed though...best autotune I've found so far was for the same T12x20 but with a lookup gap of 2. This only yields a little over 50kh (over 80% down from before), whereas my R9 290s have only decreased by 50% (from ~340kh to ~170kh).
My 750ti is even worse - previously it was running at 150kh, now I can't get it over 10kh!
March 08, 2014, 07:30:07 AM
Any news on a new cudaminer for Maxwell Nvidia 750, or the new Nvidia drivers, or the new drivers or best cuda 5.5? .
I would have news about the drivers, but if I told you they'd have to kill me.
Christian
March 08, 2014, 07:20:25 AM
Any news on a new cudaminer for Maxwell Nvidia 750, or the new Nvidia drivers, or the new drivers or best cuda 5.5? .
Now I'm 300 kh / s with t10x24 and overclock +135 mhz gpu +600 mem, driver 334.89 WHQL Windows 7 64-bit.
Now I'm 300 kh / s with t10x24 and overclock +135 mhz gpu +600 mem, driver 334.89 WHQL Windows 7 64-bit.
March 08, 2014, 06:23:22 AM
I just ran the blake algo on the blakecoin pool
The gtx780ti seems to run at 971Mhash/s, however all shares are rejected.
Debug gives this reject reason: H-not-zero
on an other topic: MRC finally reached Nfactor=11
The gtx780ti seems to run at 971Mhash/s, however all shares are rejected.
Debug gives this reject reason: H-not-zero
on an other topic: MRC finally reached Nfactor=11
Jump to: