Topic: Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux) - page 305. (Read 6590718 times)

Hello,

for whose with 1060 3G that doesn't work, what drivers do you use? Are you dual mining?

most of my rigs are in 382.53 and one is on 388.00 I think, all on cuda 8.0,  and I have no problem in ETH only.

the new issue with claymore.

As soon as I start claymore, it started crashing.

All my AMD rigs started crashing on "C:\WINDOWS\System32\DriverStore\FileRepository\igdlh64.inf_amd64_250db833a1cd577e\igdrclneo64.dll" (in event logs)

reinstalling AMD drivers via safe mode didn't help
Restoring windows to previous good known version didn't help

Finally, just deleted igdrclneo64.dll which fixed the issue for me on my all rigs - do this at your own risk, highly recommend creating a system restore point before trying this.

this trick worked for me/

-1. It is good as it is. If you want subfolder or unique filenames, use a script. The man has put enough options to control the log file already.

It is a fact that when you do OC settings you are adding power and stress to the card, and most likely your hardware cound't handle it. with the steps I posted you could probably get an idea on why a single Card (assuming it runs successfully) with OC settings or vice versa (Assuming it failed)

I just upgraded to 11.5 my few of my rigs GTX 1070 and with both increases hashrate, both are sometimes unstable under linux

What you mean by "Disabled all power states in the video card BIOS except the top"? Can you share some picture?

Little update on testing my BIOS modded rx580 mini rig.  This is dual mining keccac and always the same bat file

v11.0 - 61.3 mhs with heavy fluctuation into the mid 50s
v11.5 - 61.9 mhs with heavy fluctuation as above -oldkernels 1 option used
v11.5 - 62.2 mhs fluctuation of a tenth at most(62.15 to 62.25), seems stable on the new kernels for some reason?

Heavy fluctuation would be down into the mid 50s.  Sometimes for a few minutes.  Each card would randomly drop 3-10%.

Here's the difference between when 11.5 was unstable (4%+ rejected no matter how much I slowed, plus freezing) with the new kernels and now:

- Not running HWinfo
- I didn't run GPUZ before, but it's important NOT to run it while mining
- updated to the new video drivers that came out march 19, minimal install which may be very important.  I had the bloat installed before.
- disabled QuickEdit mode in console
- Disabled all power states in the video card BIOS except the top

The card speeds and voltages are the same as before when I had instability.  I can't think of anything else that has changed.

Be careful updating to the new AMD drivers.  They will block all overdriving software until you run pixel clock patcher.  The patcher wasn't required before for all cards, now it will be.  I hope some here find this useful and Claymore as well.  Happy mining.  

PS will report back after running 24 hours on the new kernels.  I've only been running an hour but I can tell something has changed.

I was posting a while back that I've got stability issues on mixed polaris 570&580 rigs.
Somehow similarly I have one miner with 8*570*4GB Asus (boxy looking and not the Strix one) and 4*580*8GB Nitros
It was just giving me pain - miner was crashing probably every few hours (the only one out of  10+ of the similar build, except the mixed 570&580)
In some logs I could see that GPU0 (my display card) would fail so...I had a display dummy also mem (I think Elpida) set at 1920
So I rolled back to 11.2 removed the display dummy and increased the mem clock to 1950 and now it's fine.

https://forum.ethereum.org/discussion/14274/gpuminer-cu-k1-failed-4-unspecified-launch-failure

help pls .s

mining.help also has a pretty good description of how to control individual card settings and claymores control in his opinion isn't stable.

If you haven't already goto mining.help has a complete description and software tools that help you set up claymores for max performance including a 1 click bios upgrade which worked brilliantly for me.

Keccak is not Keccakc

I can try that, but what will I learn?  The system runs stable for 24+ hours if I don't modify the GPU settings or if I modify the settings using Afterburner.  But putting those same settings in the Claymore config causes a crash on startup. 

Doesn't the fact that I can run stable with no GPU modifiers (or modifying GPU with Afterburner) indicate that hardware and drivers are working correctly?
 - Dave

Start at mining only one GPU installed in PCIe x16 slot, and set aside or exclude all other GPU's for the meantime.

If this runs successfully,  you may now install other GPU's one by one, from there you can actually troubleshoot which might be the culprits.

It maybe a Riser, PSU, Drivers, etc.

I'm having a problem with a crash or lockup when starting.  I've tried 11.3, 11.4 and 11.5.  All exhibit the same issue.

Claymore runs great when I don't set any GPU parameters in the config file.  However, as soon as I add any clock or voltage adjustments to the GPUs, one of 3 things happens when I start the software:
1. BSOD with Thread Stuck in Device Driver error
2. System just locks up (no mouse movement or num lock), requiring a hard power cycle
3. It starts properly but eventually exhibits failure 1 or 2 after an hour or two.

When I say the problem happens with any GPU parameters, I've tested this by using all stock settings (the same that the GPUs default to when I don't set anything in the config), but I still have the problem.  So, the problem is not settings outside of a stable operating range.

The crashes usually occur after "POOL/SOLO Version", but not all the time.

I checked the logs but nothing stands out.  The logs look normal then just stop when the crash happens. 

I'm running a rig with 3xRX570 and 3xRX580 using the blockchain driver on Win10.

I've tried running -oldkernels 1, but that actually made it worse.  I couldn't start mining at all.  I always get OpenCL errors, even with no clock or voltage adjustments.

I tried -gser1 and -gser 2, but neither of those helped.

Is there something else I should be looking at?
 - Dave

see https://ethermine.org/miners/0xB529d2FaC7dA5902Cba304dA4901315A82668332
I can see you have a balance of 0.00052 ETH

oho my Rutherford.... what the hell i was doing. OK even though i was mining on etherminer should my wallet be visible there ?

Ok i'll bite you're not mining at nanopool you're mining at ethermine.   If you really want to mine at nanopool change your bat file
to where i have it boldfaced.
start C:\GPU_Miner_v10.5\EthDcrMiner64.exe -epool eth-eu1.nanopool.org:9999 -ewal 0xB529d2FaC7dA5902Cba304dA4901315A82668332.Vismai -epsw x
pause"

Your wallet will only be in the log a couple of times at the top.  It should literally take minutes for them to make a page.

This is my bat file

"

setx GPU_FORCE_64BIT_PTR 0
setx GPU_MAX_HEAP_SIZE 100
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_SINGLE_ALLOC_PERCENT 100
start C:\GPU_Miner_v10.5\EthDcrMiner64.exe -epool eu1.ethermine.org:4444 -ewal 0xB529d2FaC7dA5902Cba304dA4901315A82668332.Vismai -epsw x
pause"

and I found this in the logs while was trying to gain some insight on what's happening

"23:07:02:504   1e0   watchdog - thread 0 (gpu0), hb time 2734
23:07:02:508   1e0   watchdog - thread 1 (gpu0), hb time 1016
23:07:03:497   3e38   ETH: put share nonce aeb50a38014a8cf8
23:07:03:501   3e38   ETH round found 1 shares
23:07:03:505   4344   ETH: 03/16/18-23:07:03 - SHARE FOUND - (GPU 0)
23:07:03:510   4344   send: {"id":10,"method":"eth_submitWork","params":["0xaeb50a38014a8cf8","0x98bd38500076ce467e3f8792bb54df7452fe408c7a7f947b46b6ce8a99753788","0x4e0abf8513087c241fad4701cafc18baa839e21c8e44917cf5e75d9ad634003f"]}
"

now my bigger question why does it has some other wallet addresses in the last line of the log.
I am using Claymore's dual miner version 10.5

please help, this is frustrating. I am watching from 3 days for my address to appear on nanopool.