Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
MP 1,1-5,1 Activate AMD hardware acceleration
- Thread starter h9826790
- WikiPost WikiPost
- Start date
- Sort by reaction score
- Status
You mean rename EHC1 and EHC2 to EH01 and EH02?
I haven't done that, and recommend to connect all affected USB device via a USB 3.0 hub to work around the issue.
Last edited:
Well, that will not working out. The affected USB ports are the cMP own ports, and untill the system is not booted up, I have no keyboard+mouse.
Sorry at first ... i haven't read the 160 pages ... I use two OC instances (one with OCLP and the other with Martins Package). I think, that my graphics card (RX580) works better (in Starcraft) with Martins Package. I have Legacy WLAN on Board. Is there a possibility to activate the Card in Martins Package? In OCLP the WLAN works fine.
Hey folks. Just took the plunge and have 0.8.7 successfully running on my 5,1 w/Mojave 10.14.6, Apple-recommended RX 560, but no HEVC hardware accel. I checked in VideoProc Converter and also tried using Compressor (carefully following Martin Lo's Compressor settings, of course.) but no dice. Also in Compressor, I'm getting lots of "failed jobs" depending on the source video and settings.
Any tips for getting HEVC hardware encoding going on my set-up? (And what's up with Compressor being so flakey now?)
UPDATE: I've since updated to Monterey and now have HEVC encoding enabled. (Yay!) However, unless I'm mistaken, it seems Opencore hwaccel is limited to 8-bit HEVC encoding, no? 🧐
Any tips for getting HEVC hardware encoding going on my set-up? (And what's up with Compressor being so flakey now?)
UPDATE: I've since updated to Monterey and now have HEVC encoding enabled. (Yay!) However, unless I'm mistaken, it seems Opencore hwaccel is limited to 8-bit HEVC encoding, no? 🧐
Last edited:
Polaris GPU has no 10bit HEVC encoding. Not OC related, but hardware limit.
Ok - thanks for clarifying. 👍
So far everything else is going well with my first install, except for no wifi (2012 MP legacy card), even after following your wifi video tutorial. Legacy Bluetooth, however, is working. For wifi to work, do I need to copy over Bluetooth kexts from a Mojave installer? Or are these included in your opencore package? My *uneducated* guess is the required kexts got deleted in the Monterey install... Seems if Bluetooth works, I should be able to get wifi back, too... Any ideas?
I've got a strange issue. Suddenly my 5,1 was booting up with the Finder having no menubar. Double clicking in the Finder wasn't working either. Otherwise the mac was functional. A warm restart would always cure it. The issue would only appear about one out of 3/4 cold starts making it difficult to troubleshoot. I tried re-installing macOS, any menubar items, etc.
I finally decided to revert back to the 0.8.0 w the 12.x config list V2 (from 0.8.7) package here. Issue solved!
This is completely repeatable on my system. macOS 12.6.2. I've gone back and forth from 0.8.0V2 to 0.8.7 and the issue will not occur with 0.8.0V2. My .plist is Martin's stock with only Thunderbolt 3 enabled.
Any ideas?
I finally decided to revert back to the 0.8.0 w the 12.x config list V2 (from 0.8.7) package here. Issue solved!
This is completely repeatable on my system. macOS 12.6.2. I've gone back and forth from 0.8.0V2 to 0.8.7 and the issue will not occur with 0.8.0V2. My .plist is Martin's stock with only Thunderbolt 3 enabled.
Any ideas?
May be some updated kext causing the issue.
If you have time to run more test, when that happen, can you check Consoles to see if any possible related entry?
I'd be happy to run some tests. What should I be looking for in the Console? I checked Crash Reporter and there is nothing after a bad 0.8.7 cold boot.
Since my install (0.8.7 on Monterey 12.6.2) I've noticed the Finder now asks for admin permission before copying files, creating new folders, etc. to an external eSATA drive. This OWC external drive also has a USB-C connection, so if I connect it via USB-C instead (via Sonnettech PCIe card), then there is no issue. I've ruled out this being a drive issue by recreating the behavior with another external drive.
Is there something I need to do to restore full compatibility of permissions when using an eSATA connection?
Update: Here's a clue - I can check the "ignore ownership on this volume" in the Sharing & Permissions form for the drive with success. But again, it's not an issue with the drive and I'd need to do this for any drive I connect via eSATA. (Not particularly ideal.)
Is there something I need to do to restore full compatibility of permissions when using an eSATA connection?
Update: Here's a clue - I can check the "ignore ownership on this volume" in the Sharing & Permissions form for the drive with success. But again, it's not an issue with the drive and I'd need to do this for any drive I connect via eSATA. (Not particularly ideal.)
Thanks for the report. That's fixed few hours ago, you may re-download the latest one. Or just chage that number inside the config.
Edited
some kind of mysticism, today my safari breaks down all the time - it doesn’t print text in the browser, then it freezes-
downloaded via chrome everything is ok
Last edited:
May I know for all website, or just a specific website?
I use Brave normally, but just tested Safari, it seems functioning correctly to me.
I don’t know, but after the latest updates on iOS, my iPad is very slow in safari - constant lags, the NHL website generally works very slowly. and safari in the mail server, constantly reloads and refreshes the page - I think the latest updates do not work correctly with all sites.
As for cMP 4.1/5.1, I feel that it is living its last days and it’s time to go down in history - at least I already feel great inconvenience in working with large projects in archicad, and especially in twinmotion, UE and a little bit in Blender - the processor is already too outdated - and where you need to compile shaders - the processor does not correspond to the level of tasks - for example, yesterday I set up every scene with reflections and light with such lags and for several minutes, I got tired of suffering - saved everything for windows and rendered on windows (although also not a rocket at all)
Hi!
Did you ever solve this? I have the exact same problem and would like to know if there's a woraround...
Thanks!
I have HW GPU Decoding on my RX 6600 8GB on latest Ventura 13.1, but I have a problem. I am not able to HW Decode video higher than 4k. For example I tried 5.3K 60Fps HEVC from GoPro and it uses the CPU to decode the video. I think this is because there is a limitation imposed by the OS (Max Resolution or Max Level). The GPU is able to decode 5.3K 60, so it is just a limitation imposed. Is there a way to force HW Decode for every video?
Moreover, I am not able to encode video at 10bit HEVC using Hardware Encode, but it uses CPU Encoding, even if NAVI has 10bit HW Encoding.
Thanks!
Moreover, I am not able to encode video at 10bit HEVC using Hardware Encode, but it uses CPU Encoding, even if NAVI has 10bit HW Encoding.
Thanks!
Maybe this will be useful to Martin:
with the latest config 0.8.8, I have a kernel task process at 56% - this process uses a lot of processor resources and reduces performance. at the same time, if I use my individual config made with cdf instuctions - everything is fine.
As far as I can guess, if something does not work correctly with addresses or something is wrong in the smbios version or somewhere else there is a conflict or contradiction, a non-stop cyclic task is launched. I’m attaching a screenshot of my normal work (I didn’t take a screenshot of working with Martin’s config, but there it’s just that the kernel task is 50 or even 65%)
All previous tested versions of Martin's config worked well
I am attaching my working config for Martin’s understanding, but I don’t attach Martin’s config, since Martin has it
with the latest config 0.8.8, I have a kernel task process at 56% - this process uses a lot of processor resources and reduces performance. at the same time, if I use my individual config made with cdf instuctions - everything is fine.
As far as I can guess, if something does not work correctly with addresses or something is wrong in the smbios version or somewhere else there is a conflict or contradiction, a non-stop cyclic task is launched. I’m attaching a screenshot of my normal work (I didn’t take a screenshot of working with Martin’s config, but there it’s just that the kernel task is 50 or even 65%)
All previous tested versions of Martin's config worked well
I am attaching my working config for Martin’s understanding, but I don’t attach Martin’s config, since Martin has it
No, the problem is not Davinci Resolve. Decoding 8bit 5.3k 60 is not working in every app (such as QuickTime, IINA)
Thanks for the report. So, no matter how long you wait, Kernel task still stay at ~50%?
My package aim for providing some sort of one stop solution, may be some kext / function is not required in your setup, which cause that extra CPU usage. I only have time to test if the package can boot and function as expected. No time to monitor such issue on my own cMP. It's good to know that you can create your own config which fit your own cMP.
Anyway, may I know if the whole EFI folder is the same (especially the same kexts are used), but you just re-do the config by yourself?
Last edited:
that's why professionals every resolution (12K, 8K or 6K) always processed in a 4K timeline (if the hardware allows it), but mostly in HD, and only switched back to the desired resolution for export.
.
No, the problem is not encoding, the problem is decoding, so even using a lower res timeline doesn’t solve the problem. Is there a way to force hw amd decoding for video more than 4K?