Hello, hope you guys stay healthy wherever you are…
Im new on mining and HiveOs, and I would like to ask a question.
I have 3 GPU 3080 TI, 2 Asus Tuf, and 1 MSI Gaming Trio.
GPU 0 : MSI Gaming Tro
GPU 1 : Asus Tuf
GPU 2 : Asus Tuf
Both 3 GPU driver already up to date, but getting different result hashrate.
Currently My GPU 1 had higher rate, average at 90 MH/s. I tested my GPU1 overclock setting on both GPU0 and GPU2 but the result are different, it stay around 80 MH/s on GPU 2, and below 60 MH/s on GPU0. So based on trial and error, my current overclock setting in the picture is what i found best for each individual card. So my question is:
Why GPU 1 and GPU 2 cant have same hashrate with same overclock setting? Since both of them are Asus Tuf
What is Throttle : Power Limit mean and how to solve it on my GPU 0? I browse the forum earlier and I found out that my GPU 0 (MSI Gaming Tri) got that Throttle ; Power Limit.
Below are my GPU Overclock Setting, Nvidia Card Info, and Miner Stat :
Thank you for the advice, done setting PL to 380 and no more throttle. Any advice on same Asus Tuf card but different hashrate on same overclock setting?
i always run max fan to rule out any thermal issues at first. i would try more core as well, make sure youre restarting the miner after making changes as well
This can maybe help you, I flashed it on my Asus TUF 3080 Ti (non-oc version).
Working with same clocks as Asus bios but more stable and faster a bit.
Read all the threads posted in video description and flash at your own risk!
My card also can’t go above 2400 memory without hash dropping significantly…
Red Panda did a better video based on that video. fyi, and to toot my own horn, I’m the “Hero” in Donkey’s video there (MatHelm644), lol. What really makes Red Panda’s video better is that he actually gave me (and Donkey) a shoutout. But the bios took my evga xc3 3080ti from 77 to 88+ (with lowered timings and now mirrors my evga ftw3 3080ti).
I just wanted to see how it did with the same timings as the FTW3. But in trying to get it to 77 mh, I ended up going the other way, as in down on the timings. Plus it’s gotta be easier on the gpu I think. Then I noticed that the rig I had moved it to was only at nvidia 470.86. I had used the -nvidia-driver-update command in hiveshell, but Hiveos as it turns out had dropped the 495.86 version (on which the XC3 was doing 77-79mh). So I found that I had to do the “-nvidia-driver-update 495.86” command to get it back to it’s normal low.
Point is it kind of has me approaching OC from a lowering perspective, I guess you’d call it. Because I found that with certain conditions, lower actually got a better hash number. Whereas before I jacked the numbers up as far as I could, then would back off a bit for longevity sake. So now I have to at least look/test over time to see how low I can go. But this may only apply to LHR gpus, and has a lot to do with the miner as well.
Great find!