Ryzen 7000 Testing

Including the slight discrepancy between
hw.cpufrequency: 4500000000 and
machdep.tsc.frequency: 4499995964 ?
yes it is normal also in others system without problems

In the past data hub and with a bad bootloader)was possible to interact with data hub and destroy a perfect working audio)

@CaseySJ if your system goes to sleep in a correct way
Sleep / wake I mean
could you test Realtek 2.0 audio before and after the sleep/wake process?
If it is possible without any AppleALC Alcu kext

thank you
 
SpeedKeeper makes a huge improvement!!

Details:
  • YouTube playback from Safari near perfect thru HDMI/DP
  • YouTube playback from Safari 'virtually' perfect through external USB DAC
  • YouTube playback from Firefox was previously horrible regardless of output port, but is very very good now (I think I heard a brief stutter)
  • Playback from Music App is also 'virtually' perfect through HDMI/DP and very very few stutters through external USB DAC
However, Line Out is improved, but still not listenable.

According to the author's description on GitHub:

View attachment 7801

Speedkeeper maintains an 8 to 10 percent load on CPU. Just something to be aware of.
View attachment 7802
Hey Casey,

Maybe you have this fixed already.

I'm sure you already tried but I had an audio issue on a 5950x - did you try disabling Global C-State Control and also make sure Eco mode is not enabled?
 
Code:
Last login: Tue Oct 11 17:41:47 on console
mario@marios-Mac-Pro ~ % sysctl -a | grep tsc
kern.hv.clock.tsc_base: 199249903721
kern.hv.clock.tsc_clock_last: 410232008090
net.pktsched.netem.heap_size: 2048
net.pktsched.netem.sched_output_ival_ms: 1
net.pktsched.verbose: 0
machdep.tsc.nanotime.tsc_base: 17408352592
machdep.tsc.nanotime.ns_base: 118063027738
machdep.tsc.nanotime.scale: 1160804406
machdep.tsc.nanotime.shift: 0
machdep.tsc.nanotime.generation: 3
machdep.tsc.frequency: 3699992241
machdep.tsc.deep_idle_rebase: 1
machdep.tsc.at_boot: 0
machdep.tsc.rebase_abs_time: 46662160556
machdep.cpu.tsc_ccc.numerator: 0
machdep.cpu.tsc_ccc.denominator: 0
mario@marios-Mac-Pro ~ % sysctl -a | grep freq
net.link.fake.switch_mode_frequency: 10
hw.busfrequency: 408000000
hw.busfrequency_min: 408000000
hw.busfrequency_max: 408000000
hw.cpufrequency: 3700000000
hw.cpufrequency_min: 3700000000
hw.cpufrequency_max: 3700000000
hw.tbfrequency: 1000000000
machdep.tsc.frequency: 3699992241
mario@marios-Mac-Pro ~ %

here it works fine both via HDMI than Realtek USB 2.0 audio or any USB cheaper or not audio (without any kext)
@etorix
 
  • Like
Reactions: etorix
Several good suggestions in previous posts. Will try them all when I get home.
 
How does the latest PopOS run? Anyone tried? Same audio and pciex issues?
I’ve only tried Ubuntu (latest LTS release) and it works fine. No audio issues or PCIe issues.
 
  • Like
Reactions: tomnic and etorix
  • Like
Reactions: ExtremeXT
...
@CaseySJ if your system goes to sleep in a correct way
Sleep / wake I mean
could you test Realtek 2.0 audio before and after the sleep/wake process?
If it is possible without any AppleALC Alcu kext

thank you
This system does indeed sleep and wake properly. Just tested audio as follows:
  • After waking up from sleep with AppleALCU enabled
  • After waking up from sleep with AppleALCU disabled
In both cases, audio condition is unchanged -- still a lot of static.
 
  • Like
Reactions: fabiosun
So this is a power state issue…
What happens if you disable C6 state in BIOS? (I'm suggesting this because TrueNAS Core, which is based on FreeBSD, requires disabling C6 to run properly on Ryzen. Maybe Darwin has the same underlying issue?)
On this board we have the ability to disable global C-states rather than a particular C-state. Alas, disabling global C-states does not improve audio.
 
***
If a simple little utility like SpeedKeeper can make such a significant improvement, it seems someone with kext experience would be able to develop a more technically sound solution. This seems to be a solvable problem...
***

The best part of running all these audio tests is having the pleasure of listening and re-listening and re-listening to Disturbed's cover of The Sound of Silence! :)

 
It seems that SpeedKeeper needs to take priority in the scheduler somehow to make the fix perfect. The audio was great when Mac idle (youtube on Chrome). When Cinebench runs the breaking gets worse, single thread cinebench has moderate distortion and multithread cinebench breaks audio even more.
 
Was this tool developed to help with running mac on ryzen APUs? Is this caused by iGPU? Then would non ASUS boards (gigabyte?) be better since they can actually diable the iGPU?
 
Seems like new bios today:
BIOS 0705
"Update AGESA version to ComboAM5PI 1.0.0.3 patch A
Improve system performance
Improve GPU compatibility for GeForce RTX 40 series
 
Can you try 54000000?
Just tried it, but alas no luck. AmdTscSync.kext does load and run, but also fails to yield an improvement.
 
Seems like new bios today:
BIOS 0705
"Update AGESA version to ComboAM5PI 1.0.0.3 patch A
Improve system performance
Improve GPU compatibility for GeForce RTX 40 series
This was released last week in some regions including the U.S. My board was upgraded to 0705 last week.
 
Was this tool developed to help with running mac on ryzen APUs? Is this caused by iGPU? Then would non ASUS boards (gigabyte?) be better since they can actually diable the iGPU?
Please see post below: "There's someone in the Discord server with a Gigabyte motherboard, even with disabled iGPU the internal audio is still garbage."
 
  • Like
Reactions: ExtremeXT
@PoMpIs

Did you have any trouble installing Monterey from scratch on your AM5 system (which I believe is on its way back to where it came from :))?

I’ll provide more details of my misadventures with Monterey later this evening! Will also attempt Ventura soon.
 
  AdBlock Detected
Sure, ad-blocking software does a great job at blocking ads, but it also blocks some useful and important features of our website. For the best possible site experience please take a moment to disable your AdBlocker.