Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 943. (Read 2347664 times)

hehehe ... aaand you wonder why i wont change from linux? ...

and its a damn site easier to compile stuff too ...

which distro did you use? ... redhat based or debian based? ...

#crysx

no problem mate ...

apologies for confusing you earlier - as i misread your question ...

#crysx

I just switched my desktop (the one with the GTX970 G1 Gaming) from windows to linux and got a very nice surprise: on windows, when mining, the mouse is lagging and the machine is pretty much unusable. On linux, no such thing! The machine is happily mining Quark at 17MH while I watch a 1080p video on Youtube.

What is this witchery? Not that I'm complaining...

BTW, it's about time for my donation, so here's a beer SP: 3bb1e5e50d3967cfc8f5e14aa700c316fc69c45380fcad938aed2d67fc62615c

Thanks man. know it's clear.

ok- i see that i misread about x15 ... but the rules of engagement are the same when it comes to the various algos ...

xintensity -X 28 works for MY cards ... you need to adjust the intensity settings to suit your own cards ...

if you are getting a 'scanhash' error - this 'usually' means that the card has run out of available memory for the process to happen ... drop the -X to 27 ( for example ) ...

this may or may not work well for your card - if you still get those errors - the keep droppign it until it processes the shares ...

the best way to do this is to run ccminer-spmod without ANY intensity settings and see if all is ok with the default settings built in to the card ... THEN start to adjust using -X ...

the -X is NOT static - as it changes with the algo change ... some algos use a higher -X setting - some lower ...

believe it or not - compiling ccminer in cuda 7 actually allows me to up the intensity to -X 29 ... but the hashrate is appalling ...

btw - i get about 2.8MH per card also in x11 - but i DO NOT oc - like many others here including sp ... you are bound to get a lot more hashrate if you adjust clock settings also ...

#crysx

seems like misunderstanding.
i also don't mine x15 algo.  it's about XIntensity
You say that for 750 Ti card -X 28 is the best, but with that command
./ccminer -a x11 -o stratum+tcp://drk.ghash.io:3333 -u** -p ** -X 28
i receive
so i can use only default Intensity which is hardcoded to miner. but with such default config i can get only 2.8mh  seems like to low.

all algos use different settings ...

sp ( as with the othe devs in their fork of ccminer ) has built in intensity settings for each algo ... what they are - well only sp knows .. which is why i want to collate a full document on all settings - not just intensity ...

with x11 - i use -X 28 and that seems to be fine ... at the max hashrate the cards can go ...

with x15 - im not sure as i never mine with x15 ...

you need to change the settings for every algo and see which ones come out best ...

#crysx

Thanks for reply.
checked -X 15 in x11 algo and seems like it even more slow than without it.

according to this forum with stock kernal i have to receive 3+mh in x11.
Any advices, what config i have to use?

wow - what a difference going DOWN in intensity on quark ...

-------

[2015-09-17 17:37:13] accepted: 409/414 (98.79%), 35567 kH/s yes!
[2015-09-17 17:37:25] GPU #0: GeForce GTX 750 Ti, 5935 Temp= 79C Fan= 65%
[2015-09-17 17:37:25] accepted: 410/415 (98.80%), 35571 kH/s yes!
[2015-09-17 17:37:33] GPU #4: GeForce GTX 750 Ti, 5960 Temp= 80C Fan= 65%
[2015-09-17 17:37:34] accepted: 411/416 (98.80%), 35574 kH/s yes!
[2015-09-17 17:37:39] GPU #1: GeForce GTX 750 Ti, 5936 Temp= 76C Fan= 62%
[2015-09-17 17:37:40] GPU #2: GeForce GTX 750 Ti, 6031 Temp= 78C Fan= 63%
[2015-09-17 17:37:40] GPU #3: GeForce GTX 750 Ti, 5924 Temp= 79C Fan= 64%
[2015-09-17 17:37:40] GPU #2: GeForce GTX 750 Ti, 6008 Temp= 78C Fan= 63%
[2015-09-17 17:37:41] accepted: 412/417 (98.80%), 35574 kH/s yes!
[2015-09-17 17:37:42] GPU #5: GeForce GTX 750 Ti, 5877 Temp= 78C Fan= 63%
[2015-09-17 17:37:43] quark.eu.nicehash.com:3345 quark block 2190162

-------

thats huge ...

#crysx

in v67 - lyra2v2 is broken ... you need to use the previous version ...

the intensity levels that you are using are the ones we are trying for quark - not lyra2v2 ...

it seems the best intensity levels for lyra2v2 is using the xintensity -X 15 ... at least - for our cards ( gigabyte 750ti oc lp ) ...

and the 'memory' error is due to the memory allocation ON the card ... not the computer ...

when you set an intensity value - it uses the on-board memory of the video card and max's the memory when increasing the the intensity too much ( increasing the cuda threads it uses to process data inside the video cards memory ) ...

the devs can explain that side a hell of a lot better than i can ...

#crysx

Seems like lyra2rev2 not working.
finally fixed all issues and no accepted shares:




also maybe you can advice, i try to add "-i 22.9" or "-i 24" error appeared:


p.s. not sure what exactly "memory" used but my pc has 32gb ram

I use .67 A good result for the GTH 960.





Ps. I do not know how correctly sent to the creator of the beer, let it be income for the day. For _SP transaction ID
7d52b98dc6448bdfdcd64fafc3fffb20a2d641a8e867ac591878cdd4e2b298bc

no - will try that now ...

tanx ...

#crysx

did you try -i 22.9?

the cards are gigabyte 750ti oc lp ... the latter one is c6.5 ... no oc - no intensity increase ...

what would be the optimum intensity setting? ...

initially i used -i 24 ... and that was stable ...

-------

[2015-09-17 16:40:57] accepted: 387/395 (97.97%), 35161 kH/s yes!
[2015-09-17 16:41:08] GPU #0: GeForce GTX 750 Ti, 5835 Temp= 80C Fan= 65%
[2015-09-17 16:41:09] accepted: 388/396 (97.98%), 35159 kH/s yes!
[2015-09-17 16:41:20] quark.eu.nicehash.com:3345 quark block 2190008

-------

current hashrate using just the settings earlier ...

#crysx

Don't use cuda 7.0.
Cuda 7.5 should work, but slower than 6.5

If those are gigabyte oc cards, they will do 6150 in x86 windows and cuda 6.5 (+200khash)

there are a LOT of cpu validattion errors on the one that is compiled with c7 and quark ...

c7 x11 has no issue running - but much less hashrate than the one compiled with c6.5 ...

i cant compile with c7.5 at all - i dont have a spare machine to build ... or time for that matter ...

#crysx

working ...

f22x64 c7 ...

*** ccminer 1.5.67-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, tsiv and pallas.

[2015-09-17 16:08:39] Starting Stratum on stratum+tcp://quark.eu.nicehash.com:3345
[2015-09-17 16:08:39] NVML GPU monitoring enabled.
[2015-09-17 16:08:39] Binding thread 1 to cpu 1 (mask 2)
[2015-09-17 16:08:39] Binding thread 2 to cpu 2 (mask 4)
[2015-09-17 16:08:39] 5 miner threads started, using 'quark' algorithm.
[2015-09-17 16:08:39] Binding thread 3 to cpu 3 (mask
[2015-09-17 16:08:39] Binding thread 0 to cpu 0 (mask 1)
[2015-09-17 16:08:39] Binding thread 4 to cpu 0 (mask 10)
[2015-09-17 16:08:40] Stratum difficulty set to 0.04
[2015-09-17 16:08:41] quark.eu.nicehash.com:3345 quark block 2189889
[2015-09-17 16:08:46] accepted: 1/1 (100.00%), 2277 kH/s yes!
[2015-09-17 16:08:53] quark.eu.nicehash.com:3345 quark block 2189890
[2015-09-17 16:08:53] GPU #1: GeForce GTX 750 Ti, 5735 Temp= 55C Fan= 37%
[2015-09-17 16:08:53] accepted: 2/2 (100.00%), 23505 kH/s yes!
[2015-09-17 16:08:57] GPU #0: result for nonce $023D4C7C does not validate on CPU!
[2015-09-17 16:09:09] GPU #2: GeForce GTX 750 Ti, 5661 Temp= 51C Fan= 36%
[2015-09-17 16:09:10] accepted: 3/3 (100.00%), 23932 kH/s yes!
[2015-09-17 16:09:10] accepted: 3/4 (75.00%), 23932 kH/s nooooo
[2015-09-17 16:09:10] reject reason: Share above target.
[2015-09-17 16:09:12] quark.eu.nicehash.com:3345 quark block 2189891
[2015-09-17 16:09:12] GPU #4: GeForce GTX 750 Ti, 5664 Temp= 54C Fan= 37%
[2015-09-17 16:09:12] GPU #3: GeForce GTX 750 Ti, 5657 Temp= 58C Fan= 38%
[2015-09-17 16:09:12] GPU #2: GeForce GTX 750 Ti, 5663 Temp= 52C Fan= 36%
[2015-09-17 16:09:12] GPU #0: GeForce GTX 750 Ti, 5616 Temp= 56C Fan= 38%
[2015-09-17 16:09:12] GPU #1: GeForce GTX 750 Ti, 5745 Temp= 58C Fan= 38%

-------

f20x64 c6.5

*** ccminer 1.5.67-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, tsiv and pallas.

[2015-09-17 16:08:15] Starting Stratum on stratum+tcp://quark.eu.nicehash.com:3345
[2015-09-17 16:08:16] NVML GPU monitoring enabled.
[2015-09-17 16:08:16] Binding thread 0 to cpu 0 (mask 1)
[2015-09-17 16:08:16] 6 miner threads started, using 'quark' algorithm.
[2015-09-17 16:08:16] Binding thread 5 to cpu 1 (mask 20)
[2015-09-17 16:08:16] Binding thread 4 to cpu 0 (mask 10)
[2015-09-17 16:08:16] Binding thread 3 to cpu 1 (mask
[2015-09-17 16:08:16] Binding thread 2 to cpu 0 (mask 4)
[2015-09-17 16:08:16] Binding thread 1 to cpu 1 (mask 2)
[2015-09-17 16:08:17] Stratum difficulty set to 0.04
[2015-09-17 16:08:17] quark.eu.nicehash.com:3345 quark block 2189887
[2015-09-17 16:08:24] quark.eu.nicehash.com:3345 quark block 2189888
[2015-09-17 16:08:31] quark.eu.nicehash.com:3345 quark block 2189889
[2015-09-17 16:08:31] GPU #3: GeForce GTX 750 Ti, 5880 Temp= 63C Fan= 49%
[2015-09-17 16:08:32] GPU #4: GeForce GTX 750 Ti, 5940 Temp= 65C Fan= 51%
[2015-09-17 16:08:32] GPU #2: GeForce GTX 750 Ti, 5960 Temp= 62C Fan= 49%
[2015-09-17 16:08:32] GPU #0: GeForce GTX 750 Ti, 5945 Temp= 65C Fan= 51%
[2015-09-17 16:08:33] GPU #1: GeForce GTX 750 Ti, 1695 Temp= 52C Fan= 46%
[2015-09-17 16:08:33] GPU #5: GeForce GTX 750 Ti, 2127 Temp= 55C Fan= 48%
[2015-09-17 16:08:34] accepted: 0/1 (0.00%), 28628 kH/s nooooo
[2015-09-17 16:08:34] reject reason: Job not found.
[2015-09-17 16:08:34] accepted: 0/2 (0.00%), 28628 kH/s nooooo
[2015-09-17 16:08:34] reject reason: Job not found.
[2015-09-17 16:08:39] GPU #5: GeForce GTX 750 Ti, 5878 Temp= 63C Fan= 49%
[2015-09-17 16:08:39] accepted: 1/3 (33.33%), 29517 kH/s yes!

-------

so its working with quark and x11 but not with lyra2v2 ...

#crysx

sorry - that was my fault - i didnt change the port :| ...

trying again ...

#crysx

Works on windows.. Try to do a full clean, and a full rebuild.

Seems to be some rejects. 100%

Is the --benchmark mode showing "does not validate on the cpu?"