Topic: [ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX] - page 702. (Read 3426976 times)
March 14, 2014, 12:16:16 PM
Binary only releases of heavycoin for CUDA miners would not work out due to GPL requirements.
People would just compile their own binaries without the fee and post them online.
Christian
March 14, 2014, 12:14:53 PM
personally when i go to get mine, they will all be on powered risers and they will be left at their default power usage. all having their own molex power.
my entire system with 3 750Ti's on full blast should pull around 140W
that's with it maxed out! i would rather save the power bill and lose some performance. They've already shown the 750Ti will hit 1.4GHz with NO power adjustments. so why change it at all?
my entire system with 3 750Ti's on full blast should pull around 140W
that's with it maxed out! i would rather save the power bill and lose some performance. They've already shown the 750Ti will hit 1.4GHz with NO power adjustments. so why change it at all?
March 14, 2014, 12:13:44 PM
Hi there, I've been using the 2-28 cudaMiner on my EVGA Superclocked version GTX 780. So far, I've only got it to work by using the 'auto' setting, every parameter I've tried to put in has led to it crashing, any tips on what else I can try? I only get 1.6 kH this way, I read about people getting up to 5 kH with this card. I don't want it screaming 24/7, and I like being able to use the desktop at the same time, but I should be able to do better than this.. here is my bat file:
cudaminer.exe -s 10 --algo=scrypt-jane:YAC -d 0 -l auto -o *server* -u *worker* -p *password*
Using x64 cudaminer on WIndows 8.1, latest Nvidia drivers.. Any ideas??
cudaminer.exe -s 10 --algo=scrypt-jane:YAC -d 0 -l auto -o *server* -u *worker* -p *password*
Using x64 cudaminer on WIndows 8.1, latest Nvidia drivers.. Any ideas??
March 14, 2014, 12:11:28 PM
Even at the default TDP of 38.5w (4 cards @ 154w), that's still dangerously close to the limit. Try to connect as few cards as possible to each power lead, or use a higher rated connector designed for more power, like the 6-pin or 8-pin PEG connectors.
I don't get why some are voluntarily going from 38.5 W to 65 W for maybe 10-15% higher kHash/s rates.
This kills Maxwells excellent power efficiency.... and apparently also PSU cable connectors.
Christian
March 14, 2014, 11:45:21 AM
@cbuchner1: I read that cudaminer is not yet optimized for the maxwell core. It does support it of course, but there might be some room for improvements. Is that true?
so far I haven't seen improvements by switching to CUDA 6.0 RC targeting the Compute 5.0 architecture in code - and it hurt performance for non-Maxwell chips (Compute 3.0 specifically). I'll wait for CUDA 6.0 final release then.
the good performance you're already getting from Maxwell comes from architectural differences (bigger cache, better instruction scheduling capabilities). So they're already exploited by the current code.
Christian
Ok, thanks for the statement!
March 14, 2014, 11:45:13 AM
PSA
I got a nice refresher course in power usage today. I was running 4x750 ti cards flashed with an unlocked TDP VBIOS, using powered USB risers, connected to a single 4-connector Molex cable. One unlocked card can pull 65.5w from the PCI express slot. Typical molex connectors have a max power handling of 187w (11A). With 4 unlocked cards, I was shoving 262w@12v down one cable. The end result:
http://m.imgur.com/vDqGNmh,787qoci
Even at the default TDP of 38.5w (4 cards @ 154w), that's still dangerously close to the limit. Try to connect as few cards as possible to each power lead, or use a higher rated connector designed for more power, like the 6-pin or 8-pin PEG connectors.
I got a nice refresher course in power usage today. I was running 4x750 ti cards flashed with an unlocked TDP VBIOS, using powered USB risers, connected to a single 4-connector Molex cable. One unlocked card can pull 65.5w from the PCI express slot. Typical molex connectors have a max power handling of 187w (11A). With 4 unlocked cards, I was shoving 262w@12v down one cable. The end result:
http://m.imgur.com/vDqGNmh,787qoci
Even at the default TDP of 38.5w (4 cards @ 154w), that's still dangerously close to the limit. Try to connect as few cards as possible to each power lead, or use a higher rated connector designed for more power, like the 6-pin or 8-pin PEG connectors.
March 14, 2014, 11:39:55 AM
@cbuchner1: I read that cudaminer is not yet optimized for the maxwell core. It does support it of course, but there might be some room for improvements. Is that true?
so far I haven't seen improvements by switching to CUDA 6.0 RC targeting the Compute 5.0 architecture in code - and it hurt performance for non-Maxwell chips (Compute 3.0 specifically). I'll wait for CUDA 6.0 final release then.
the good performance you're already getting from Maxwell comes from architectural differences (bigger cache, better instruction scheduling capabilities). So they're already exploited by the current code.
Christian
March 14, 2014, 11:37:09 AM
Today, I finally ordered my rig (5x EVGA 750Ti superclocked) which I will run under Linux. Since the overclocking possibilities are not good/existing under Linux (as far as I know), I have a question:
@cbuchner1: I read that cudaminer is not yet optimized for the maxwell core. It does support it of course, but there might be some room for improvements. Is that true? What is your oppinion for the 750Ti models? Do you think, you will be able to get a few more khash out of these cards?
@cbuchner1: I read that cudaminer is not yet optimized for the maxwell core. It does support it of course, but there might be some room for improvements. Is that true? What is your oppinion for the 750Ti models? Do you think, you will be able to get a few more khash out of these cards?
March 14, 2014, 11:27:02 AM
Christian, is this your blog? If not, they didn't use quotes. ;-)
http://cryptomining-blog.com/1424-gpu-miner-for-heavycoin-is-possible-and-is-being-worked-on/
the introductory statement " The author of CudaMiner has reported some success already..." should make it clear that the text below is a quote.
Christian
March 14, 2014, 11:11:00 AM
Yeh I'm up for that. I'm even up for adverts as long as their not huge and make noise
March 14, 2014, 11:09:17 AM
is ready!
Christian, did you think about adding a fee in the program, similarly to all those protoshares miners for ypool do?
That would make it more profitable for you to share the software earlier perhaps.
Say a 2.5% fee. That would be fair.
Christian, did you think about adding a fee in the program, similarly to all those protoshares miners for ypool do?
That would make it more profitable for you to share the software earlier perhaps.
Say a 2.5% fee. That would be fair.
as I stated, I haven't even begun thinking about modalities for a release yet. I am still in development mode here.
But I appreciate your input.
Whatever I do, this has to be agreed on with the other Christian
Christian
I feel you guys should add a fee in also. It is up to use,the public, to decide if we want to use miner.
I understand that it can be edited out but I feel that most people would appreciate the work from the authors and leave it alone. If it wasn't for their work we would not be making the profit.
I have been using the YPool Cudaminer that has a 3% fee. I know someone went in and changed it to that because original it was 6%. That was a little unfair on top of the 5% fee from Ypool in my opion. Well the author changed it to 6% because he had it at 10% to begin within. But the point is the person deserves some kind of copensation for their work and time put into the work involved.
JMHO.
Agreed.
I also have no issue with paying a fee.
March 14, 2014, 11:07:13 AM
is ready!
Christian, did you think about adding a fee in the program, similarly to all those protoshares miners for ypool do?
That would make it more profitable for you to share the software earlier perhaps.
Say a 2.5% fee. That would be fair.
Christian, did you think about adding a fee in the program, similarly to all those protoshares miners for ypool do?
That would make it more profitable for you to share the software earlier perhaps.
Say a 2.5% fee. That would be fair.
as I stated, I haven't even begun thinking about modalities for a release yet. I am still in development mode here.
But I appreciate your input.
Whatever I do, this has to be agreed on with the other Christian
Christian
I feel you guys should add a fee in also. It is up to use,the public, to decide if we want to use miner.
I understand that it can be edited out but I feel that most people would appreciate the work from the authors and leave it alone. If it wasn't for their work we would not be making the profit.
I have been using the YPool Cudaminer that has a 3% fee. I know someone went in and changed it to that because original it was 6%. That was a little unfair on top of the 5% fee from Ypool in my opion. Well the author changed it to 6% because he had it at 10% to begin within. But the point is the person deserves some kind of copensation for their work and time put into the work involved.
JMHO.
March 14, 2014, 11:02:10 AM
I know Yeh. Theirs no asking Christian a question anymore because there's 2 of them! Damn youuuuuuuu!
Glad to see you have some proper help now
Glad to see you have some proper help now
March 14, 2014, 11:00:11 AM
I would be curious to know if it is more profitable to use this new GPU miner on heavycoin in tandem with your CPU against just continuing to mine heavycoin on your CPU and devoting your GPUs to another Scrypt, Scrypt-N, Keccak etc. coin that is profitable.
Obviously this depends on your CPU of course. I'm using a 3930K, which gets about 400 khash with its overclock and I have a pair of 780ti cards mining another Scrypt-N coin.
Will be interesting to see how this all turns out! As always, thanks for your excellent work Christian(s)
Obviously this depends on your CPU of course. I'm using a 3930K, which gets about 400 khash with its overclock and I have a pair of 780ti cards mining another Scrypt-N coin.
Will be interesting to see how this all turns out! As always, thanks for your excellent work Christian(s)
March 14, 2014, 10:31:12 AM
we've accelerated 4 out of 5 hashing algos for HVC, but the last one (Groestl) is causing us a bit of trouble.
Our first port is actually slower than the CPU - at least on my system.
Running Groestl on CPU and the rest on GPU I am getting an 800% performance boost vs. running the CPU alone.
Multiple CPU threads can even share a GPU
Note that overall HVC is *not* CPU-only. They really need to change their advertising, and change it quickly. Whatever they did to SHA256 to turn it into a "GPU and ASIC resistant" HEFTY1 algorithm - it didn't work as they intended.
Christian
Christian, is this your blog? If not, they didn't use quotes. ;-)
http://cryptomining-blog.com/1424-gpu-miner-for-heavycoin-is-possible-and-is-being-worked-on/
March 14, 2014, 09:57:03 AM
does anyone have managed to run 6 x 750Ti under debian with cudaminer? I read somewhere it's hard to run more than 4 cards under linux, but havent tested it myself yet.
https://litecointalk.org/index.php?topic=16800.0
ive got a 6x 750ti rig running right now using this. gigabyte windforce gpus on a z87-g45 mobo... trying to get a 7th gpu to recognize.
Does your 750Ti's does have an additional power connector (6-pin pci-e power)?
yes they do
March 14, 2014, 08:32:56 AM
is ready!
Christian, did you think about adding a fee in the program, similarly to all those protoshares miners for ypool do?
That would make it more profitable for you to share the software earlier perhaps.
Say a 2.5% fee. That would be fair.
Christian, did you think about adding a fee in the program, similarly to all those protoshares miners for ypool do?
That would make it more profitable for you to share the software earlier perhaps.
Say a 2.5% fee. That would be fair.
as I stated, I haven't even begun thinking about modalities for a release yet. I am still in development mode here.
But I appreciate your input.
Whatever I do, this has to be agreed on with the other Christian
Christian
March 14, 2014, 08:17:36 AM
But adding a fixed fee can be removed by editing the sources.... But, you can make a donation, if you want
Christian
I'd be happy with a fixed fee and binary release only.
If sources are to be made available, I hereby pledge 5% of my earnings
- Myagui
March 14, 2014, 08:04:44 AM
What hash rates are you getting on your hardware? Just want to compare... something.
AMD Phenom II X6 1055T - CPU standalone ca 95 kHash/s on 6 threads
AMD Phenom II X6 1055T using the first 3 GPUs in the machine and 6 CPU threads: close to 800 kHash/s (Groestl remains on CPU)
cannot get GPU use readings, as GPU-z has decided to always crash on start after the previous driver update (revision 334 drivers I think). Currently running R 335 drivers from nvidia.
Nice, imagine that paired with a better cpu...:p
My X79 rig is ready!
Christian, did you think about adding a fee in the program, similarly to all those protoshares miners for ypool do?
That would make it more profitable for you to share the software earlier perhaps.
Say a 2.5% fee. That would be fair.
But adding a fixed fee can be removed by editing the sources.... But, you can make a donation, if you want
Christian
March 14, 2014, 07:59:01 AM
What hash rates are you getting on your hardware? Just want to compare... something.
AMD Phenom II X6 1055T - CPU standalone ca 95 kHash/s on 6 threads
AMD Phenom II X6 1055T using the first 3 GPUs in the machine and 6 CPU threads: close to 800 kHash/s (Groestl remains on CPU)
cannot get GPU use readings, as GPU-z has decided to always crash on start after the previous driver update (revision 334 drivers I think). Currently running R 335 drivers from nvidia.
Nice, imagine that paired with a better cpu...:p
My X79 rig is ready!
Christian, did you think about adding a fee in the program, similarly to all those protoshares miners for ypool do?
That would make it more profitable for you to share the software earlier perhaps.
Say a 2.5% fee. That would be fair.
Jump to: