CCminer(SP-MOD) Modded GPU kernels.

[sp_]

Sp just to see... where is 1.7.6 plz  I don't see it on github.
750ti oc 1345 core  7 mh quark

1.7.6 is only if you build yourself (the latest open sourcecode on github)

Release 1.7.5(19 mar 2016) windows binaries is here:

https://github.com/tpruvot/ccminer/releases

Why don't you run it and report your hashrate.

The last time I did a comparission with the 1.7.5 windows binary and the sp-mod private #5 I got these results:

Sp-Mod #5 default intensities hashrates in MHASH/s. (980ti 1290 coreclock/1650 memclock)

nist5: 43.80 (+17%)
quark: 27.85 (+61,91%)
lyra2v2 16.10 (+16.6%)
x11:   12.85 (+18%)
x13:   10.27 (+19.97)
x15:    8.83 (+18.52 )
neoscrypt: 0.99 (+73.68%)

[fenomenhaa]

Sp mod 5 working stable in my rigs without any problem. i'm sure about that.(Celeron rigs too)But about performance some performance lost according to advertised from Sp.Esp. neoscyrpt.But they are still faster from opencource ones..

[Amph]

there is a way to change back the p state from 0 to 2 without using nvidia inspector? with command prompt for example?

[tbearhere]

there is a way to change back the p state from 0 to 2 without using nvidia inspector? with command prompt for example?

When I reboot my computer mine go back to the p2 state Amph.  980ti and 970 gtx.

[bathrobehero]

there is a way to change back the p state from 0 to 2 without using nvidia inspector? with command prompt for example?

Inspector works from command line as well so you can cook up a bat that runs your preferred settings on boot. I haven't tried it though.

[sp_]

I submitted a bugfix (less duplicate shares) and a small skein optimalization@github

[Longsnowsm]

SP_ I know your working on the multi algo miner, but I hope that doesn't mean the single algo ccminer is going away...  I have looked at the Claymore miner and I personally don't want to go that direction where I am overdriving my GPU's for a little more hashing.  So I hope the standard option will be available.  I am about to stand up my Nvidia rig and will be looking to get one of your private versions of the ccminer for that rig.  I appreciate all your doing to make the fastest optimized miner for the nvidia gpu's!

[tbearhere]

1. What is your commandline parameters.
2. What is your operation system.
3. How much memory do you have.
4. Wich cards are connected

My miner software  sp-mod #5 is working on your 4core rig.
You need to troubleshoot on your own why it doesn't work on the 2 core rig.

Try --cpu-affinity 0 to enable the first core only.
Try --cpu-priority 0 to give the minerless cpu and the system more resources.
Try -C (cpu-mining) to avoid blocking sync call.

Are you overclocking your cards? remove the oc.
Do you use powered risers?
Is your PSU big enough?
Do you have a computer virus on your pc?

Just did a test run with that and cpu usage went down great... but quark is a real problem. For some reason quark with crash after 3 minutes of hashing any where and any way All other algo's are ok.
EDIT: I forgot to try the  -C  
The -C brought the cpu up to 100% but finally its stable. Ya  
EDIT: Even lyra2v2 and neoscrypt smoothed out. Wow finally... Ya

[tbearhere]

there is a way to change back the p state from 0 to 2 without using nvidia inspector? with command prompt for example?

Inspector works from command line as well so you can cook up a bat that runs your preferred settings on boot. I haven't tried it though.

For my one rig it would have to do this...

cd C:\Program Files\NVIDIA Corporation\NVSMI
nvidia-smi -ac 3505,1455 -i 0,1,2,4

It would be nice to have that work on boot-up or built into the bat or conf file.

[kahir]

I'm really excited for the new upcoming NVidia lineup of cards coming in the next months the specs change is huge developers saying gain in performance is really big I wonder how can that reflect on the mining performance .... we will see 200% gains 

[sp_]

I'm really excited for the new upcoming NVidia lineup of cards coming in the next months the specs change is huge developers saying gain in performance is really big I wonder how can that reflect on the mining performance .... we will see 200% gains 

Etherum needs fast random access.. DDR ram is not fast in random access.

[sp_]

SP_ I know your working on the multi algo miner, but I hope that doesn't mean the single algo ccminer is going away...  I have looked at the Claymore miner and I personally don't want to go that direction where I am overdriving my GPU's for a little more hashing.  So I hope the standard option will be available.  I am about to stand up my Nvidia rig and will be looking to get one of your private versions of the ccminer for that rig.  I appreciate all your doing to make the fastest optimized miner for the nvidia gpu's!

The reason why we do dual mining is that the etherum hashrate on the gtx 970 is almost the same as the 980ti. While the gpu is waiting for the memory accesses to finish, the 2700cores in the 980ti are doing almost nothing. But if the cores are solving another POW algorithm, your return can increase with 25% without loosing any Etherum hashrate.
Dual mining works best for the highend cards. Gtx 970,980,980ti,1070,1080.

[Amph]

there is a way to change back the p state from 0 to 2 without using nvidia inspector? with command prompt for example?

When I reboot my computer mine go back to the p2 state Amph.  980ti and 970 gtx.

yes i know, but i want to avoid a reboot each tiem i need to revert it back, so iw as searching for something more fast

[fenomenhaa]

@Sp private 5 don,'t work on x14 algo, tried private 4 and built 78 also all not working,tprovot miner works well on this algo, pls fix private revions for x14 algo.Thx...

[bensam1231]

for gods sake, remove overclock and verify the problem persists :/
1345 is way above normal clock for 750Ti

No it is not.. I have a 750ti that at stock clocks is 1345 core.
You can oc any gigabyte 750ti to 1345 and it stays stable.. been doing it for 2 years and that is nvidias 750ti  sweet spot.. but don't go over that.

And without any oc'ing on my 2 core rig it sp5 pr still crashes. It is in the build so. Trying  the cpu commands.

Same cards from same manufacturers WILL NOT always clock the same. Silicon lottery is random. You can have bum cards from the same manufacturer that are the same make and model. Even if they're somewhat around the same dispersion, they generally wont be identical and you can always get a lemon that doesn't clock anywhere close to the others.

Don't go throwing around experience like you can't be wrong. That's a dangerous thing to do...


And Tbearhere if you never reboot your machines because you don't want to revert your memory clocks you'll run into problems after a GPU driver soft reset/lock with miners... not that you ever do that I'm sure.

[sp_]

@Sp private 5 don,'t work on x14 algo, tried private 4 and built 78 also all not working,tprovot miner works well on this algo, pls fix private revions for x14 algo.Thx...

x14 and x17 was broken. I fixed it in release 79 opensource.

[fenomenhaa]

@Sp private 5 don,'t work on x14 algo, tried private 4 and built 78 also all not working,tprovot miner works well on this algo, pls fix private revions for x14 algo.Thx...

x14 and x17 was broken. I fixed it in release 79 opensource.

Realase 79 also broken on x14 just tried....

[sp_]

Confirmed broken. Benchmarkmode works, poolmode crash. Will look into it.

[sp_]

Submitted a fix.

Wil build release 80

[Amph]

ok, i've found how to reset to p2 state without rebooting your computer and without using inspector, just this nvidia-smi -rac

very useful, was a pain in the ass to reset each time i need to use other algo, which otherwise would result in a drivers crash with p0 state, like quark...probably due to the high overclock...