CCminer(SP-MOD) Modded GPU kernels.

[sp_]

sp did you optimize x11 in release 9
I've jumped to 7700 khash now (with 200 MHZ OC) on gtx970

You should get higher rates, try the ccminer52.exe file.

x11, x13 and x15 should be faster in release9.

[sp_]

This version has a -i parameter wich was added by tpruvot. On the 970 and 980 cards might get a boost if you change to higher intensity. -i 20 -i 21 -i 22?

[jpouza]

jpouza: From the picture I see that you get a couple of does not validate on the cpu. Did you clock it too high? 9600 KHASH is i monster hashrate btw

Need to check the  "does not validate" , seams to work fine on the 750ti. This might be a problem with the ccminer52.exe file. Is the ccminer.exe performing the same as the 52 version?

I've slowed down the clock and everything is fine now, still get 9,2-9,4 MH/s @ X11.

I will test the intensity (-i) parameter later.

Cheers

[jpouza]

-i parameter does not work in the command line of the miner...

About the mining I will keep my hardware cool at 55-60ºC max. temps, this now I can get with 9MH/s and lower clocks with stock voltages, the effort to push the OC to get 9,3-9,5MH/s I think it´s not worth at all, better stay in the safe ranges with a little (about 1MH/s with my 3x980) less hash power

Cheers

[alexbg21]

Palit GTX 970 - X11 - Latest 9 release by SP_ : 7900 Mhs

[SavellM]

So for Maxwell which would be better to mine on SP fork or tpruvot fork?

Sorry just getting back into this.

I'm mining X11 if that helps.

[sp_]

My fork is currently around 50KHASH faster on the 750ti than the tpruvot fork @x11. But tpruvot will probobly merge my kernal changes into his fork, and then they will perform the same.
This is the power of opensource.

[SavellM]

So what would the benefits of either fork?

tpruvot just supports more algos?

[Bombadil]

So what would the benefits of either fork?

tpruvot just supports more algos?

Tpruvot also makes sure the linux version is working correctly.
There is no real benefit, sp_ is continously working on hashrate improvements, while tpruvot is making side-features like the api & extranonce support (and also speed improv).

You guys should work on the same codebase, IMHO

[sp_]

My fork is missing some of his bugfixes, and his fork is missing some optimalizations. This is software under development, and the exe files are beta versions that contain bugs.

Note that ccminer has become 30% faster than in september for the new maxwell cards. Most of the kernal improvements are done by me, but some by tpruvot, tsiv and klaus_t.

To follow the development and the contributors check github:

https://github.com/sp-hash/ccminer/network


And please donate.. We work hard to give you free Hashpower.


Old performance article:

http://cryptomining-blog.com/3503-crypto-mining-performance-of-the-new-nvidia-geforce-gtx-980/

[SavellM]

Cheers sp_ I will

I appreciate all your hard work especially on new architecture (Maxwell).

[Epsylon3]

the fix is here : https://github.com/tpruvot/ccminer/commit/d7c13b061162f7382ee7b4ebd7acf2f34ba4a2a2

yaamp ignores the new request id 3 and send a global one (id=null) which was not handled

[SavellM]

sp_ will you incorporate this fix yourself?

As I also mine on Yaamp.com so this would help out greatly.

[sp_]

thanks. I checked in the fix.
I will checkin some more kernal improvements (x13).

[sp_]

New build from github

http://www.filedropper.com/release10

Contains 2 bugfixes from the tpruvot (stratum disconnects)
X13 and x15 is a bit faster. (hamsi+8,3%)
fixed linuxbuild (untested)

The sourcecode is available here:

https://github.com/sp-hash/ccminer

[SavellM]

Thanks sp_, I was just LITERALLY asking for a new build!

You rock!

[ryen123]

Release 10 works perfectly. No problems whatsoever.

Hashrates (750Ti O/C)
X11: 3000 kH/s
X13: 2320 kH/s
X15: 2000 kH/s

[jpouza]

Release 10 works perfectly. No problems whatsoever.

Hashrates (750Ti O/C)
X11: 3000 kH/s
X13: 2320 kH/s
X15: 2000 kH/s

What are your clocks with the 750Ti?

I can get only 2950-2970KH/s with 1246MHz Core / 1580MHz Memory

[Epsylon3]

i found another side effect of the extranonce subscribe (its the possible 0khs accept on start)...

Stratum Request id was not tested on share answers (it was tested nowhere in fact before my recent changes)

Just commited

Code:

[2014-11-24 19:34:42] < {"id":3,"result":true,"error":null}
[2014-11-24 19:34:42] accepted: 1/1 (100.00%), 0.00 khash/s yay!!!

[ryen123]

Release 10 works perfectly. No problems whatsoever.

Hashrates (750Ti O/C)
X11: 3000 kH/s
X13: 2320 kH/s
X15: 2000 kH/s

What are your clocks with the 750Ti?

I can get only 2950-2970KH/s with 1246MHz Core / 1580MHz Memory

1210MHz Core / 1600MHz Memory