For some reason I can not get all my cards to run the same. They are all the same exact cards. I have 2 plus my mobo being powered by a 1k watt PSU and then I have the other 2 plugged into an 850w PSU I have 3 of cards plugged into one of these if it matters Amazon.com
why dont u try with absolute core clock values.
First. Are those adapters not known to cause fires as they have to much power flowing through them?
Second you need to get your thermals down. The cards are throttling due to heat. It shouldn’t take much i run my Dell 3090 at 2850 on the memory and it holds 125.9mh/s at 48 degrees. As soon as the core clock starts creeping up past 50 degrees on any of my 3090’s they start throttling. It’s not the core temperatures throttling it though just using it as a reference. The memory temperatures are causing the throttling, I bed you can’t even touch the back of those cards
I have not read anywhere about those adapters causing fires. If you have can you link me to it. You don’t even plug power into it so I’m not sure how it could have to much power. I have 2 8 pins into each card and a molex into the riser and then the riser just plug into that adapter using a usb. I am going to change the gpus around some to try and cool them down. Maybe even change the thermal pads
the problem is if its not good quality molex you might have problems and the dude is saying if you split from one cable into two and from next one into another two you end up making from one cable 3 separate ones right. well one cable can provide 190 watts alone i think but when you split it in two and then split it in two again, the 3 connections might end up pulling more that 190 watts that the initial one can hold. that will cause heat in the cable and can melt it.
So I took the cards apart and re did the thermal paste I watched a video about my exact card and they said the pads on the alienware version are good so I didnt change those. But now I am getting higher temps but the MH/s is better now. I added a screenshot of what I am getting now. Are my cards running way too hot right now?
You may want to run the overclocks in windows first for each card if possible as you can see the memory tempature in windows 3090’s throttle at 110 degrees on memory. 3090’s have huge potential, but ventilation and cooling is key.
How would I do that. Right now im using hive os on a usb with linux. Would I need to install windows on my mining rig using a different hard drive? and then once I had windows what would I use to check my temps
You could move a card to your desktop, but it kind of defeats the purpose unless you just want to double check it is indeed thermal throttling/try to find a sweet spot for each card. As for mining on windows I have no experience with it, all I know is that you are able to see Nvidia memory temperatures on windows
That being said I am fairly certain thermal throttling is your issue. That or not enough power, but the psu’s you have should be enough. In order for me to achieve my overclocks I have an 18 gpu rig, with only 8 gpus in it, and a less powerful card blowing onto the back of each of my 3090’s. They are great cards, but need tons of ventilation.
Do you know the best or easiest way to tell which card is which. I have them all plugged into one of these https://www.amazon.com/gp/aw/d/B08X4KXRB5?psc=1&ref=ppx_pop_mob_b_asin_title
This topic was automatically closed 416 days after the last reply. New replies are no longer allowed.