virtual memory set to 40GB. I set to 24GB first, then 40GB. Didn't improve the hashrate for the three GPUs.
I met the same problem, I made your way and it worked
Topic: Need help!h110 pro btc+ problem (Read 168 times)
What program are you using DTSM or BGMiner? I noticed I can get this exact issue with DTSM, where rather than crashing the miner, driver crashes causing some cards to go to crap solutions (using 1060 3GB's).
I also installed the three cards to other motherboard with new risers, the problem persists.
Did you try that three troubled card one by one? Or you put all together on the other motherboard? Better do not use a riser (while troubleshooting process). Plug the cards directly to PCIe 16x (one by one).
I switch the position of the three cards. It seems the problem is from the GPUs, not risers. I also installed the three cards to other motherboard with new risers, the problem persists. I guess all three cards are defective?
Afterburner shows the 3 GPUs with low hashrate only 405MHz memory clock. It seems the memory of the three GPUs are not working correctly...
That doesn't necessarily point to memory issues. It could be that the memory clock is low because it is not being used for mining as compared to your other GPUs. The fact that it is sitting idle means that there is no need to use more memory.
You can try first what the other posters have mentioned. Isolate your problematic GPUs first.
Another thing that may you can try besides all above advice you had.
You have 13 cards installed on your system, and then three cards were troubled (4,7,12). Try to unplug another three cards that worked fine, so you will running with ten cards (including cards 4,7,12), then check those three troubled cards (4,7,12). Just suspecting, this is a power issue.
You have 13 cards installed on your system, and then three cards were troubled (4,7,12). Try to unplug another three cards that worked fine, so you will running with ten cards (including cards 4,7,12), then check those three troubled cards (4,7,12). Just suspecting, this is a power issue.
Afterburner shows the 3 GPUs with low hashrate only 405MHz memory clock. It seems the memory of the three GPUs are not working correctly...
run GPU-z and see under statistic there is tab with real clock, GPU power and reason why the GPU is "capped" that might help
there is few reason why driver would drop clock, temp, power, windows utilization...
there is few reason why driver would drop clock, temp, power, windows utilization...
Try to minimize the possible problems by deduction.
These are pretty much the basics, other possibility is PSU's connectors and rails setup might affecting the problematic GPUs.
- Reshuffle the GPUs and see if GPU 4, 7 & 12 are fine on other slots.
- Look at the performance of the normal cards to the PCIe slots where those slow GPUs are previously inserted.
- Do all GPUs got the same bios versions and OC settings?
These are pretty much the basics, other possibility is PSU's connectors and rails setup might affecting the problematic GPUs.
I also tried set PCIE Link Speed in bios from AUTO to GEN1, and disabled audio not luck.
virtual memory set to 40GB. I set to 24GB first, then 40GB. Didn't improve the hashrate for the three GPUs.
I would do 8GB of memory, but what is your virtual memory set to?
I am setting up a rig with h110 pro btc+, 13 1050Ti, G3900 CPU, 4GB DDR4, Corsair 1000W PSU.
After installing windows 10 and driver, use afterburner overclock 75% TDP, +100 core, +300 mem, hashing zcash.
The problem is 3 GPUs, GPU 4,7 and 12 only do ~20 sol/s, others hashing fine 150-180sol/s.
Does anyone know what could be the problem?
After installing windows 10 and driver, use afterburner overclock 75% TDP, +100 core, +300 mem, hashing zcash.
The problem is 3 GPUs, GPU 4,7 and 12 only do ~20 sol/s, others hashing fine 150-180sol/s.
Does anyone know what could be the problem?
Jump to: