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

[cbuchner1]

Currently using:

cudaminer.exe -H 2 -d 0,1 -l K16x16,K16x16 -i 1,0

Should we setup another bat file and define -d 2,3 ?

the shortest command line to mine on all 4 devices would be

Code:

cudaminer.exe -H 2 -l K16x16 -i 1,0

options that are only given once will be replicated across all devices found in the machine,
the -i 1,0 makes only GPU #0 use interactive mode, all others use -i 0

[fruitsdemers]

Every pool still reports only half of my keccak hashrate for some reason.
It doesn't seem to happen with amd cards though so I wonder if this is a problem on my end...

[cbuchner1]

Nvidia will give you some big $$...would turn those 40w/350kh/s 750ti's into proper mining cards if you did.

EDIT: and the linux oc soft? I want precision x for linux :p -will donate if these things come out-

Not sure about this big $$ part. There is no contract or agreement between me and nVidia. What I am doing is a hobby only.

The only thing that made me big $$ were their GPUs (hello MaxCoin). Oh, and I got a free 750Ti to play with, but no earlier than anyone else.

[cbuchner1]

Every pool still reports only half of my keccak hashrate for some reason.
It doesn't seem to happen with amd cards though so I wonder if this is a problem on my end...

does anyone else have such problems? Are you mining MAX pools or Helix pools?

[fruitsdemers]

Every pool still reports only half of my keccak hashrate for some reason.
It doesn't seem to happen with amd cards though so I wonder if this is a problem on my end...

does anyone else have such problems? Are you mining MAX pools or Helix pools?

Helix at the moment but I remember the maxcoin pools doing the same thing.
It's actually slightly higher than half but it's consistently about 40% lower than it should...

I'm currently using your latest build with the T2880x24 command on a 780ti.

[cbuchner1]

Helix at the moment but I remember the maxcoin pools doing the same thing.
It's actually slightly higher than half but it's consistently about 40% lower than it should...
I'm currently using your latest build with the T2880x24 command on a 780ti.

we could try tinkering a bit with the difficulty target in the cudaminer code (artificially lowering it).
If the pools still accept these shares, you would see a higher hash rate and higher payouts

Christian

[ivanlabrie]

EDIT: and the linux oc soft? I want precision x for linux :p -will donate if these things come out-

happy with a basic command line tool, too?

Yeah, sure thing

Getting 248mh/s per GTX 780 at 1201mhz core.

[cvax]

I just got my batch of 8 Zotac 750 Ti's and am trying them out. How are you guys getting -l T5x24 to work? When I try that I get a huge error list in screenshot.


I am trying to mine Yacoin and here is my command line.
cudaminer.exe --algo=scrypt-jane -H 1 -L 2 -i 1 -l t5x24 -C 1 -m 1 -o stratum+tcp://yac.coinmine.pl:9088

Any ideas?

Auto tuner said t8x8.

[cbuchner1]

I am trying to mine Yacoin and here is my command line.
cudaminer.exe --algo=scrypt-jane -H 1 -L 2 -i 1 -l t5x24 -C 1 -m 1 -o stratum+tcp://yac.coinmine.pl:9088

I get this (and bluescreens) when I dial up the core or mem clocks too much. Try starting out with stock clocks.

Also go with the autotune suggestion. t5x24 might require too much memory for your card.

Yacoin currently takes 4 MB per hash, the lookup gap reduces this to 2MB.  You run 5x24x32=3840 hashes in parallel, requiring 7.5 GB of memory.
Yikes.

Here's how I run yacoin on 750 Ti:

Code:

cudaminer.exe -H 0 --algo=scrypt-jane -d gtx750ti -L 2 -l t64x1 -i 0 -o stratum+tcp://yac.coinmine.pl:9088 -O secret:secret

Christian

[ManIkWeet]

I get this (and bluescreens) when I dial up the core or mem clocks too much. Try starting out with stock clocks.

Also go with the autotune suggestion. t5x24 might require too much memory for your card.

Christian

You -should- really work on making these errors a bit more clear somehow...

[cvax]

I am trying to mine Yacoin and here is my command line.
cudaminer.exe --algo=scrypt-jane -H 1 -L 2 -i 1 -l t5x24 -C 1 -m 1 -o stratum+tcp://yac.coinmine.pl:9088

I get this (and bluescreens) when I dial up the core or mem clocks too much. Try starting out with stock clocks.

Also go with the autotune suggestion. t5x24 might require too much memory for your card.

Yacoin takes 4 MB per hash, the lookup gap reduces this to 2MB.  You run 5x24x32=3840 hashes in parallel, requiring 7.5 GB of memory.
Yikes.

Christian

Hi Christian,

I got the t5x24 from some of the other users postings here. Hmm.
I am currently on stock core and mem as I literally just plugged in the card. Mind if I ask you how exactly you have your Yacoin parameters set up? Are you using -b 4096 like you have in the cudaminer readme.txt? Thanks a bunch.

Edit: I will try your settings.

[cbuchner1]

Hi Christian,

I got the t5x24 from some of the other users postings here. Hmm.
I am currently on stock core and mem as I literally just plugged in the card. Mind if I ask you how exactly you have your Yacoin parameters set up? Are you using -b 4096 like you have in the cudaminer readme.txt? Thanks a bunch.

I edited above posting with my parameters.

Hmm, I am indeed missing the -b 4096 there. You can even go to -b 32768 for maximum performance if you don't run video output on that card.

[cvax]

I edited above posting with my parameters.

Hmm, I am indeed missing the -b 4096 there. You can even go to -b 32768 for maximum performance if you don't run video output on that card.

Damn, with your settings (no -b 4096 yet) I am only getting 2.36-2.44 khash/s. I recall you said you got something like 3.0 khash/s?

[gfgrimm]

I've now got 6 750 Ti's running on Win8 x64. I'm pretty disappointed.

1 Cards, OC'd, in x16 slot (running at x16): 320-330 kH/s [Total of 320-330 kH/s]
2 Cards, OC'd, in 2 x16 slots (each running at x8): 300 kH/s each  [Total of ~600KH/s]
4 Cards, OC'd, running on x1-x16 risers (each running at x1): ~270-280 each [Total of ~1100 kH/s]
6 Cards, stock, running on x1-x16 risers (each running at x1): 230-240 each (one runs at ~210 and is operating at PCIe 1.1 x1...whereas all the other cards are running at PCIe 2.0 x1) [Total of ~1430kH/s]

I haven't been able to get 7 Cards to show up in Windows...figure that it could be a MoBo/BIOS address space limitation.

I can't really apply any overclock at all when I have 6 cards loaded. MSI AB, GPU-Z, etc., all feel a little unstable when using. Applying even just +100MHz to MEMCLK will cause cudaminer to crash.

I have also noticed that my little G3220 is way too puny to run -H 1 flag.

Here are the flags I'm running -H 2 -i 0 -l T25x16

Will I see improved scaling in Ubuntu with cudaminer?

I'm thinking of returning all of these cards because 230 kH/s per card is not so great. If I was even hitting 280, I'd maybe be able to swallow this.

What is the primary factor here causing a scalability issue? Is it the PCIe operating mode? Is it Windows 8? Is it the Nvidia Driver? Is cudaminer not written to scale well past 4 GPUs?

Yes, I did try running 2 or 4 cards each in a separate cudaminer instance. While the system stability improved, the results yielded the same aggregate hash rate, and any attempts to boost any clocks cause a quick cudaminer crash.

If I can't get more out of these cards by Monday, I'll have to send them back to Newegg.

Am I misunderstanding something? With:

1 Cards, OC'd, in x16 slot (running at x16): 320-330 kH/s [Total of 320-330 kH/s]
2 Cards, OC'd, in 2 x16 slots (each running at x8): 300 kH/s each  [Total of ~600KH/s]
4 Cards, OC'd, running on x1-x16 risers (each running at x1): ~270-280 each [Total of ~1100 kH/s]

That's 7 cards (with OC) at best case 292KH/s ave, or 285KH/s ave? More than the 280 you said you might be able to swallow? Are you factoring in power usage into your valuation of the cards? It seems they use even less than first thought.

I've got 5 x Gigabyte 750 Ti waiting to plug into a Win 7 x64 system but I need risers before I get going, so I'm very interested in the experiences and thoughts of yourself and j0achim (and indeed any others with 5+ cards in a single rig!).

You do misunderstand. Each line is a different configuration to track the scalability of performance--I tried 1 card, 2 cards, 4 cards, and then 6 cards. The performance I listed above is what I got.

The degradation of performance and overclockability seems to be linear with the # of cards plugged in at once...

[myagui]

Helix at the moment but I remember the maxcoin pools doing the same thing.
It's actually slightly higher than half but it's consistently about 40% lower than it should...
I'm currently using your latest build with the T2880x24 command on a 780ti.

we could try tinkering a bit with the difficulty target in the cudaminer code (artificially lowering it).
If the pools still accept these shares, you would see a higher hash rate and higher payouts

Christian

Just to add that I too have a similar experience, with the latest & greatest release from today (and also around Max release). I'm seeing the same 40% less hashrate reported by the pool with Helixcoin. Nothing seems to change if I setup separate workers for my GTX770 & the GTX 750 TI, or if I keep them on a single launch command.

The only thing odd that I've noticed, is that the hashrate reported by the pool get's closer to the value reported by Cudaminer, when higher Vardiff kicks in (at say 128). At Vardiff 32 & 64 the discrepancy is much larger, and it's not a case of just the expected variations in share findings or luck, as I see the same hashrate reported over long periods of time (as long as the Vardiff remains the same).
Is it possible that the "changes" or "fix" that you worked on shortly after the Maxcoin launch have left some gremlin vestiges? 

PS: Again an awesome bump for the 'green team' with that latest update with the hand from nvidia. Big thanks for the update Christian. I'm surely going to donate as soon as I'm fortunate to get a break on a new coin/algo launch.
Cheers!

[bacon_wrangler]

I have 1.8 helixcoins!  .... 

Today's cudaminer release won't help me, as I have a compute 3.0 card, correct?

[myagui]

I have 1.8 helixcoins!  .... 

Today's cudaminer release won't help me, as I have a compute 3.0 card, correct?

Correct. Helped me with 750 TI, but no changes with 770 for example.

[trancelord]

Ok, how do I set my gtx 680

cudaminer to change pools if 1 cant connect for a certain amount of time?

Would like to do multiple pools.



I got this so far, but i want to add more pools

cudaminer.exe -o stratum+tcp://minercrew.org:3336 -u Miner.user -p password
cudaminer.exe -d 0,1,2 -i 1,0,0 -l auto,S27x3,28x4 -C 0,2,1

HElp will be appreciated and maybe i can donate??  

[cvax]

Zotac 750 Ti
Stock = 2.42kh/s on Yacoin, 265 kh/s on scrypt

Overclocked +135 core + 600 mem with increased TDP bios
Cudaminer 2/18 = 3.15kh/s on Yacoin, 305kh/s on scrypt
Cudaminer 2/28 = 3.35kh/s on Yacoin!  

cudaminer.exe --algo=scrypt-jane -H 2 -L 2 -i 0 -l t64x1 -b 4096 -m 1

This is inside a PCI-E 16x slot as the only card right now. I will try scaling up with a few more cards now too.

[fruitsdemers]

Helix at the moment but I remember the maxcoin pools doing the same thing.
It's actually slightly higher than half but it's consistently about 40% lower than it should...
I'm currently using your latest build with the T2880x24 command on a 780ti.

we could try tinkering a bit with the difficulty target in the cudaminer code (artificially lowering it).
If the pools still accept these shares, you would see a higher hash rate and higher payouts

Christian

Just to add that I too have a similar experience, with the latest & greatest release from today (and also around Max release). I'm seeing the same 40% less hashrate reported by the pool with Helixcoin. Nothing seems to change if I setup separate workers for my GTX770 & the GTX 750 TI, or if I keep them on a single launch command.

The only thing odd that I've noticed, is that the hashrate reported by the pool get's closer to the value reported by Cudaminer, when higher Vardiff kicks in (at say 128). At Vardiff 32 & 64 the discrepancy is much larger, and it's not a case of just the expected variations in share findings or luck, as I see the same hashrate reported over long periods of time (as long as the Vardiff remains the same).
Is it possible that the "changes" or "fix" that you worked on shortly after the Maxcoin launch have left some gremlin vestiges? 

PS: Again an awesome bump for the 'green team' with that latest update with the hand from nvidia. Big thanks for the update Christian. I'm surely going to donate as soon as I'm fortunate to get a break on a new coin/algo launch.
Cheers!

Thank you for corroborating.
Over longer periods of time, the hashrate reported by the pool seems to shrink even more to perhaps 1/5th of what's being reported directly from the client. After trying all manners of shenanigans, I've noticed that restarting my cudaminer every few minutes seems to attenuate, if not eliminate, the problem a little so it definitively feels like it's some kind of error with vardiff...

Unless this can be somehow fixed, it's currently more profitable to mine vertcoin.