AB350 Fatal1ty k4 AGESA 1.0.0.6a CPU Temp?
Printed From: ASRock.com
Category: Technical Support
Forum Name: AMD Motherboards
Forum Description: Question about ASRock AMD motherboards
URL: https://forum.asrock.com/forum_posts.asp?TID=5641
Printed Date: 26 Jun 2024 at 11:57pm
Software Version: Web Wiz Forums 12.04 - http://www.webwizforums.com
Topic: AB350 Fatal1ty k4 AGESA 1.0.0.6a CPU Temp?
Posted By: teelily88
Subject: AB350 Fatal1ty k4 AGESA 1.0.0.6a CPU Temp?
Date Posted: 20 Jul 2017 at 2:35am
|
" rel="nofollow - I updated BIOS to 3.0 for the AGESA 1.0.0.6a. All is well (well, cant reach 3200 still but im just 1 notch under so thats fine by me) except for the CPU temp in bios. Its reading +20c. HWinfo shows tdie properly for cpu (averages 30c with normal use) AND under the MOBO the CPU temp shows similar average to tdie around 30c. However, inside BIOS, its reporting +20c. I use a custom curve for the fans, and even pushing them up +20c, they are jumping all over the place. Its like the mobo is reading directly from the tctl in hwinfo now, which is reporting spikes and thus spikes in the fan settings. Anyone else experiencing this or have suggestions? Fans and CPU reporting in BIOS ran just fine on bios 2.6. CPU: Ryzen 1600x MOBO: AB350 Fatal1ty Gaming K4
|
Replies:
Posted By: danielcepa
Date Posted: 20 Jul 2017 at 3:47am
|
Having the same fan issue on the AB350 Pro 4 with a R7 1700 In 2.60 the custom fan curve followed the MB CPU temp sensor resulting in a nice ramp up curve without sudden spin-ups. In 3.00 the fan curve follows the Tctl/Tdie (cpu) temp wich is notoriously jumpy with ryzen. This results in odd fan noises even when doing trivial tasks. Fan speed is all over the place with instant jumps between rpms (900-1400 in my case) This is a regression, imho. |
Posted By: teelily88
Date Posted: 21 Jul 2017 at 1:03am
|
" rel="nofollow - I agree with regression. The hard part for me is that in using an "x" cpu (1600x), its reporting a +20c offset, so at a tdie 60c, its reading at 80c and throws the fans up to 100%. with 3x 140mm noctua 2000rpm industrials for intake, that is screaming. The old fan curve was PERFECT! At least give us an option. I'm losing a lot of faith with asrock right now with this ryzen board. CPU voltage issues (MASSIVE vdroop and under powering at idle), no LLC, RAM is all over the place, now this fan issue and over reporting the temp by 20c? We knew at release that the "x" cpus over reported on tctl... its a little ridiculous.
|
Posted By: Xaltar
Date Posted: 21 Jul 2017 at 1:33am
|
This change seems to be across all boards using AGESA 1.0.0.6a. This is obviously a first implementation and there are still kinks. I got around it by setting a custom fan curve with much higher tolerances. The temps are being taken from a different sensor now (as some of you already know). I am sure ASRock is working on it for the next update. The problem is trying to keep up with bug fixes on top of AGESA version implementation and the fact that AGESA versions make sweeping changes to how things work. ASRock is not alone in this, every manufacturer is struggling to keep up and manage bug fixes. We just have to hang in there. It will get resolved, it will just take time  ------------- 
|
Posted By: rpmrg
Date Posted: 21 Jul 2017 at 8:45pm
| 3.0 is a bad release and a buggy BIOS, i reverted back to 2.60. |
Posted By: parsec
Date Posted: 21 Jul 2017 at 11:06pm
|
We have had a significant amount of complaints in the forum about the original, pre-AGESA 1.0.0.6a UEFI versions CPU temperature and fan speed control relationship on the ASRock Ryzen boards. Those users felt the CPU temperature reported in the UEFI/BIOS, and used for the fan speed control, was the incorrect one to use, and did not cause the fan speed control to increase the CPU cooler's fan speed when monitoring the CPU temperature as it should. The change we see in the Ryzen 3.00 UEFI versions (I have it with my X370 Killer SLI/ac board) seems to be a response to those complaints. I had a feeling that if this was changed, it would be met with mixed reviews. IMO, this is related to AMD's use of the Tdie and Tctl CPU temperature readings. My understanding of the two readings is they are related to the fan speed control needed for the supplied Ryzen CPU cooler vs standard PC fans used with aftermarket CPU coolers. The purpose and use of those two CPU temperature readings is still unclear IMO, and needs to be sorted out. ------------- http://valid.x86.fr/48rujh" rel="nofollow"> 
|
Posted By: AlbinoRhino
Date Posted: 22 Jul 2017 at 12:59am
|
" rel="nofollow - I'm not sure why the +20C still exists anymore. It isn't needed. Maybe it's due to all RyZen chips being unlocked which would lead to more OC'ing which then could represent a problem of a higher demographic burning up their chips. Maybe AMD is safeguarding themselves. In my Bios I have given up on setting custom curves for all of the fan headers listed there. I've set them all to "Silent Mode" which has cured the fans from ramping up and down. I would like to see the fan curves for each of the preset modes inside of the Bios. Will Asrock provide these to me for the Killer SLI A/C? |