AGESA 1.0.0.6

wardog wrote: Remember FIRST to Reset to UEFI Deaults, Save and Exit, then flash as this is major BIOS rewrite/update.

You mean that when the Bios Agesa 1.006. a comes, i should do that what you recommeded to me?

like reset uefi defaults, save and exit, then flash to the agesa 1.006 A when avalaible, then just set the advance memory off, save and exit, and try xmp profile 2.1 with the new bios update?


by the way, if i try that what i describe above in this text, but with this 1.006, the profiles doesn't deleted, so maybe i will be waiting till the new agesa 1.006 A.

Remember FIRST to Reset to UEFI Deaults, Save and Exit, then flash as this is major BIOS rewrite/update.

wardog wrote: x58haze wrote: Maybe the culprit of my ram, is that, i notice in Taiphon, at 1600mhz it should be running on TFAW 36, and the rc close to 64 Therefore i'm running 1533x2= 3066 mhz dual kit, 16-17-17-35, with tfaw 36, rc 40 and maybe thats the problem right? it should have a better tfwaf, rc, thats why people say that (XMP profile 2.1) are not compatible at all with Ryzen, which is better to manually overclocking the ryzen memory? That's why previously I suggested: wardog to x54haze wrote: DO go into mem config and make sure those XMP sub-timings are being properly detected and set. I no 'toopid Why this happens? Altogether now, lets sing "AGESA to the ........ Rescue?" Reset your BIOS, Save and Reboot, turn off Memory Training, THEN set XMP, reboot and your proper sub-timings should appear. Correctly, as per XMP. See what I did there? With Memory Training enabled, regardless of XMP being set, Training allows Ryzen to match an in this case "close but not exact' table entry from within the AGESA section code. I've seen this here in hous of testing, and why, x58haze, I previously ask you to ckeck the sub-timings.? As far as I'm concerned this is a Bug, and has been reported. AFAIC when XMP is manually chosen and set/selected by the User, Memory Training should be automagically Disabled so as not to override the XMP settings derived off the memory stick that was programmed there by the memory manufacturer. But is it a true "bug'? Let's contemplate AMD might purposely be allowing it, Memory Training, to occur behind our backs in this manner due to Ryzens finicky on again off again affairs as they relate to a previous BIOS clocks the same mem higher for some whereas the current BIOS clocks them lower for others. If so, AMD has some way yet to go with future AGESAs and compatibility. Which is what may be allowing this, compatibility.

Well in my case i followed your Advise, about setting all to default, by pressing f9, then saving, then putting off the (advance memory training) then saving, and boot to bios again to active XMP 2.1 to 3200

and the thing is, that is not working it still putting the TFAW to 24, and the RC to 40, while Thaipon runner shows that these memories at 3200 should be running with sub-timmings of RC 64 and TFAW 36

Therefore, if i put xmp profile 2.1 then set to 3066 16-17-17-35 as you recommend before, and only changes the sub-timming TFAW to 36 and rc to 64, instead of using 3200, well pc boot as usual, but problem of input lag persist
check the timmings:


Edited by x58haze - 14 Jul 2017 at 10:31am

Jakob wrote: Hey there, wardog! Thank you for the explanation. Would it be possible to use the free version of taiphoon to alter the SPD-settings (XMP won't help me, I'm afraid) of my not-booting-by-itself RAM so it would maybe boot. Then, I could work the BIOS, see what happens.. Thanks again!

Well, that's soley my assumption. Yet, and as to the reasons why, I may have let this happen too. What with Ryzens inherent finickiness towards memory and ALL the past DDR4 that Users had/have in the hands, compatibility would be my priority also. AMD _and_ the board manufacturers also!

The Freeware version of Thaiphoon Burner doesn't.

But PM me and I'll hook you up with a discounted price to purchase.

x58haze wrote: Maybe the culprit of my ram, is that, i notice in Taiphon, at 1600mhz it should be running on TFAW 36, and the rc close to 64 Therefore i'm running 1533x2= 3066 mhz dual kit, 16-17-17-35, with tfaw 24, rc 40 and maybe thats the problem right? it should have a better tfwaf, rc, thats why people say that (XMP profile 2.1) are not compatible at all with Ryzen, which is better to manually overclocking the ryzen memory?

That's why previously I suggested:

wardog to x54haze wrote: DO go into mem config and make sure those XMP sub-timings are being properly detected and set.

I no 'toopid


Why this happens? Altogether now, lets sing "AGESA to the ........ Rescue?"



Reset your BIOS, Save and Reboot, turn off Memory Training, THEN set XMP, reboot and your proper sub-timings should appear. Correctly, as per XMP.

See what I did there?




With Memory Training enabled, regardless of XMP being set, Training allows Ryzen to match an in this case "close but not exact' table entry from within the AGESA section code.

I've seen this here in hous of testing, and why, x58haze, I previously ask you to ckeck the sub-timings. 


As far as I'm concerned this is a Bug, and has been reported. AFAIC when XMP is manually chosen and set/selected by the User, Memory Training should be automagically Disabled so as not to override the XMP settings derived off the memory stick that was programmed there by the memory manufacturer.

But is it a true "bug'? Let's contemplate AMD might purposely be allowing it, Memory Training, to occur behind our backs in this manner due to Ryzens finicky on again off again affairs as they relate to a previous BIOS clocks the same mem higher for some whereas the current BIOS clocks them lower for others. If so, AMD has some way yet to go with future AGESAs and compatibility. Which is what may be allowing this, compatibility.