Help OCing ETH mining on Claymore (RX 580s)

[igottrades]

Hi everyone. Been reading the forum and finding a ton of very useful info to get my mining rig up and running. I've learned quite a bit, but still ironing out the last kinks. Would love suggestions for what I am doing wrong here. I have some OC settings on, but starting to get a random stop in mining with Claymore sometime in middle of the night. Otherwise seems to run smooth. The command line window just stops producing any new text and logs don't show any errors. The rig stays on and doesn't actually crash or blue screen, etc.

My setup:
- 6x MSI RX 580 Armor 8Gb OC (default settings are 1366 core clock speed and 2000 mem clock speed)
- 2x Rosewill Hive 1000S PSUs (1000 watts each); for now GPUs are split evenly; 3 on each PSU
- Claymore Dual Miner v11.9
- Windows 10
- Virtual memory set to 26000 - 32000
- Command console disabled quick edit mode
- Used PBE to change voltage straps for each card individually (unplugged the rest and edited each one).
- Rig plugged into its own electrical socket. Using Kill-A-Watt to monitor energy and rig is running at ~930 watts (pretty small variations) at 8.29 Amps.
- I am reaching 183.5-183.9 Mh/s total. There are 4 GPUs running at a pretty stable 31 Mh/s, but 2 lag with 30.1 Mh/s and 29.3 Mh/s. I have tried everything to get them to catch up with the others at 31 Mh/s, but I just can't without getting memory errors in HWInfo. My current setup gives me 0 errors for the 4 beast GPUs and about 10-50 errors over a long period on the lagging GPUs.
- Using "z" command in Claymore to auto-detect intensity every now and then, but these cards seem to run best at 10.
- AMD Blockchain drivers
- HWInfo for monitoring
- Afterburner to bump up fan speeds manually, since Claymore fan settings seem to not work properly (has anyone figured out how to get these params to actually work on AMD GPUs?). I run fans at about 50-60%, since that keeps the GPUs in the low 50s for temp.

Here is the text from the bat file I'm launching Claymore with (took out the pool & miner wallet):

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
EthDcrMiner64.exe
-epsw x
-esm 0
-mode 1
-asm 1
-estale 0
-r 1
-tt 77
-tstop 77
-fanmin 45
-fanmax 90
-cvddc 825,825,825,825,825,825
-cclock 1150,1150,1150,1150,1150,1150
-mclock 2145,2250,2250,2250,2250,2080

My config file also has all of the extra commands (everything below the executable).

As you can see above, GPU1 and GPU6 are the laggards.

Any comments on something I may be doing wrong? Is Claymore stopping every now and then due to the OC settings? I have all the sleep and power options turned off in Windows. Thanks for any advice ahead of time.

[NCarter84]

Try bumping your cvddc to 850 (across the board).

Also, ddu the Blockchain drivers and install newer drivers. Those drivers are over a year old now.

There is a tool on here to change all cards to compute mode, I suggest using that after installing new drivers! Don't forget to run the dag patcher.

[fgm]

-cvddc is useless under 0.950 without -mvddc, check with hwmonitor or gpuz

[evlo]

If you will be able to run 825mV@1100 stable, that means at leat 5 days without even miner crash in my book, then your cards are great. 850 is usually lowest you can get with most cards.
Maybe someone with time to really fine tune each cards to the max can get lower, but i think at most you can get 5W per card, so ~5%, very noticeable on bigger rigs, but very time consuming.

Identify which cards are crashing claymore at 825 and bump only these to 850 - that is most ideal approach.

BTW I did not know about "Command console disabled quick edit mode", thanks for the tip.