CH7 Bios 2703 Issues

[RSFoxx]

I apologize in advanced for being very vague here but i'm seeing more and more performance degradation going from bios 2501 to 2606 to 2703, with my 3700x. I don't believe that was the intention but my cinebench scores and my 3Dmark scores are getting lower and lower with every update. Also, ram timing support seems to be getting worse as well. What has me triggered enough to make this thread though, is the fact that PBO enabled/disabled/auto doesn't seem to make any change with custom ram timings including FCLK at 1600 or 1800mhz. If I run DOCP (3200mhz CL14) my CPU speeds boosts to a decent 4,350mhz. However, if I run custom timings, my CPU sits at an all core of 4,275mhz without change. Making adjustsments to PBO has no effect. Any help would be much appreciated.

So far, bios 2501 has been the best performer for me with CB R15 score of 2151. Bios 2703 gives me a CB R15 of 2043.

[RSFoxx]

Ok, so I took some screenshots to better show what is going on. I'm focusing on ram and ram timings only because of this issue i'm seeing. I'm using the G.Skill TridentZ Neo 3200mhz CL14 Samsung B-die.



When I run Default bios settings with DOCP @ 3200mhz PBO on Auto I see normal operation with 1 or 2 cores hitting 4,350+ frequency.



If I go back into the bios and adjust ram frequency and timings to 3600mhz CL14 in 1T with default settings, my 3700x runs all core frequency of 4.2 GHZ. Also, adjusting PBO to disable, auto, or enabled doesn't change the frequency.



I also included this screenshot of ryzen master for any additional timings or settings i'm missing.



Here are my CB R15 scores for reference.

[adr82]

IIRC, 2501 is the latest BIOS which still had AGESA 1.0.0.2. That's the version that seems to generally boost higher than any of the 1.0.0.3 variants in later BIOS releases, so it'd make sense that it would give you the best benchmark results.

The rest of what you're seeing I would be very tempted to chalk up to general Asus incompetence. I've also seen major changes in the apparent level of DRAM overclockability between BIOS versions, although with my C7H 2406/2501 seem to be worse than 2606/2701/2703.

I think it's often very difficult or at least time consuming to debug anything like this, because what you have is essentially one black box (AGESA) inside another black box (the BIOS), with the added complication that the BIOS is also quite buggy because Asus either can't or don't want to devote any significant resources to it. In other words, some of your benchmark losses could be down to changes in boosting caused by AGESA, some of it could be down to intentional BIOS changes, and some of it could be down to unintentional BIOS changes (bugs).

[JkFoobar]

It's working as intended. AMD dropped the power limits, which means clocks.

https://www.overclock.net/forum/28085580-post549.html

https://www.overclock.net/forum/11-a...read-4275.html

[RSFoxx]

It's working as intended. AMD dropped the power limits, which means clocks.

https://www.overclock.net/forum/28085580-post549.html

https://www.overclock.net/forum/11-a...read-4275.html

Thank you for not reading my posts and posting someone else's 2 cents that doesnt pertain to my issue.

[adr82]

You might want to try out the modded BIOS versions from this post, as they contain AGESA 1.0.0.2. This would let you compare standard 2703 vs modded 2703, and if you saw any improvements in the modded version that would then point to AGESA 1.0.0.3 as the cause. Similarly if you didn't see any change between the two BIOSs with different AGESAs, that would indicate that the BIOS itself is to blame. Of course it's entirely possible both are playing into the behaviour you're seeing, but it still might be interesting to do the comparison!

[RSFoxx]

You might want to try out the modded BIOS versions from this post, as they contain AGESA 1.0.0.2. This would let you compare standard 2703 vs modded 2703, and if you saw any improvements in the modded version that would then point to AGESA 1.0.0.3 as the cause. Similarly if you didn't see any change between the two BIOSs with different AGESAs, that would indicate that the BIOS itself is to blame. Of course it's entirely possible both are playing into the behaviour you're seeing, but it still might be interesting to do the comparison!

Thank you for this and your last post. I was going to revert back to 2501 but I might actually test this first just to see. It's certainly not operating as intended.

[Nick Name]

Does it happen if you don't use DOCP and manually set all RAM settings?

[RSFoxx]

Does it happen if you don't use DOCP and manually set all RAM settings?

I mean, that is why I created this thread. When I run DOCP I have no issues. If I manual set ram frequency and timings, the cpu is stuck at an all core frequency of 4.2ghz.

[Nick Name]

I mean, that is why I created this thread. When I run DOCP I have no issues. If I manual set ram frequency and timings, the cpu is stuck at an all core frequency of 4.2ghz.

Sorry, I meant when you change the frequency to 3600MHz is it after setting DOCP or are you setting that value to Manual and then setting it to 3600MHz?