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

[sp_]

sp work is not always easy to reuse, i mostly search the best performance on linux and that doesnt always match with windows :p
But i can generally reuse a part of its work, or at least i can see where there are rooms of improvement.

I recently tried the most profitable algos (quark, x11, neoscrypt) in your fork compared to sp_'s and they are significantly slower.
I mine on linux only, with compute 5.2 cards and cuda 6.5.
I made the latest 10% enhancement over his neoscrypt kernel (which is actually djm34 code): I think you can safely use it on your fork, the only caveat is that it doesn't include compute <5.0 code.

It's DJM34's code with modding. 7-10% faster than the first opensource version. You added another 10% so then it's 20% faster than the first version. If people choose to publish their work, other people will help to make it faster. I prefer opensource, but I understand why people keep it private.

Here is my first improvement (neoscrypt +7%):



https://github.com/sp-hash/ccminer/commit/48072790339fc021f096e20765ed8cd03701c155

[kenshirothefist]

It's really great to see constant development and open-source engagement in ccminer! Here are some tips for the devs (already sent tip to tpruvot a few days ago):

pallas: https://blockchain.info/tx/538749267db068bb5a0417363087c60f59aba890ed22a3a83d93eab7e73c81ca
sp_ https://blockchain.info/tx/76597bd418e5f8306fe65a9405199a8b7def0327c3ad16e54ac905891c781549
djm34: https://blockchain.info/tx/e33fc9cd80d41c3bcaf2f02c6e60e88647401d600ef4ba633a71dcf043d8e0ff

Keep up the good work and try to keep this project opensource and as much as possible in a single fork - it's much easier for the users this way...

Best regards,
kenshirothefist

[pallas]

It's really great to see constant development and open-source engagement in ccminer! Here are some tips for the devs (already sent tip to tpruvot a few days ago):

pallas: https://blockchain.info/tx/538749267db068bb5a0417363087c60f59aba890ed22a3a83d93eab7e73c81ca
sp_ https://blockchain.info/tx/76597bd418e5f8306fe65a9405199a8b7def0327c3ad16e54ac905891c781549
djm34: https://blockchain.info/tx/e33fc9cd80d41c3bcaf2f02c6e60e88647401d600ef4ba633a71dcf043d8e0ff

Keep up the good work and try to keep this project opensource and as much as possible in a single fork - it's much easier for the users this way...

Best regards,
kenshirothefist

Thanks, much appreciated! :-)

[TheFiendishOne]

@epsylon3 per the official websites mining page https://www.dashpay.io/mining/

I downloaded the ccminer optimized for GTX 750 Ti and AVG is picking up that ccminer as a threat "ADP.ARES.Generic"

I want to assume this is a false positive, but figured I'd get some feedback first?  Please get back to me. 

[Epsylon3]

i guess it is...

https://www.virustotal.com/fr/file/def08b2e8d5dc3dbf21dfea6e1eec17718ea9a1ffb204063de68a3cece1f5d93/analysis/

https://www.virustotal.com/fr/file/166483ed270b2ed9f690eb86ed2bed5b4979e0468dafc71eb6f2075d83677c1b/analysis/

or its only on your machine

[djm34]

It's really great to see constant development and open-source engagement in ccminer! Here are some tips for the devs (already sent tip to tpruvot a few days ago):

pallas: https://blockchain.info/tx/538749267db068bb5a0417363087c60f59aba890ed22a3a83d93eab7e73c81ca
sp_ https://blockchain.info/tx/76597bd418e5f8306fe65a9405199a8b7def0327c3ad16e54ac905891c781549
djm34: https://blockchain.info/tx/e33fc9cd80d41c3bcaf2f02c6e60e88647401d600ef4ba633a71dcf043d8e0ff

Keep up the good work and try to keep this project opensource and as much as possible in a single fork - it's much easier for the users this way...

Best regards,
kenshirothefist

thanks a lot

[sp_]

It's really great to see constant development and open-source engagement in ccminer! Here are some tips for the devs (already sent tip to tpruvot a few days ago):
pallas: https://blockchain.info/tx/538749267db068bb5a0417363087c60f59aba890ed22a3a83d93eab7e73c81ca
sp_ https://blockchain.info/tx/76597bd418e5f8306fe65a9405199a8b7def0327c3ad16e54ac905891c781549
djm34: https://blockchain.info/tx/e33fc9cd80d41c3bcaf2f02c6e60e88647401d600ef4ba633a71dcf043d8e0ff
Keep up the good work and try to keep this project opensource and as much as possible in a single fork - it's much easier for the users this way...
Best regards,
kenshirothefist

Thanks alot.

A single fork is not that easy since my fork only support compute 5.0 and up devices:

here is the thread:

https://bitcointalk.org/index.php?topic=826901.0

1.5.63(sp-MOD) is available here: (31-08-2015)

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

The sourcecode is available here:

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

[crispycracker]

Good day all.
Been trying to assemble your latest (ccminer 1.6.6) for a Kepler (GTX 690), (compute 3.5) .
Getting the following error :
'keccak256_cpu_init' at line 299 : invalid device symbol.
Any advice ?  Or is it simply a Maxwell only scypt?

[sp_]

Good day all.
Been trying to assemble your latest (ccminer 1.6.6) for a Kepler (GTX 690), (compute 3.5) .
Getting the following error :
'keccak256_cpu_init' at line 299 : invalid device symbol.
Any advice ?  Or is it simply a Maxwell only scypt?

you need to add compute 3.5 in the projectfile/makefile.

[pallas]

Hi tpruvot,

I've seen a ton of commits and that you updated the linux branch, should we switch to it? What are the news?

Thanks a lot!

[Epsylon3]

i just have updated the cuda 7.5 branch after a windows test... no improvements since the RC

[ngdias]

Is ccminer dropping support for 3.0 cards? Newer versions of the miner don't seem to work, but I also couldn't find anything explicit in readme files.

[Epsylon3]

which one did you test / are failing ?

[sp_]

i just have updated the cuda 7.5 branch after a windows test... no improvements since the RC

I removed all the max limit registers and most of the launchbounds, it helped a bit, but still slower than 6.5

[ngdias]

which one did you test / are failing ?

tpruvot's or djm34. maybe both? Lost support around the time Vertcoin forked or a bit before that.

[Epsylon3]

super precise answer, well better than nothing i will pass...

[antimony121]

the latest version off github (1.6.6dev) doesn't seems to quit properly after setting --retries=0

Code:

[2015-09-24 17:47:31] Stratum authentication failed
[2015-09-24 17:47:31] ...terminating workio thread

and it just get stuck there instead of quitting.

[Epsylon3]

thx for the report, will try that...

edit: indeed, fixed in linux branch : https://github.com/tpruvot/ccminer/commit/41707760f560234b20485555560a547c28d10289

[antimony121]

Thanks for the very quick fix! 

thx for the report, will try that...

edit: indeed, fixed in linux branch : https://github.com/tpruvot/ccminer/commit/41707760f560234b20485555560a547c28d10289

[antimony121]

hello, not sure if it is an related issue, but when the network timeouts (when my connection gets dropped..) it doesn't terminate properly

it would segmentation fault a few times under my script and quit, but eventually it just get stuck there

Code:

[2015-10-01 18:48:59] Stratum connection failed: Could not resolve host: keccak.hk.nicehash.com
[2015-10-01 18:48:59] ...terminating workio thread

when i press ctrl-c it still doesn't quit

Code:

^C[2015-10-01 18:58:08] SIGINT received, exiting

still stuck there and not taking me back to the command prompt

Thanks for the very quick fix! 

thx for the report, will try that...

edit: indeed, fixed in linux branch : https://github.com/tpruvot/ccminer/commit/41707760f560234b20485555560a547c28d10289