Closed goodbest closed 3 years ago
I wonder the VRAM usage is about the same ? I didn't run any 3D APP, just chrome with youtube. that's it
This may be a system issue, not related to the patch.
For the record, I'm using the kext all the time. Before 20:19, the VRAM is 99% but I did nothing other than Chrome.
After 20:19, I reboot the device, and start playing StarCraft2 (with metal engine). The VRAM remains ~30%, but the GPU processor reaches 99%(which means the game uses a lot calculations)
VRAM:
GPU Processors
Cool, can you confirm that Metal isn’t disabled? By running a Geekbench Metal compute?
EDIT: ah didn’t see the comment in the other thread. Seems like you’ve already tried it.
@osy86 , @goodbest
I have run the Geekbench 5 test, looks good
Thanks, I’ll include this in the next release.
My VRAM usage is also topping up when I use Chrome. Might be a bug, but haven't noticed any side effects yet.
https://github.com/osy86/HaC-Mini/issues/316#issuecomment-731743555
The true explanation why the patch is working remains unknown yet, since
21
is a magical value in the dyld. The following is the logic why I try the number 0.The original Big Sur use an statement:
if somevalue <21, then do sth.
Glitches are everywhere.the current kext is patched to do the statement:
if somevalue >=21, then do sth.
The logic is illustrated here: https://osy.gitbook.io/hac-mini-guide/details/metal-driver-fix However it's working partially in big sur.So I replace the value 21 to 0:
if somevalue >=0, then do sth.
And it's working now, with no side effect found yet.BTW, since I suppose a value between 0 and 21 would be better, I tried with different numbers for HVK. The patch is working with the patch
if somevalue >=19
, but no withsomevalue >=20
. However, I think it's still better to useif somevalue >=0
if there's no obvious side effect.This PR will close https://github.com/osy86/Polaris22Fixup/issues/10 https://github.com/osy86/HaC-Mini/issues/316
Potentially close: https://github.com/osy86/HaC-Mini/issues/392