Show Posts
|
Pages: [1]
|
Holy Balls!!
Just WOW
I just want to say that whomever wrote that Radeon VII C-mode kernel should be freaking proud. Squeezing 100 MH/s out of that card at below stock, with ancient documentation, using the buggiest OpenCL compiler in existence. It's just, I'm in awe.
The text released alongside this miner is also ridiculous, I don't think I have ever seen a piece of software that is this well documented.
Gonna have some fun tonight messing with Linux and re-tuning the RVIIs.
Holy shit they also added the log rotation feature I requested.
|
|
|
Hi, just switched from PhoenixMiner. TRM seems much more performant on my Radeon VII's so kudos!
A small feature request: a way to limit the size of the logfile (and when the size limit is reached overwrite old entries like a circular buffer).
Cheers!
|
|
|
@PhoenixMiner I am recently experiencing a significant slowdown on my Radeon VII cards!
OS: Arch Linux Driver: opencl-amd 19.30.934563-1 (which is on your 'good drivers' list I think) PhoenixMiner version: 5.2e Algorithm/coin: Ethash/Ethereum
I used to use PhoenixMiner 4.7c which worked great up until a few weeks ago when It would not start anymore citing some DAG allocation error (on a card with a 16GB framebuffer LoL). I then switched to 5.2e and everything was well again (same MH/s).
Today I went to check on the rig and the hashrates have dropped dramatically! I tried auto-tune etc, same result. I read somewhere here that PhoenixMiner is switching to another method to allocate the DAG due to 4GB coming up shortly? Maybe this new method kicked in recently and it is more compute heavy? I am sure you know that the Radeon VII is (the only?) consumer GPU that is actually core-limited instead of memory-limited when mining ethash.
Radeon VII used to do ~92 MH/s out of the box, now it barely reaches 80 MH/s. The rigs could run stable for many months without monitoring, so having to monitor and make multiple changes in past week has been a pain. I would rather not go through the hassle of switching to another miner.
|
|
|
|