Overclock.net banner
3341 - 3360 of 3365 Posts

·
Registered
Joined
·
105 Posts
Unfortunately the loss in performance when pushing past 1900 FCLK still exists and the "tweaks" that worked with the A.A4 BIOS to get back the performance no longer work with the A.A0 BIOS.

Also no change in WHEA 19s at 4133/2067 and no change with regards to audio anomalies.

Going to play a little more then back to my 24/7 settings .....

** EDIT **
Preliminary testing indicates lowering CCD from previous voltages is assisting sound anomalies, i.e. before I was using 1.050/1.060 (CCD/IOD) now dropped CCD to 1.020 and sound is now behaving even after several reboots, but LatencyMon still cries.

Though dropping CCD may result in some performance loss.

All in all, does not look that this agesa is going to get me running 4133/2067 as 24/7 settings ...

** EDIT 2**
Tweaking PBO, rather than leaving it on "Motherboard" gains back some of the performance when compared to what was occurring with A.A4

Setting "CPU VDDP" to 0.880v now allows 4133/2067 to post without "07" error every time ive rebooted (done around 30 reboots), this is an improvement from when we did not have access to "CPU VDDP", of course needs more thorough testing but looks like a positive result.

Latency seems to be a little worse, I get 51.6 - 51.8 ns with the settings below

View attachment 2517378
How do you get your cldo vddp so low? Lowest I can set in bios is 1V since a couple of versions ago, or is it somewhere there under AMD oc / ryzen master?
 

·
Registered
Joined
·
105 Posts
Oh, and would you mind a short summary of anything special you do with UBU? For some reason the bioses I update doesn't work after flashing now (won't boot).. at the end, you use the save as mod option? And then just rename it to original name? Only thing I haven't tried for a while is using original name I think, feels weird even in bios flashback it doesn't work though
 

·
Registered
AMD 5900X, MSI 570X Unify, GTX1080, F4-3600C16D-32GTZN
Joined
·
108 Posts
Oh, and would you mind a short summary of anything special you do with UBU? For some reason the bioses I update doesn't work after flashing now (won't boot).. at the end, you use the save as mod option? And then just rename it to original name? Only thing I haven't tried for a while is using original name I think, feels weird even in bios flashback it doesn't work though
Save as mod option then delete the mod_ stuff so the file's name is equal to the original. MSI accepts the modded bios when using the original filename.

I used UBU to update drivers and microcode after my modifications.
 

·
Registered
Joined
·
1,779 Posts
Sounds like something funky is going on with your BIOS.

Ive always had such voltage options for CLDO VDDP and I change them in the "regular" OC section where the other voltages are.

I didnt understand what is the "UBU".

The last few flashes ive done is by using "AfuWin64"

Here - BIOS ⁄ UEFI Utilities » AMI
"Aptio_V_AMI_Firmware_Update_Utility"

Tick "Program All Blocks"

Back to stress test results ....

Couple of hours of Y-Cruncher also passed

ZenTimings_Screenshot_uga2.png
 

·
Registered
AMD 5900X, MSI 570X Unify, GTX1080, F4-3600C16D-32GTZN
Joined
·
108 Posts
Sounds like something funky is going on with your BIOS.

Ive always had such voltage options for CLDO VDDP and I change them in the "regular" OC section where the other voltages are.

I didnt understand what is the "UBU".

The last few flashes ive done is by using "AfuWin64"

Here - BIOS ⁄ UEFI Utilities » AMI
"Aptio_V_AMI_Firmware_Update_Utility"

Tick "Program All Blocks"

Back to stress test results ....

Couple of hours of Y-Cruncher also passed

View attachment 2517393
There is no CLDO_VDDP in the Vermeer side of the bios. It's still there in the bios for older generations.
 

·
Registered
Joined
·
1,779 Posts
There is no CLDO_VDDP in the Vermeer side of the bios. It's still there in the bios for older generations.
Huh ???

You got something mixed up ??

In our BIOS "VDDP Voltage" is translated in ZenTimings as "CLDO VDDP"

And I am guessing that this is what @kmellz is referring to ?.?
 

·
Registered
Joined
·
105 Posts
There is no CLDO_VDDP in the Vermeer side of the bios. It's still there in the bios for older generations.
Huh ???

You got something mixed up ??

In our BIOS "VDDP Voltage" is translated in ZenTimings as "CLDO VDDP"

And I am guessing that this is what @kmellz is referring to ?.?
Ah, was mixing things up, there's VDDP at the top, then at the bottom a similarly named chipset soc voltage cldo, which at first glance I though was listed there in zentimings
 

·
Registered
Joined
·
1,779 Posts
Ah, was mixing things up, there's VDDP at the top, then at the bottom a similarly named chipset soc voltage cldo, which at first glance I though was listed there in zentimings
(y)(y)
 

·
Registered
Joined
·
112 Posts
Now on Bios A91 and been shifting films from one usb drive to another nearly all day and no probs at all.
Now gotta see if mem does the same as on A4.2
Smiffy
 

·
Registered
Joined
·
1,779 Posts

·
Registered
Joined
·
137 Posts
Tunned PBO+CO on the new BIOS and boost clocks are lower again! Now getting less than 680 points single thread on CPU-Z benchmark... Going back to AA4 or AA1.

I also did some game benchmark tests comparing the optimal CO between A82 (AGESA 1.1.0.0) and AA4 (1.2.0.3A) and it seems game performance actually increase few % (minimum frame times). So I think I will keep the newer AGESA for now even though single core boost is lower. 1.1.0.0 gets consistently CPU-Z 1T scores of 700 while 1.2.0.3 gets around 690 points.
I can get back single core boost setting higher negative CO for my best cores but then it's not stable (need to go from -7 to -12 at least to get same boost and closer to 700 CPU-Z points)...

Anybody seeing boost clocks regression with this newer AA0?

p.s. all BIOSes where flashed used AFUWIN to make sure the whole ROM is programmed
 
  • Rep+
Reactions: Cidious

·
Registered
Joined
·
150 Posts
After update to AA0 from A9 and redoing my settings, I am noticing higher cpu temperatures in gaming.
Have not been running any benches to compare performance. But my temps are definitely at least 5C higher.
 

·
Registered
Joined
·
558 Posts
Tunned PBO+CO on the new BIOS and boost clocks are lower again! Now getting less than 680 points single thread on CPU-Z benchmark... Going back to AA4 or AA1.

I also did some game benchmark tests comparing the optimal CO between A82 (AGESA 1.1.0.0) and AA4 (1.2.0.3A) and it seems game performance actually increase few % (minimum frame times). So I think I will keep the newer AGESA for now even though single core boost is lower. 1.1.0.0 gets consistently CPU-Z 1T scores of 700 while 1.2.0.3 gets around 690 points.
I can get back single core boost setting higher negative CO for my best cores but then it's not stable (need to go from -7 to -12 at least to get same boost and closer to 700 CPU-Z points)...

Anybody seeing boost clocks regression with this newer AA0?

p.s. all BIOSes where flashed used AFUWIN to make sure the whole ROM is programmed
You summed it up. They went for stability. And generally better performance. But less peak. I think this was all in the light of killing the USB bug.
 

·
Registered
Joined
·
1,779 Posts
Tunned PBO+CO on the new BIOS and boost clocks are lower again! Now getting less than 680 points single thread on CPU-Z benchmark... Going back to AA4 or AA1.

I also did some game benchmark tests comparing the optimal CO between A82 (AGESA 1.1.0.0) and AA4 (1.2.0.3A) and it seems game performance actually increase few % (minimum frame times). So I think I will keep the newer AGESA for now even though single core boost is lower. 1.1.0.0 gets consistently CPU-Z 1T scores of 700 while 1.2.0.3 gets around 690 points.
I can get back single core boost setting higher negative CO for my best cores but then it's not stable (need to go from -7 to -12 at least to get same boost and closer to 700 CPU-Z points)...

Anybody seeing boost clocks regression with this newer AA0?

p.s. all BIOSes where flashed used AFUWIN to make sure the whole ROM is programmed
I noted something also, I got back most of lost boost by tweaking PBO and tweaking telemetry.

On my 5600x I am using [email protected], [email protected], [email protected], then in CPU VDD telemetry I have set 105A for the first value and 14A as the offset
 

·
Registered
Joined
·
22 Posts
CPU VDD telemetry I have set 105A for the first value and 14A as the offset

Is that 14A as in 1400mA ? Or did you mean 14mA in the CPU VDD Telemetry Offset?

Using EDC 210A is interesting. What benefits have you found using 210A as opposed to say 140A (which I am using)?
 

·
Registered
Joined
·
1,779 Posts
Is that 14A as in 1400mA ? Or did you mean 14mA in the CPU VDD Telemetry Offset?

Using EDC 210A is interesting. What benefits have you found using 210A as opposed to say 140A (which I am using)?
14mA

:)

Look here
 

·
Registered
Joined
·
24 Posts
Tunned PBO+CO on the new BIOS and boost clocks are lower again! Now getting less than 680 points single thread on CPU-Z benchmark... Going back to AA4 or AA1.

I also did some game benchmark tests comparing the optimal CO between A82 (AGESA 1.1.0.0) and AA4 (1.2.0.3A) and it seems game performance actually increase few % (minimum frame times). So I think I will keep the newer AGESA for now even though single core boost is lower. 1.1.0.0 gets consistently CPU-Z 1T scores of 700 while 1.2.0.3 gets around 690 points.
I can get back single core boost setting higher negative CO for my best cores but then it's not stable (need to go from -7 to -12 at least to get same boost and closer to 700 CPU-Z points)...

Anybody seeing boost clocks regression with this newer AA0?

p.s. all BIOSes where flashed used AFUWIN to make sure the whole ROM is programmed
yes my 5800x used to boost 4850mhz on default but now it does at 4800 i think they are trying to make more stable cpus across the board since mine is really bad silicon i have now more stability on stock freq. of course with memory oc
 

·
Registered
Joined
·
22 Posts

·
Registered
Joined
·
1,779 Posts
I thought so.
However, your post said you used 14A for CPU VDD Telemetry offset. That's very different to 14mA.
That is correct, to the magnitude of 1000

😂😂

however I did direct you to a thread where a discussion takes place with regards to the "mA" and "A" here (at the bottom of that particular post) that discusses the usage of mA as the input value.

Please spend some time and test it yourself, you may come to the similar conclusion that its looking like the "mA" value is actually amps,

so sorry for the misinformation
 
3341 - 3360 of 3365 Posts
Top