CCminer(SP-MOD) Modded GPU kernels.

[flipclip]

Not sure it's related, but during the last week/couple weeks I, at times, get 1/4 hashrate from every card, on ccminer restart.
Never happened before.
It could be due to the driver, the kernel or whatever; or it could be some recent modification to ccminer...

When you say "ccminer restart".  Is that using ctrl-c to close the miner down?  
With the 1/4 hashrate, if you go into nvidia-settings are the cards stuck with a 0 performance level (i.e. lowest)?
I was noticing something like the above (one of my 750Ti's {I run two, but of different manufacturers} would get stuck with a 0 performance level after a ctrl-c of ccminer), and I just reverted to using the kill command, instead of ccminer's internal stop command.  Haven't had the issue since.
(Also I sent you a message pallas.)

[pallas]

Not sure it's related, but during the last week/couple weeks I, at times, get 1/4 hashrate from every card, on ccminer restart.
Never happened before.
It could be due to the driver, the kernel or whatever; or it could be some recent modification to ccminer...

When you say "ccminer restart".  Is that using ctrl-c to close the miner down?  
With the 1/4 hashrate, if you go into nvidia-settings are the cards stuck with a 0 performance level (i.e. lowest)?
I was noticing something like the above (one of my 750Ti's {I run two, but of different manufacturers} would get stuck with a 0 performance level after a ctrl-c of ccminer), and I just reverted to using the kill command, instead of ccminer's internal stop command.  Haven't had the issue since.
(Also I sent you a message pallas.)

Yes by clicking ctrl-c.
I will look at performance levels when that happens again, didn't think of it, thanks.
I've seen the PM just didn't have time to try it ;-)

[Cryptozillah]

I have some stability problems with two of my "leftovers" nvidia rigs.
One of them are powering down the whole system after mining for a couple of hours and the other one is ccminer just crashing on randomly.

I have identical hardware in all my the rigs (NVIDIA/AMD), mobos, cpu, ram, risers and psus.

In one rig i have one EVGA FTW 750Ti and the other rig i have one EVGA SC 750Ti + Gigabyte Windforce 750Ti.
The one powering down is the rig with two cards.

At the same time i have two rigs running five 750Ti:s each at 36Mhs Quark totally stable for weeks.

This is really annoying because they are also crashing with stock settings on the gpu:s. Temps are really low also so i dont get this.

All systems have fresh Win 8.1 installations.
I am running the 355.60 Driver on all rigs and i am using the 61 build due to best performance for me with that one.

Anyone have a clue what could be the problem ?
Can i find logs from ccminer somewhere ?

Start swapping hardware one component at a time between a good rig and a bad rig until the good rig goes bad
and the bad rig goes good. Then you've identified the faulty component.
Some more info would be helpfull. Is there any error message when ccminer crashes? Can mining be resumed or
does the system need a reboot?

Here's something that can be added to a bat file to check for a crash.

tasklist /FI "IMAGENAME eq ccminer.exe" 2>NUL | find /I /N "%1">NUL
if "%ERRORLEVEL%" NEQ "0" goto restart_ccminer_becauuse_it_died

I will take a screenshot next time ccminer stop mining, cant remember what it say when it happends.
If i just close the miner and start it up again i will get really bad hashrate. So a reboot is needed after every "crash"

I have been testing replacing some hardware like psu and ram but still the same problems with these two rigs. I am starting to think that these gpu:s are someway faulty. All these three cards are used cards that i have been buying lately.

I would suggest swapping the graphics card to check if it makes any different.
And I have been mining with release 66, so far is stable for me.

Yeah i guess that is what i need to do.
I am using r61 on all rigs so that build should not be the problem.

I am watching some of my 750Ti cards pushing above 7.3Mhs as we speak. I see no increase in speed with the other releases for my 750Ti cards.

[Cryptozillah]

I forgot to say that its like 50/50 that ccminer stop mining and just stay there. Or it will just crash and close itself.
I need to reboot after either of that happends to get full hashrate again.

[djm34]

Not sure it's related, but during the last week/couple weeks I, at times, get 1/4 hashrate from every card, on ccminer restart.
Never happened before.
It could be due to the driver, the kernel or whatever; or it could be some recent modification to ccminer...

it could be due to tasks requiring "some cpu power" are running in the background (anti-virus, botnet (lol), stuff like that...).
I usually use an elevated priority (cpu-priority 2 or something) to make sure it doesn't happen)

[pallas]

Not sure it's related, but during the last week/couple weeks I, at times, get 1/4 hashrate from every card, on ccminer restart.
Never happened before.
It could be due to the driver, the kernel or whatever; or it could be some recent modification to ccminer...

it could be due to tasks requiring "some cpu power" are running in the background (anti-virus, botnet (lol), stuff like that...).
I usually use an elevated priority (cpu-priority 2 or something) to make sure it doesn't happen)

Not likely in my case: it's a vanilla Ubuntu machine used only for mining, well updated and cared for ;-)

[sp_]

I submitted a bether launchconfig for blake-512. Another quark speedup on the 750ti

building release 67

[joblo]

Not sure it's related, but during the last week/couple weeks I, at times, get 1/4 hashrate from every card, on ccminer restart.
Never happened before.
It could be due to the driver, the kernel or whatever; or it could be some recent modification to ccminer...

When you say "ccminer restart".  Is that using ctrl-c to close the miner down?  
With the 1/4 hashrate, if you go into nvidia-settings are the cards stuck with a 0 performance level (i.e. lowest)?
I was noticing something like the above (one of my 750Ti's {I run two, but of different manufacturers} would get stuck with a 0 performance level after a ctrl-c of ccminer), and I just reverted to using the kill command, instead of ccminer's internal stop command.  Haven't had the issue since.
(Also I sent you a message pallas.)

Yes by clicking ctrl-c.
I will look at performance levels when that happens again, didn't think of it, thanks.
I've seen the PM just didn't have time to try it ;-)

I have had cards do that, usually following a crash, both Windows and Linux.
I have also noticed that Ctrl-C can also cause it. Killing the task from a seperate shell seems to have
fewer problems.

[sp_]

release 66 had a bug that could cause the miner to hang on stratum failure. (disconnects)

I believe T Nelson fixed it with this commit:

https://github.com/sp-hash/ccminer/commit/331a437d6a0093050eeacd25d1707c1a1b4cd386

[sp_]

-NVLM working in linux (merged by flipclip) (tpruvot's implementation)
-New time limit option (merged by flipclip) (tpruvot's implementation) (mine for a given number of seconds and then exit)
-Corrected hang on stratum failures (T Nelson)
-Fixed multiple -i and -X values. (merged by flipclip)   (You can now set individual intensites for each of the gpu's)
-Faster lyra2v2 on compute 5.2 devices.
-Faster quark on 750ti.

1.5.67(sp-MOD) is available here: (14-sep-2015)

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

The sourcecode is available here:

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

[pallas]

Not sure it's related, but during the last week/couple weeks I, at times, get 1/4 hashrate from every card, on ccminer restart.
Never happened before.
It could be due to the driver, the kernel or whatever; or it could be some recent modification to ccminer...

When you say "ccminer restart".  Is that using ctrl-c to close the miner down?  
With the 1/4 hashrate, if you go into nvidia-settings are the cards stuck with a 0 performance level (i.e. lowest)?
I was noticing something like the above (one of my 750Ti's {I run two, but of different manufacturers} would get stuck with a 0 performance level after a ctrl-c of ccminer), and I just reverted to using the kill command, instead of ccminer's internal stop command.  Haven't had the issue since.
(Also I sent you a message pallas.)

Yes by clicking ctrl-c.
I will look at performance levels when that happens again, didn't think of it, thanks.
I've seen the PM just didn't have time to try it ;-)

I have had cards do that, usually following a crash, both Windows and Linux.
I have also noticed that Ctrl-C can also cause it. Killing the task from a seperate shell seems to have
fewer problems.

I experienced it right now, after ctrl-c.
Cards stay in P8 (power saving, idle state)

[Slava_K]

Quark +50kH on 750Ti
-160kH on 980GTX.

[sp_]

Quark +50kH on 750Ti
-160kH on 980GTX.

Hmm. Looks like I need to revert the last change on the compute 5.2 devices.

[flipclip]

Not sure it's related, but during the last week/couple weeks I, at times, get 1/4 hashrate from every card, on ccminer restart.
Never happened before.
It could be due to the driver, the kernel or whatever; or it could be some recent modification to ccminer...

When you say "ccminer restart".  Is that using ctrl-c to close the miner down?  
With the 1/4 hashrate, if you go into nvidia-settings are the cards stuck with a 0 performance level (i.e. lowest)?
I was noticing something like the above (one of my 750Ti's {I run two, but of different manufacturers} would get stuck with a 0 performance level after a ctrl-c of ccminer), and I just reverted to using the kill command, instead of ccminer's internal stop command.  Haven't had the issue since.
(Also I sent you a message pallas.)

Yes by clicking ctrl-c.
I will look at performance levels when that happens again, didn't think of it, thanks.
I've seen the PM just didn't have time to try it ;-)

I have had cards do that, usually following a crash, both Windows and Linux.
I have also noticed that Ctrl-C can also cause it. Killing the task from a seperate shell seems to have
fewer problems.

I experienced it right now, after ctrl-c.
Cards stay in P8 (power saving, idle state)

Make and model of the card?  Maybe it is specific to a card and brand.  I'm almost positive it only affected one of my cards and not the other.

[sp_]

Quark +50kH on 750Ti
-160kH on 980GTX.

Hmm. Looks like I need to revert the last change on the compute 5.2 devices.

I have reverted it now for compute 5.2. Looks bether.

[joblo]

Not sure it's related, but during the last week/couple weeks I, at times, get 1/4 hashrate from every card, on ccminer restart.
Never happened before.
It could be due to the driver, the kernel or whatever; or it could be some recent modification to ccminer...

When you say "ccminer restart".  Is that using ctrl-c to close the miner down?  
With the 1/4 hashrate, if you go into nvidia-settings are the cards stuck with a 0 performance level (i.e. lowest)?
I was noticing something like the above (one of my 750Ti's {I run two, but of different manufacturers} would get stuck with a 0 performance level after a ctrl-c of ccminer), and I just reverted to using the kill command, instead of ccminer's internal stop command.  Haven't had the issue since.
(Also I sent you a message pallas.)

Yes by clicking ctrl-c.
I will look at performance levels when that happens again, didn't think of it, thanks.
I've seen the PM just didn't have time to try it ;-)

I have had cards do that, usually following a crash, both Windows and Linux.
I have also noticed that Ctrl-C can also cause it. Killing the task from a seperate shell seems to have
fewer problems.

I experienced it right now, after ctrl-c.
Cards stay in P8 (power saving, idle state)

Make and model of the card?  Maybe it is specific to a card and brand.  I'm almost positive it only affected one of my cards and not the other.

Don't think so. It's happened to me on 780ti, 750ti & 980, all EVGA.

[sp_]

All systems have fresh Win 8.1 installations.
I am running the 355.60 Driver on all rigs and i am using the 61 build due to best performance for me with that one.

Can you try build 67 please.

Do you have a cudart32_65.dll in the ccminer folder?

Downgrading the driver could increase by 200-300KHASH per 750ti. (Hot tip )

Don't forget to donate guys.

[pallas]

Make and model of the card?  Maybe it is specific to a card and brand.  I'm almost positive it only affected one of my cards and not the other.

Two different brands, same behaviour.

[sp_]

Why do you need to break with ctrl-c ? Seems to work fine in windows.

[t-nelson]

Are there more details about this alleged ctrl+c issue?  It's not clear to me what you're doing, let alone what you expect.

Ctrl+C doesn't do anything but signal some loops to break in the control logic.  Which loops depends on if it's the first or second time you've pressed it.  It doesn't actually touch the GPUs at all.

I think you're either observing the intended behavior or a driver bug.

If you press ctrl+c once, then see GPUs drop to power-saving, that's expected behavior.  You've told the GPU worker threads to stop after they're done with current work.  So as the threads complete, you will see GPUs drop to low power modes, since they have nothing to do.  App is just waiting for the rest of the threads to finish before exiting.

If you're running ccminer, killing it with ctrl+c, then starting it again and seeing low performance until a reboot.  That sounds like a driver bug to me.