Question Ryzen 2600 with AGESA 1.0.0.3 ABBA BIOS?

[VforV]

Hello.

I've just ordered an upgrade for my PC which looks like this:
CPU: Ryzen 2600 BOX
MB: MSI B450M MORTAR MAX
RAM: Crucial Ballistix Sport LT 16GB Kit (8GBx2) DDR4 3000 CL15
CPU Cooler: ARCTIC AC Freezer 34 eSports DUO White

I plan to OC as much as I can the CPU and the RAM, I've seen some tutorials for both with the MSI MB, but so far searching the net I could not find an answer to my question: should I update the MB's BIOS to the latest AGESA 1.0.0.3 ABBA version?

I know I don't need it for Ryzen 2600, but I've seen on the MSI BIOS page that it brings "Improved memory compatibility" (well, the 2 versions before the last). Link > https://www.msi.com/Motherboard/support/B450M-MORTAR-MAX#down-bios

If someone knows or has done this BIOS update to AGESA 1.0.0.3 ABBA version on their Ryzen 2000 CPU and is willing to share his knowledge I'd be grateful.

 

[VforV]

I'm using a bios that says AGESA 1.0.0.6 with my 2700x currently, the mem hits some kinda hard wall at around 3000MHz. It was released 2019/01/16
There has been 5 bios updates since this one came out on the website including ABB and ABBA.
I'd love to get 3200 speed mem running, but I have my doubts.
It seems like the IMC on the ryzen 1000/2000 is just limiting me.

I looked at your MB BIOS page and you are on an older version than me. I'm on ComboPI 1.0.0.3AB (not ABB as I said above), yours would be the equivalent F41. MSI just updated their BIOS page with the latest ComboPI1.0.0.4 Patch B (for Ryzen 3950x).

Also your MB and RAM are different than mine, so you get different results in OC.

Today I'll start OC-ing RAM with my current BIOS, I can't wait to see what I can do with that.

 

[Soulkeeper]

I just tried the latest bios. It boots 3133 and 3200 speed ram without issue, but I can't get them stable.
Constant errors in memtest instantly. Ohh well
I'll have to experiment more to be certain I can't get them stable, but it's definitely not a voltage issue.
Atleast the exact same settings used in the older bios version at 3000 mem are stable
Also something changed with their smartfan I had to add acpi_enforce_resources=lax to my kernel boot params otherwise os fan control/temps were not available.

I guess i'll have to get a 3950x and/or different mem and/or a different motherboard if I want 3200 speed mem.

 

[DrMrLordX]

Your IMC performance will do worse under any UEFI released with Matisse compatibility. Stay on PinnaclePi.

 

[Soulkeeper]

This newest bios isn't bad.
I managed to get 3133 stable with my hynix at 14-17-14-17-35 1.38v set 1.40v read, but I had to enable geardown mode with 2T. vsoc is 0.90v
3200 would make it half way thru memtest before errors/reboot.

 

[DrMrLordX]

You may lose enough performance with geardown mode enabled and 2T that the clock bump isn't worth it.

 

[Soulkeeper]

Yeah, i'll have to do some benchmarks to compare.
My main goal here is to get the IF running as fast as possible, then I can compare actual results. I still havn't given up on 3200 i'll give it another round of testing ... takes hours as you probably know.

 

[VforV]

Yeah, i'll have to do some benchmarks to compare.
My main goal here is to get the IF running as fast as possible, then I can compare actual results. I still havn't given up on 3200 i'll give it another round of testing ... takes hours as you probably know.

I don't know if this helps, but I what I found out OC'ing my 3000MHz CL15 kit is that it has a sweet spot at 3600MHz CL16 and lots of sub-timmings @ 1.4v, anything lower frequency or higher, and/or lower timings with those other frequencies either brought instability or worse performance, no matter the voltage.

I've seen my kit of RAM perform a lot better though on some Ryzen 3000 systems, but that's because they have a better memory controller, so all in all I'm in the predicted range of performance for my kit on Ryzen 2600, which is fine for me.

If you haven't read any of these, maybe these guides can help you too >
https://github.com/integralfx/MemTestHelper/blob/master/DDR4 OC Guide.md#voltage-scaling
and this one (I can't paste a simple link, only like this :/)

Spoiler

https://www.reddit.com/r/overclocking/comments/ahs5a2

 

[Soulkeeper]

I just wanted to update you guys.
I got 3200 stable !
14-18-14-18-36-54 2T GDM on @1.42v(1.44 actual)
this is 16GBx4 hynix

This voltage is high, but if It fries I guess i'll have to upgrade ...

 

[VforV]

This voltage is high, but if It fries I guess i'll have to upgrade ...

In theory it should not "fry"... all my research the past weeks and OC-ing taught me that most RAM kits can go max 1.45v and live with it, while some premium ones (mostly Samsung B-die) can go up to 1.5v max.

My MSI MB shows in BIOS the RAM voltage at 1.45v in white color and from 1.46v up in red... so it's exactly as people say from experience from OC-ing. My RAM is Micron E-die.

edit: RAM heatsinks and good case cooling should be desired when OC-ing.

 

[Soulkeeper]

Passed a single memtest run and It was "stable" doing distributed computing and using the computer for several hours then I got a reboot.
upped the voltage a bit, got another reboot. Loosened the timings same. ohh well
It's definitely stable enough to benchmark anything, no ecc errors, just a random reboot that seems to crop up after a few hours. I don't want to boost the voltage anymore, I guess i'm stuck.
Atleast this bios is far better than the older ones i've tried, I could never even get benchmark stable with them.