@Claymore
After 9.2 release everything become crazy in my rigs. As you remember I reported to you 9.2 gave gpu0 hangs in all my rigs. Yesterday When I try to find a solution, I have a new problem too. 9.0 and 9.1 began to hang my system. But when I try to mine with 5 vga, there is no problem.( -di 01234 or -di 12345 or -di 02345 any combination less then 6 work without problem)
I can mine with 6 vga ethereum Claymore 7.4. I can mine zcash with 6 vga Claymore 8.0. I checked everything but I couldn't find a solution.
here what i do when things go south.
boot to safe mode. run ddu. reinstall fav drivers 15.12 for pre rx, 16.10.1 or so (NOT the 16.12.x) for rx or rx/r9 rigs.
9 times out of 10 it worked for me.
then use reflect free to make an image backup of the working OS install. goes into the weeds again, restore the image from reflects rescue disk/usb.
I always use ddu. This is really different problem.
rig1= works with 9.0 and 9.1 without problem. 9.2= gpu0 hangs problem.
rig2= works with 9.0 and 9.1 without problem. 9.2 = gpu0 hangs problem.
rig3= system locks with 9.0 and 9.1 if I use more then 5 vga. ( This problem began yesterday. Before yesterday it has just problem with 9.2 like rig1 and rig2) 6 vga can mine zcash with 8.0 and can mine ethereum with claymore 7.4
Ever thought about the fact that perhaps a more efficient miner stresses the gpu more?
In other words, a shitty miner will only stress the gpu a little bit, the gpu will be able to retain higher clocks without any issues, simply because it's not being utilized like it "should"..
Then there's the modified straps story.. These days, with tools like polarisbioseditor it's easy for beginners to open up their rom, copy some straps, reflash.
But what most of you guys don't seem to realize is the fact that this has an enormous effect on your gpu. Duh :p
Lower straps will stress your gpu once again a lot more. This in combination with a more efficient miner and high clocks will most probably crash your GPU/System/Driver.
So.. by simply lowering your clocks, you will solve a lot of these "mysteries"..
It's nice to know exactly "what" you're doing and to understand what is exactly happening when you're "mining".. isn't it?
Hey Evilope,
Here is what I think, please give your thoughts.
When we use tighter straps , it will stress the card for certain but I dont think that stress would be able to change hardware. May be some diodes get stuck with some charge and need some more time to reset back to original discharged state. To me zcash miner seemed to keep a residual stress on the card so much that my ETH mining and XMR mining gets trashed big time. but surprisingly zec miner is the most stable for me. for 72 hrs no crashes.
What I do generally is
1) go to amd settings and clear shedder cache of the gfx cards (works sufficiently for 8/10)
2) just take my machine offline (power down for 1 min by unplugging. Keep power button pressed and possible discharge the mobo by taking out the battery).(works sufficiently for 8/10)
3) DDU and nuke the driver (works sufficiently for 10/10)
Decreasing some clocks on memory and core gives similar stability (many times more power efficiency). 5%-10% decrease in hashrate and 15-20% decrease in power consumption!
Works well