Hi, hoping for some help with 1 of my 3090's I use for mining. Using v5.5c.
The problem I'm having is as soon as it hits around 60c temp, the performance drastically goes down. For some reason the card stops pulling the usual ~350 watts and only draws ~270 watts after 60c ish temperature is hit. I have a 2nd 3090 which pulls much higher, has no monitors plugged into it and draws around 400 watts. That card will happily go up to 66-67c temps and remains at the 400 watts draw. I'm really scratching my head with the first 3090. It hurts watching it's hshrate go from 120mh to 90mh just because the temp climbs from 56c to 60c lol. Meanwhile the 2nd card doesn't seem to care about temperature and will chug away doing 122mh with 400 watts.
If you need any extra information such as logs, OC settings etc then please let me know and I can provide them.
Yes, a log will be helpful, but first make sure you add the command-line option
-hstats 2 to see the actual clocks, and the reason for throttling (we assume that there is throttling involved), and because 60C is too low for temperature-related throttling, it is probably power-related but we need the log for more information.
If the card is throttling because of the power limit, you can increase it with this option
-powlim +20,0 (the 0 is to leave the power limit of the second card at default value as it doesn't have throttling issues).
Hi Phoenix, thanks for the response. I've added in the command line options you suggested and ran the miner for a couple minutes. Sadly the -powlim argument didn't really help. I wasn't sure what the most convenient way of sharing the log would be so I just went ahead and threw it in a pastebin. You can see that here:
The exact cards are EVGA 3090 FTW Ultra's and I'm using EVGA's Precision X1 overclocking software. The OC settings I'm running are the same for each card and they are as follows.
Power and Temp target slider is maxed at 119% for power and 91c for temp. The fans are also all set to 100%.