[ANN] ccminer 2.3 - opensource - GPL (tpruvot)

[tbearhere]

CCMINER INTENSITY-

I have an Ubuntu 14.04.1 build, x64, and have had success with x13 and an intensity of 20.5.  I have since tried various intensity settings with x11 and Lyra2RE algorithms.  Algo x11 maxes out with "-i 20.5"' but the improvement is very small compared to the default intensity of 19, only about +25kh/s for the complete rig (6 x 750ti).  Lyra2RE does not improve with an increased intensity, or a lowered intensity.  With Lyra2RE, an intensity setting of "-i 21" or greater resulted in HUGE kh/s values and no accepted shares.  Algos x11 and x13 generated "out of memory" errors at settings of "-i 20.7".  

I have used the setting of "./ccminer -a x13 -d 0,1,2,3,4,5 -i 20.5,20.5,20.5,20.5,20.5,20.5 -o stratum+tcp://URL" in my launch string.  

So, overall, x13 mining with "-i 20.5" gives +125kh/s on my 6-card rig (14,300kh/s), x11 mining with "-i 20.5" improves by 25kh/s on the same rig (17,900kh/s).  For Lyra2RE, the intensity settings had little or no effect on accepted hash rate.              
--scryptr

So you have to set intensity for each card.

Apparently not!   I just redid my command line, using minimal instruction flags.  In testing x15, I used the command line "./ccminer -a x15 -i 20 -o stratum+tcp://URL".  On my 6-card 750ti rig, it gave an improvement of +70kh/s, or 12,350 kh/s vs 12,280kh/s at the default setting of 19 (no flag).  Any setting greater than 20 resulted in a memory error.

I once had a mixed bag of nVidia cards, and was in the habit of setting each card individually.       --scryptr

ok thank you

[scryptr]

CCMINER INTENSITY-

I have an Ubuntu 14.04.1 build, x64, and have had success with x13 and an intensity of 20.5.  I have since tried various intensity settings with x11 and Lyra2RE algorithms.  Algo x11 maxes out with "-i 20.5"' but the improvement is very small compared to the default intensity of 19, only about +25kh/s for the complete rig (6 x 750ti).  Lyra2RE does not improve with an increased intensity, or a lowered intensity.  With Lyra2RE, an intensity setting of "-i 21" or greater resulted in HUGE kh/s values and no accepted shares.  Algos x11 and x13 generated "out of memory" errors at settings of "-i 20.7".  

I have used the setting of "./ccminer -a x13 -d 0,1,2,3,4,5 -i 20.5,20.5,20.5,20.5,20.5,20.5 -o stratum+tcp://URL" in my launch string.  

So, overall, x13 mining with "-i 20.5" gives +125kh/s on my 6-card rig (14,300kh/s), x11 mining with "-i 20.5" improves by 25kh/s on the same rig (17,900kh/s).  For Lyra2RE, the intensity settings had little or no effect on accepted hash rate.              
--scryptr

So you have to set intensity for each card.

Apparently not!   I just redid my command line, using minimal instruction flags.  In testing x15, I used the command line "./ccminer -a x15 -i 20 -o stratum+tcp://URL".  On my 6-card 750ti rig, it gave an improvement of +70kh/s, or 12,350 kh/s vs 12,280kh/s at the default setting of 19 (no flag).  Any setting greater than 20 resulted in a memory error.

I once had a mixed bag of nVidia cards, and was in the habit of setting each card individually.       --scryptr

ok thank you

MORE ON CCMINER INTENSITY-

Just for thoroughness, I also ran some tests on Keccak.  The Keccak hash rate for the full rig at default was 889,000kh/s to 910,000kh/s, with fluctuation.  With flag settings in the range of "-i 22.5" to "-i 23.5", the 6-card rig was able achieve hash rates of 925,000kh/s to 935,000kh/s.  The Kecccak hash rate fluctuated too much to make a good average in a brief time.  I was able to generate acceptable hash with a setting of "-i 24.5"' but the results appeared to be slightly lower than max, perhaps 910,000 or so.  A setting of "-i 25" resulted in an error.  Overall, a hash rate increase of 20,000kh/s to 30,000kh/s is possible.      
--scryptr

[Epsylon3]

https://docs.google.com/spreadsheets/d/1dI1Cc3JhhsA-UdIRlndvQ9huX8FmoJT2TXFDB0wdp54/edit#gid=1086083191

[Shivoreun]

Hi to all.
Can anyone, please, help me with miner config... The commandline is:
ccminer.exe -a lyra2 -o stratum+tcp://pool.verters.com:4444 -u snpminer.1 -p x

And the miner answer is:
[2014-12-24 15:19:48] NVAPI GPU monitoring enabled.
[2014-12-24 15:19:48] Starting Stratum on stratum+tcp://pool.verters.com:4444
[2014-12-24 15:19:48] 1 miner thread started, using 'lyra2' algorithm.
[2014-12-24 15:19:48] Stratum difficulty set to 0.1
[2014-12-24 15:19:48] pool.verters.com:4444 lyra2 block 215189
Cuda error in func 'keccak256_cpu_init' at line 299 : invalid device symbol.

Trying to run it on GT 630. What I'am doing wrong?
Thanks in advance.

[Epsylon3]

which card ? its for Direct Compute 3.0 mini

[Shivoreun]

Ok, thanks, found a problem. GT 630 is 2.1 only
Is there any miner, which supports 2.1 compute?..

[scryptr]

Ok, thanks, found a problem. GT 630 is 2.1 only
Is there any miner, which supports 2.1 compute?..

As far as I know, Bombadil (KBomba) compiled the last version of ccminer with Cuda 2.0/2.1 support.  Here are links:

     "http://cudamining.co.uk/url/releases/member/6" or "https://github.com/KBomba/ccminer-KBomba"
     or "http://cryptomining-blog.com/wp-content/files/ccminer-1.0-windows-compute-2.zip"
  
I am able to run x11 at about 800-850 kh/s on my 550ti with the v1.0 release.

The rest of the ccminer versions of this vintage dropped Cuda 2.0/2.1 support.

The "cudamining.co.uk" site needs an update!    Newer contributions should be available there.  However, it was a Windows-centric crowd that contributed.  I wanted to run Zelante's split-screen utility, but it was only available for Windows, Cuda 3.0+, and Cuda 6.5 toolkit.       --scryptr

[K1773R]

Ok, thanks, found a problem. GT 630 is 2.1 only
Is there any miner, which supports 2.1 compute?..

try compiling it for 2.1

[Shivoreun]

As far as I know, Bombadil (KBomba) compiled the last version of ccminer with Cuda 2.0/2.1 support.  Here are links:
     "http://cudamining.co.uk/url/releases/member/6" or "https://github.com/KBomba/ccminer-KBomba"
     or "http://cryptomining-blog.com/wp-content/files/ccminer-1.0-windows-compute-2.zip"

Tnaks. I've already tried this version, but it don't know what is Lyra2RE

[scryptr]

Ok, thanks, found a problem. GT 630 is 2.1 only
Is there any miner, which supports 2.1 compute?..

try compiling it for 2.1

I believe that versions of Ccminer greater than 1.0 began using Cuda code that is not supported by Cuda v2.0/2.1 chip architecture.       --scryptr   

[Epsylon3]

yes a big part of recent improvements are made with GPU assembler commands, which was really basic on gpus before SM 3.0

Merry XMas to all :p

[sp_]

Two of the users in my thread has found a problem in the stratum implementation. CCminer is sending low difficulty shares, and this triggers some kind of protection on the pool and the payout is reduced. Please investigate

[sp_]

I think the fix is easy.

When the pool ask for a difficulty change all threads must be suspended and restarted. We don't want old threads to submit shares after the difficulty has been changed.

[GES]

neoskrypt gtx 970 100kh?

[scryptr]

neoskrypt gtx 970 100kh?

DJM34 is officially developing neoscrypt code for Cuda.  He has given progress reports on the hashing speeds for various nVidia cards in sp_'s ccminer thread.  He also announced a crowdfunding plan.       --scryptr

[tbearhere]

neoskrypt gtx 970 100kh?

DJM34 is officially developing neoscrypt code for Cuda.  He has given progress reports on the hashing speeds for various nVidia cards in sp_'s ccminer thread.  He also announced a crowdfunding plan.       --scryptr

Where does one pay?

[djm34]

neoskrypt gtx 970 100kh?

DJM34 is officially developing neoscrypt code for Cuda.  He has given progress reports on the hashing speeds for various nVidia cards in sp_'s ccminer thread.  He also announced a crowdfunding plan.       --scryptr

Where does one pay?

you can send donation to the address for that project which is in my signature: 18UZRrybgjmR4M53Aca4Q8YBguqErZ5qaP
The miner will be release once enough money has been collected (ideally 5btc... but may-be less... considering the btc is not really high)

Current speed: gtx750ti: ~120kh/s
                     gtx780ti: ~200kh/s
                     gtx 980:  ~375kh/s

[flipclip]

I have two different nvidia 750Ti's (GIGABYTE and a EVGA "superclocked") in my rig (Ubuntu 14.04).  I was trying to adjust the intensity of the cards (just to see if it made any hashing difference) individually using -d 0,1 -i 19,21 (also --devices 0,1 --intensity 19,21) but when I checked the api call of "threads" the "I" value was 19 for both cards.  I tried a few different number combos and whatever was the first number for -i was the "I" value for both cards.  Also when I placed a high number as the first intensity (example: 28), ccminer would crash out (expected), but if that number was the second one ccminer would work fine.  It seems the second value is being skipped.  What should the command be for setting intensity per card? Or does this only work if you have two types of cards, say a 750Ti and a 980? Or is it not a per card setting at all?  Thanks.

[Epsylon3]

indeed this syntax is not yet supported, i will try to do it...

[flipclip]

Thanks for your reply.  Just wanted to make sure I wasn't doing something wrong.  I guess I can try starting up separate ccminer instances utilizing the -d switch.