CCminer(SP-MOD) Modded GPU kernels.

[chrysophylax]

Hi all!
I was wondering what was your setting with a 750 ti and 960 rigs on quark. For the moment I'm still trying to find a good intensity. I'm getting 5-5.5 MH per 750 and roughly 8-9 MH per 960...

Thanks a lot!

that depends quite a lot on the hardware itself - but our gigabyte 750ti oc lp cards max out at -i 24 ...

dont know what the -X parameter is - as ive been out commission for almost three weeks now sick - so no tests have been done on the xintensity ...

dont have any 960 cards - so cant help with that one ...

#crysx

[antho281]

thanks for the info!
Don't really know what is the X factor neither, I've read the last pages and still not sure about it.
Actually Running on G3220 @ 3.00GHz × 2 with 16 GB Ram (DDR3-1600) for both my 750 ti RIG (x6) and 960 RIG (x6)

I'll try -I 24!

[sp_]

Try -i 22.9 . You should get 6mhash with.release 61

[antho281]

I'll test, thanks a lot Sp_!
Gonna send my beer this week, no choice

[pallas]

Release 61

Just downloaded the new release and I get "Cuda error in func 'scanhash_lyra2v2' at line 127 : out of memory." on my 6X750Ti rig, running lyra2v2.
Never had this error before.

And some donations from VTC income for the hard working devs in this great community:

- @sp_ : 6fbe32c66687bd5d4c7935b7810e1e0afe8e34e4eef6a8842e754b1cd0f16f76
- @djm34 : 9bb440455556ca17d660f4b896c1c6f66e10654e8f66181e73aeed5b9d6f7c21
- @pallas : 340dc9a0730d0f42f2c1741cd12558d9b41a87c6aacf5bf6c7f22c8681fc23d5

Thanks very much sir!

[sp_]

Submitted another optimalisation in lyra2v2. This change was based on the change tvpruvot did in his fork (skein256 and bmw256),
but I also removed the stack usage, reduced the registercount and threads per block.

lyra2v2 +100KHASH on 970 and 50khash on 750ti (1.1% faster)

[chrysophylax]

Try -i 22.9 . You should get 6mhash with.release 61

would that be the sweetspot for the 750ti? ...

i just suggested what thefarm maxes out at before scanhash errors start ...

what would you suggest the -X parameter at sp? ...

#crysx

[chrysophylax]

Submitted another optimalisation in lyra2v2. This change was based on the change tvpruvot did in his fork (skein256 and bmw256),
but I also removed the stack usage, reduced the registercount and threads per block.

lyra2v2 +100KHASH on 970 and 50khash on 750ti (1.1% faster)

nice - will test today ...

will be going into the office today whether sick or not ... been away from it and working for three weeks now due to illness ...

if i cant make it through today - ill be out for another week ...

the work for the donation links are all there ... just needs community testing and use ...

tanx for the great work ...

#crysx

[sp_]

I have added some pull requests from a user called t-nelson. Hopefully the crash on ctrl+c is gone now.
Need to test some more. The head version on git might be unstable on windows.

[sp_]

Submitted another optimalisation in lyra2v2. This change was based on the change tvpruvot did in his fork (skein256 and bmw256),
but I also removed the stack usage, reduced the registercount and threads per block.
lyra2v2 +100KHASH on 970 and 50khash on 750ti (1.1% faster)

nice - will test today ...
will be going into the office today whether sick or not ... been away from it and working for three weeks now due to illness ...
if i cant make it through today - ill be out for another week ...
the work for the donation links are all there ... just needs community testing and use ...
tanx for the great work ...
#crysx

Any luck?

gtx 970 factory clocks (gigabyte windforce oc)

Original ccminer lyra2v2: 9.5MHASH
sp-mod 61+: 9.85MHASH

+3.68%

[chrysophylax]

Submitted another optimalisation in lyra2v2. This change was based on the change tvpruvot did in his fork (skein256 and bmw256),
but I also removed the stack usage, reduced the registercount and threads per block.
lyra2v2 +100KHASH on 970 and 50khash on 750ti (1.1% faster)

nice - will test today ...
will be going into the office today whether sick or not ... been away from it and working for three weeks now due to illness ...
if i cant make it through today - ill be out for another week ...
the work for the donation links are all there ... just needs community testing and use ...
tanx for the great work ...
#crysx

Any luck?

gtx 970 factory clocks (gigabyte windforce oc)

Original ccminer lyra2v2: 9.5MHASH
sp-mod 61+: 9.85MHASH

+3.68%

my apologies sp - i havent gone in today ...

have been in and out of bed with fever again ( relapse ) with my second course of antibiotics half way through and the third course ready ... so i have not left home or the bed so far ...

my main internet link is to the office is down at the moment also - so i cant remote in to try and compile ... no fun at the the moment ... if all is well tomorrow - ill go in and test ... i guess my determination was shot down in flames ... virus 1 - crysx 0 ...

is there much of an increase for the 750ti? ... what would be the best intensity setting to use? ... it seems that the -X 15 is the sweetspot for lyra2v2 on 750ti cards ... are there any things we need to look out for - or need to do - to maximize the testing? ...

i have a system that is fedora 22 x64 with cuda 7.0.28 - and that has issues compiling ... but when the issues are alleviated and it does compile - it has massive cpu validation errors ... is there anything that can be done to reduce / eliminate these errors ...

is it worth trying to 'upgrade' to cuda 7.5 and compiling with that instead? ...

those are some of the things i will be looking at in the test when i get back into the office ... hopefully tomorrow ...

#crysx

[sp_]

Cuda 7.0 is broken
Cuda 7.5 is dead slow (x11 down from 3MHASH to 1.9)

You should use cuda 6.5

You could consider to switch your rigs to windows. Less driver/compiler trouble. Easier to overclock. Bether profit switching software.
Install a free 64bit windows 7 trial version, and strip it. (make it Autoboot) Alter the bois so that the computer will restart automaticly after a powerloss.

Some of my windows rigs have been running for months.

[chrysophylax]

Cuda 7.0 is broken
Cuda 7.5 is dead slow (x11 down from 3MHASH to 1.9)

You should use cuda 6.5

You could consider to switch your rigs to windows. Less driver/compiler trouble. Easier to overclock. Bether profit switching software.
Install a free 64bit windows 7 trial version, and strip it. Autoboot

Some of my windows rigs have been running for months.

i see ... i have 6.5 installed on all the other miners - including the test machine ...

i tried that previously - a few weeks back now - and the installation and fiddling with the install as well as the screwing around with the windows validation is too much headache for me to deal with ... its also ONLY gui - which is dead slow on a remote level compared to console and has no where near the functionality that linux console has ...

i wont be playing with windows for a long time yet sp - especially due to its licensing costs and bullshit associated with microsoft ( yes i know - you CAN use it without one - but i cant afford to have that in the farm ) ... its too much effort for use in thefarm - way too much effort ...

linux oc and components of cuda will catch up soon enough ...

ive just got one of the internet links back online - i will log in to the test system and recompile ... will let you know in about 30mins ...

#crysx

[chrysophylax]

hi sp ...

initial results of fresh compile and mining at givemecoins using only -X 15 and --diff .5 with no oc or firmware tweaking ... just stock ...

*** ccminer 1.5.61-git(SP-MOD) for nVidia GPUs by sp-hash@github ***
        Built with the nVidia CUDA SDK 6.5

  Based on pooler cpuminer 2.3.2 and the tpruvot@github fork
   CUDA support by Christian Buchner, Christian H. and DJM34
  Includes optimizations implemented by sp , klaust, tpruvot and tsiv.

[2015-08-24 16:56:18] tid(0x0000047e) XIntensity set to 15, 983040 cuda threads
[2015-08-24 16:56:23] tid(0x00000515) Starting Stratum on stratum+tcp://mining-ace.granitecoin.com:6001/


[2015-08-24 16:57:55] tid(0x00000515) accepted: 358/360 (99.44%), 24145 kH/s yes!
[2015-08-24 16:57:55] tid(0x00000515) accepted: 359/361 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:55] tid(0x00000515) accepted: 360/362 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:55] tid(0x00000515) Stratum difficulty set to 4
[2015-08-24 16:57:56] tid(0x00000519) GPU #2: GeForce GTX 750 Ti, 4068 Temp= 59C Fan= 47%
[2015-08-24 16:57:56] tid(0x00000515) accepted: 361/363 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x0000051a) GPU #3: GeForce GTX 750 Ti, 4001 Temp= 59C Fan= 46%
[2015-08-24 16:57:56] tid(0x0000051b) GPU #4: GeForce GTX 750 Ti, 4080 Temp= 62C Fan= 49%
[2015-08-24 16:57:56] tid(0x00000518) GPU #1: GeForce GTX 750 Ti, 3989 Temp= 57C Fan= 45%
[2015-08-24 16:57:56] tid(0x00000515) accepted: 362/364 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x00000515) accepted: 363/365 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x00000515) accepted: 364/366 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x00000515) accepted: 365/367 (99.46%), 24145 kH/s yes!
[2015-08-24 16:57:57] tid(0x00000517) GPU #0: GeForce GTX 750 Ti, 4010 Temp= 61C Fan= 48%
[2015-08-24 16:57:57] tid(0x00000515) accepted: 366/368 (99.46%), 24145 kH/s yes!
[2015-08-24 16:57:58] tid(0x00000519) GPU #2: GeForce GTX 750 Ti, 4092 Temp= 60C Fan= 47%
[2015-08-24 16:57:58] tid(0x00000515) accepted: 367/369 (99.46%), 24146 kH/s yes!
[2015-08-24 16:58:01] tid(0x0000051c) GPU #5: GeForce GTX 750 Ti, 4012 Temp= 61C Fan= 48%
[2015-08-24 16:58:01] tid(0x00000515) accepted: 368/370 (99.46%), 24146 kH/s yes!

when the mining stabalizes in the next 15-30mins - that will show a more accurate figure ...

i like the temp and fan displays too - nice ... why not shorten T=xx F=xx ? ...

what is this new part also - tid? ...

#crysx

[sp_]

hi sp ...

initial results of fresh compile and mining at givemecoins using only -X 15 and --diff .5 with no oc or firmware tweaking ... just stock ...

*** ccminer 1.5.61-git(SP-MOD) for nVidia GPUs by sp-hash@github ***
        Built with the nVidia CUDA SDK 6.5

  Based on pooler cpuminer 2.3.2 and the tpruvot@github fork
   CUDA support by Christian Buchner, Christian H. and DJM34
  Includes optimizations implemented by sp , klaust, tpruvot and tsiv.

[2015-08-24 16:56:18] tid(0x0000047e) XIntensity set to 15, 983040 cuda threads
[2015-08-24 16:56:23] tid(0x00000515) Starting Stratum on stratum+tcp://mining-ace.granitecoin.com:6001/


[2015-08-24 16:57:55] tid(0x00000515) accepted: 358/360 (99.44%), 24145 kH/s yes!
[2015-08-24 16:57:55] tid(0x00000515) accepted: 359/361 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:55] tid(0x00000515) accepted: 360/362 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:55] tid(0x00000515) Stratum difficulty set to 4
[2015-08-24 16:57:56] tid(0x00000519) GPU #2: GeForce GTX 750 Ti, 4068 Temp= 59C Fan= 47%
[2015-08-24 16:57:56] tid(0x00000515) accepted: 361/363 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x0000051a) GPU #3: GeForce GTX 750 Ti, 4001 Temp= 59C Fan= 46%
[2015-08-24 16:57:56] tid(0x0000051b) GPU #4: GeForce GTX 750 Ti, 4080 Temp= 62C Fan= 49%
[2015-08-24 16:57:56] tid(0x00000518) GPU #1: GeForce GTX 750 Ti, 3989 Temp= 57C Fan= 45%
[2015-08-24 16:57:56] tid(0x00000515) accepted: 362/364 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x00000515) accepted: 363/365 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x00000515) accepted: 364/366 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x00000515) accepted: 365/367 (99.46%), 24145 kH/s yes!
[2015-08-24 16:57:57] tid(0x00000517) GPU #0: GeForce GTX 750 Ti, 4010 Temp= 61C Fan= 48%
[2015-08-24 16:57:57] tid(0x00000515) accepted: 366/368 (99.46%), 24145 kH/s yes!
[2015-08-24 16:57:58] tid(0x00000519) GPU #2: GeForce GTX 750 Ti, 4092 Temp= 60C Fan= 47%
[2015-08-24 16:57:58] tid(0x00000515) accepted: 367/369 (99.46%), 24146 kH/s yes!
[2015-08-24 16:58:01] tid(0x0000051c) GPU #5: GeForce GTX 750 Ti, 4012 Temp= 61C Fan= 48%
[2015-08-24 16:58:01] tid(0x00000515) accepted: 368/370 (99.46%), 24146 kH/s yes!

when the mining stabalizes in the next 15-30mins - that will show a more accurate figure ...
i like the temp and fan displays too - nice ... why not shorten T=xx F=xx ? ...
what is this new part also - tid? ...
#crysx

Thread id. I think I should remove it. The temps and fanspeed is only working in the linux version. The changes in the output is written by t-nelson, and I have merged his changes into my fork.

[sp_]

initial results of fresh compile and mining at givemecoins using only -X 15 and --diff .5 with no oc or firmware tweaking ... just stock ...

I don't need to use --diff 0.5 in the latest version. Depends on the pool you are mining. On nicehash or p2pool.pl , run without the --diff parameter.

[chrysophylax]

initial results of fresh compile and mining at givemecoins using only -X 15 and --diff .5 with no oc or firmware tweaking ... just stock ...

I don't need to use --diff 0.5 in the latest version. Depends on the pool you are mining. On nicehash or p2pool.pl , run without the --diff parameter.

with giveme coins - it still requires it ...

they havent changed anything in their stratum code as far as i know ...

here is the latest ...

-------

[2015-08-24 17:29:24] tid(0x0000051c) GPU #5: GeForce GTX 750 Ti, 4012 Temp= 67C Fan= 54%
[2015-08-24 17:29:24] tid(0x00000517) GPU #0: GeForce GTX 750 Ti, 4010 Temp= 68C Fan= 56%
[2015-08-24 17:29:24] tid(0x0000051a) GPU #3: GeForce GTX 750 Ti, 4009 Temp= 67C Fan= 54%
[2015-08-24 17:29:24] tid(0x0000051b) GPU #4: GeForce GTX 750 Ti, 4079 Temp= 71C Fan= 57%
[2015-08-24 17:29:24] tid(0x00000518) GPU #1: GeForce GTX 750 Ti, 3998 Temp= 65C Fan= 52%
[2015-08-24 17:29:24] tid(0x00000519) GPU #2: GeForce GTX 750 Ti, 4090 Temp= 68C Fan= 55%
[2015-08-24 17:29:25] tid(0x0000051b) GPU #4: GeForce GTX 750 Ti, 4075 Temp= 71C Fan= 57%
[2015-08-24 17:29:26] tid(0x00000515) accepted: 686/689 (99.56%), 24196 kH/s yes!
[2015-08-24 17:29:27] tid(0x00000517) GPU #0: GeForce GTX 750 Ti, 4012 Temp= 69C Fan= 56%
[2015-08-24 17:29:27] tid(0x00000515) accepted: 687/690 (99.57%), 24196 kH/s yes!
[2015-08-24 17:29:28] tid(0x00000517) GPU #0: GeForce GTX 750 Ti, 4006 Temp= 68C Fan= 56%
[2015-08-24 17:29:28] tid(0x00000515) accepted: 688/691 (99.57%), 24196 kH/s yes!
[2015-08-24 17:29:35] tid(0x00000517) GPU #0: GeForce GTX 750 Ti, 4011 Temp= 69C Fan= 56%
[2015-08-24 17:29:35] tid(0x00000515) accepted: 689/692 (99.57%), 24196 kH/s yes!
[2015-08-24 17:29:42] tid(0x00000517) GPU #0: GeForce GTX 750 Ti, 4010 Temp= 69C Fan= 56%
[2015-08-24 17:29:42] tid(0x00000515) accepted: 690/693 (99.57%), 24196 kH/s yes!
[2015-08-24 17:29:48] tid(0x00000517) GPU #0: GeForce GTX 750 Ti, 4011 Temp= 69C Fan= 56%
[2015-08-24 17:29:49] tid(0x0000051a) GPU #3: GeForce GTX 750 Ti, 4010 Temp= 67C Fan= 54%
[2015-08-24 17:29:49] tid(0x00000518) GPU #1: GeForce GTX 750 Ti, 3998 Temp= 65C Fan= 52%
[2015-08-24 17:29:49] tid(0x0000051c) GPU #5: GeForce GTX 750 Ti, 4011 Temp= 66C Fan= 54%
[2015-08-24 17:29:49] tid(0x00000515) accepted: 691/694 (99.57%), 24196 kH/s yes!

-------

so its settling just under the 24200 mark ... a little bit more than it would have usually done at the same level with the previous v60 ...

giveme coins is averaging about 23000 mark for this one machine ... though it is very difficult to get any stable reading from the pool ...

#crysx

[chrysophylax]

hi sp ...

initial results of fresh compile and mining at givemecoins using only -X 15 and --diff .5 with no oc or firmware tweaking ... just stock ...

*** ccminer 1.5.61-git(SP-MOD) for nVidia GPUs by sp-hash@github ***
        Built with the nVidia CUDA SDK 6.5

  Based on pooler cpuminer 2.3.2 and the tpruvot@github fork
   CUDA support by Christian Buchner, Christian H. and DJM34
  Includes optimizations implemented by sp , klaust, tpruvot and tsiv.

[2015-08-24 16:56:18] tid(0x0000047e) XIntensity set to 15, 983040 cuda threads
[2015-08-24 16:56:23] tid(0x00000515) Starting Stratum on stratum+tcp://mining-ace.granitecoin.com:6001/


[2015-08-24 16:57:55] tid(0x00000515) accepted: 358/360 (99.44%), 24145 kH/s yes!
[2015-08-24 16:57:55] tid(0x00000515) accepted: 359/361 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:55] tid(0x00000515) accepted: 360/362 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:55] tid(0x00000515) Stratum difficulty set to 4
[2015-08-24 16:57:56] tid(0x00000519) GPU #2: GeForce GTX 750 Ti, 4068 Temp= 59C Fan= 47%
[2015-08-24 16:57:56] tid(0x00000515) accepted: 361/363 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x0000051a) GPU #3: GeForce GTX 750 Ti, 4001 Temp= 59C Fan= 46%
[2015-08-24 16:57:56] tid(0x0000051b) GPU #4: GeForce GTX 750 Ti, 4080 Temp= 62C Fan= 49%
[2015-08-24 16:57:56] tid(0x00000518) GPU #1: GeForce GTX 750 Ti, 3989 Temp= 57C Fan= 45%
[2015-08-24 16:57:56] tid(0x00000515) accepted: 362/364 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x00000515) accepted: 363/365 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x00000515) accepted: 364/366 (99.45%), 24145 kH/s yes!
[2015-08-24 16:57:56] tid(0x00000515) accepted: 365/367 (99.46%), 24145 kH/s yes!
[2015-08-24 16:57:57] tid(0x00000517) GPU #0: GeForce GTX 750 Ti, 4010 Temp= 61C Fan= 48%
[2015-08-24 16:57:57] tid(0x00000515) accepted: 366/368 (99.46%), 24145 kH/s yes!
[2015-08-24 16:57:58] tid(0x00000519) GPU #2: GeForce GTX 750 Ti, 4092 Temp= 60C Fan= 47%
[2015-08-24 16:57:58] tid(0x00000515) accepted: 367/369 (99.46%), 24146 kH/s yes!
[2015-08-24 16:58:01] tid(0x0000051c) GPU #5: GeForce GTX 750 Ti, 4012 Temp= 61C Fan= 48%
[2015-08-24 16:58:01] tid(0x00000515) accepted: 368/370 (99.46%), 24146 kH/s yes!

when the mining stabalizes in the next 15-30mins - that will show a more accurate figure ...
i like the temp and fan displays too - nice ... why not shorten T=xx F=xx ? ...
what is this new part also - tid? ...
#crysx

Thread id. I think I should remove it. The temps and fanspeed is only working in the linux version. The changes in the output is written by t-nelson, and I have merged his changes into my fork.

i think you should also ...

if the thread id is useful to the general public - then leave it ... but if we cant use that info as miners ( or general public ) as info for the miners - then it doesnt need to be there in my opinion mate ...

and temp / fan in linux only? ... that would be a first for linux to have wouldnt it? ... nice merge in my opinion ... ...

#crysx

[Genoil]

...

And thanks for your continued support of ccminer. Not sure if it's relevant as you don't really seem interested, but the low CPU overhead changes you made to ethminer reduce hashrate by like 20% when you have 2+ GPUs.

https://forum.ethereum.org/discussion/2227/cuda-miner/p8

The low CPU overhead changes by sp_ didn't make it eventually into latest build in favor of tpruvot's alternative. Only sp_'s keccak optimization is still in, although it doesn't really make a difference.

As for the original reason Ethereum was referenced in this thread; I know somebody is building stratum support for ethminer. I was approached for the job last week but declined, so they found somebody else.

[sp_]

and temp / fan in linux only? ... that would be a first for linux to have wouldnt it? ... nice merge in my opinion ... ...
#crysx

It's in the windows version as well if you build 64bit and not x86.
The reason is that NVIDIA has removed the support for the API calls in the 32bit application.
On windows the 64bit builds are hashing slower than the 32 bit builds.