Sup, So PhoenixMiner ver's 5.6-5.8 are driving me crazy.. (Just about to test 5.9beta)
MY QESTION IS: Which is the best drivers to use for each of the Phoenix Miner Releases in order from Ver 5.5 onwards please..?
- A Follow up question would be, What is the best most reliable version of drivers to use on the 5.9beta release? Will I be safe with Adrenalin 21.10.2 OR should I go back to Adrenalin 20.8.1 ??
RIG SPECS
CPU: i7-7700K
RAM: 8GB
MOBO: ASUS B250 Mining Expert - Loaded Bios Optimised Defaults. Pcie link > Gen 1, etc..
PSU: x2 EVGA 1600w T2 Platinum/Titanium (I forget)
RAISERS: I just replaced All my GPU Raisers to the newer 009s Plus BLACK EDITION
OS: Win10 x64 Pro [Clean format]
GPUs RX580 8GB: x2 RX580 8GB Nitro+ (Hynix), x6 Nitro SE RX580 8GB (Samsung), x1 Power Devil RX580 8GB (Micron), x1 Asus ROG RX580 8GB (Hynix)
**Current** Drivers: Adrenalin 19.8.2 Beta
Tweeks: Applied Win10Tweeks.bat and done all the usual tweaks (Win update OFF etc..)
Set: Windows ENV_MEM Initial 60000mb > 88000mb
I am trying to pin down why my rig RXRig (x10 RX580's) keeps throwing errors and/or having the following issues:
I have managed to have a run of about 6hrs then another run for like 12hrs But crashing/BSOD not able to auto_restart miner from shell:startup > miner_start.bat
** Timeout Detection Recovery atikmpag.sys / atikmdag.sys VIDEO_TDR_FAILURE - TDS_BSOD errors.?
I had noticed a big jump in Temps, causing -tstop or -ttli to kick in.. Usually, my cards idle around hi 50's - to low - 60C deg C. (-tt 60)
** Higher than normal stale shares.
** -clKernel 1
** I use -altinit and -mi 12
** The issues began on v5.8c
I have two separate config's called from 2 different batch scripts depending on time of year/temperature outside
1) Standard config -mt 0
2) Winter (RXBoost config) config -mt 0 -vmr 25 -strap 1 -vmdag 1
All cards are now on -cclock 1150/850mv-900mv -mclock-2100-2150:900-925mv
I have down clocked my cards and undervolted
I was initially was running a stable -mclock 2170 on gpu's Samsung/Hynix/Micron cards (Mixed) ALL Were stable @ -mclock 2170 between 875-925mv voltages, with a range of -cclocks 1150,1170,1240,1150... etc... STABLE upto version PhoenixMiner v5.8
Now I have formatted my Rig to clean slate. I am running through all the various things I can think of to figure out why my rig keeps crashing.
Current testing on PNX Miner v5.6d - > 5.7 with Adrenalin 19.8.2 Beta drivers..
If still no luck I will bump up to last stable set of drivers 20.8.1 and repeat the testing steps. But as it stands. v5.8 really threw a wrench in the rig and cannot get it to run stable now at all..
Any suggestions would be appreciated...
Cheers Maj
From what I know from modding and overclocking Polaris and I did that a lot lot, 2150mhz is the limit of stable memory frequency and even tho only on good sample cards. Most will start to artifact at 2100 so pushing 2170 and getting stable with it was pretty sure a gamble.
I'm pretty sure samsung memory have zero effect on hashrates above 2000mhz. I still own one and there is no difference in performance from 2000 to 2150. I know hynix does well but didn't had many microns to make sure the mclock sweetspot.
-vmr increases memory refresh and will make the card run hotter, limiting temps to 60 will probably make cards throttle so try 65.