Quote from: PhoenixMiner on Today at 07:48:59 AM

Quote from: micdee on Today at 12:18:58 AM

Quote from: PhoenixMiner on March 15, 2021, 05:27:32 AM

Quote from: micdee on March 15, 2021, 05:11:12 AM

Tnx. I will check the data and the logfiles.
When I run the miner in the console as far as I know it always starts with the card names.. (which GPU are found for mining) In this list it shows 6 times the same (type) cards RX 580

I even had fitted the RX 5700 as a single available GPU, but then I get the miner error "NO GPU for mining"

Is there any known hardware limit for getting the RX 5700 XT's to work in Phoenixminer, e.g. minimum of RAM used, or anything else, which we might have overlooked?
The rig is fitted with only 4 Gb RAM. (We use the Asrock H110 BTC Pro+ mobo)

Perhaps the driver version may be the issue. Do not use the latest AMD Linux drivers for anything other than RX6000 AMD cards.

Well we tried a different (lower version) driver but then we had to change the kernel and still it didnt want to see the 5700 and the additions in the setting didnot seem to work, so this morning I changed to Win10 and we did see all the different cards.. and even the card settings in the confiig file (the RX*580:1160) did work. But we had the 21.1.1 drivers installed and it was not stable. So I uninstalled these and installed the 20.11.
and forced the compute mode via the AMD driver software, but I do not get it stable after a few minutes I get black screens and AMD report of a driver issue and then it all stops???

Probably one of the cards is unable to run at the clocks and voltages that are set. Check the log to see if the proper clocks and voltages are applied to each card - if there is a mix-up and the setting for 5700XT is applied to one of the Polaris cards, or vice versa, it may lead to almost instant crash. The last resort is to check the cards one by one using the -gpus command-line option. If all are running fine by themselves, you may have issues with your power supply, or the cables to the extenders.

Let me start with thanking you for your patience with me (and my friend)
After installing all available drivers of AMD (even the 21.3 which is not yet supported I found out ;-) ) it still keeps crashing

There are a few things I noticed...
With a few drivers the RX5700 card kept (re)setting its memclock to the number I set in the config file.
With last install of 20.11 drivers. I did not get a timeout on the driver (thread stuck....) but I saw an gigantic hashrate every one would like to get (from the logfile: 3: 2947100.674 MH/s) after a few of these line the log states: "
"2021.03.17:11:59:05.099: GPU3 GPU3: Starting up... (0)
2021.03.17:11:59:05.621: hwmc GPU3: set GPU clocks to 1300 MHz (Vddc 850 mV)
2021.03.17:11:59:05.624: hwmc GPU3: set VMEM clocks to 1800 MHz" <-- no vmddc mentioned

I no run the 5700 alone with following config:
" .........
-pool ssl://eu1.ethermine.org:5555
-pool2 ssl://us1.ethermine.org:5555
-wal --------------------------------------------
-prate 0.22
-cdm 2
#-cdmpass Bulldozer

-gpus 3
-tt 68
-fanmin 20
-tstop 77
-tstart 60
-amd
-acm
-cclock *:1150,rx*5700:1250
-cvddc 850
-mclock rx*580:2175,rx*5700:1800
#-mclock 2175,2175,1800,2175,2175,2175,2175
-mvddc rx*580:855,rx*5700:950
#-mvddc 860,850,900,850,850,860,850
-gt 0
-vmr rx*580:29
-logdir --------------------------------
#-minrigspeed 200
-wdog 1
-rmode 2
......"
When starting Phoenixminer this is what the log recorded:
".......
2021.03.17:13:18:28.445: main Starting GPU mining
2021.03.17:13:18:28.445: main Matched GPU1 to ADL adapter index 48 (method 1)
2021.03.17:13:18:28.514: main GPU1: AMD driver 20.11.2
2021.03.17:13:18:28.514: main GPU1: Created ADL monitor for adapter 48; overdrive version: 8 ( Cool

2021.03.17:13:18:28.514: main GPU1: using AMD driver ver 20.11.2
2021.03.17:13:18:28.576: hwmc GPU1: set auto fan: 68C target temp (min fan 20, max fan 100)
2021.03.17:13:18:28.583: hwmc GPU1: set GPU clocks to 1250 MHz (Vddc 850 mV) (NO MEN clock or voltage)??
2021.03.17:13:18:28.646: wdog Starting watchdog thread
......"
In "stand alone" mode it is already running longer without a problem than before.
We have 2x750W of power.. 4 rx580 cards on PSU1 and 2x580, 1 5700 on PSU2. I think that should be sufficient.. we do about 1030W on the wallsocket.
What might cause the incredible super fantastic hashrates?

micdee

Posts

Merits