Subject: X399 Taichi Bios 2.30 / TR 1950x
Posted: 14 Sep 2018 at 11:57pm
![]() I have had such a unpleasant experience buying my first asrock motherboard. I've had issues since day one and some have been fixed with latest bios update but why do these take so long, other manufactures had the update in April and I've only just got mine :( Another example is driver updates, I have a a320m hdv board, which on the asrock site now has quite old drivers, again I go onto other manufactures websites, select a board with the same chipset and they all have upto date drivers. It's wrong that I have given my money to asrock but have to use another manufactures website to get the latest updates. And I still have unanswered questions regarding the power management of my setup as changing power plan options within Windows does not change the processor speed at all, it always still fluctuates even though I want it to run at 100%. Asrock after sales support in my opinion is zero percent. It's unexpectable |
![]() +1 i read all the time Asrock has bad bios. i was thinking its only "bad" designed grafics wise and maybe some cryptic options without explaination.. i would be fine with that, i dont care for bios looks. but it seems it is full of bugs. my am3 board from asrock was much better bios wise. on my x470 gaming k4 the fan controller is not working in default settings (cpu fan!) im not alone see PBO has only enabe and disable (no levels) disabled i get 50 mhz less so i think this is not working as intendet. also enabled is really HOT on the cpu no LLC? RGB software is a joke too (customize adressabe features? boring pre programmed patterns) when i bought the board it had the FALSE advertising still online. i really thought i get the IR VRM parts... oh they just realised their error after the video on youtube and discussion on reddit. also PWM over temp protection is faulty (better said non existend) i wish i would have got the asus x470.. ASROCK YOU LOST A LOYAL CUSTOMER! |
![]() Don't need to get Realtek audio drivers from a third party when Microsoft Update Catalog has them available and are pushed as they are released. |
![]()
So in response to all of this I can confirm I already have bios version 4.90 this is the update that came out 9 days ago with the agesa update 1.0.0.4 The last agesa update to my board was back in march and was only 1.0.0.1 which still had these gpu problems, according to amd forum they were fixed with amd agesa update 1.0.0.2 which is what I keep saying about other manufactures applying the update to there boards april/May time. So asrock released agesa update March then September that's actually 6 months I had to wait for this update to fix my problems and I can confirm that now using the latest bios version all gpu glitches have now gone but I still have this annoying power management problem. My high performance power plan is at min/max cpu state 100%, that's not the problem My gpu shared memory has always been at 1gb I don't have any c6 state options in the bios. Pspp is already on performance. And HPET is already disabled. Still Windows power plans have no effect what's so ever. But thanks for the info regarding the realtek drivers ![]() |
Configuration is Asrock taichi x399, BIOS L3.32, Ryzen Threadripper 2950X, GSKILL F4-3600C17Q-64GTZKW (4 x 16GB samsung b-die, XMP rated to 3600MHz 17-19-19-39 2T), EVGA GeForce GTX 1080 Ti FTW3, Samsung Pro 950 512GB NVME SSD, Samsung Pro 850 2TB SATA SSD, ASUS MG279Q Monitor (27" 2560x1440 144hz Freesync model, IPS panel)
The Fail Count setting has been fixed on BIOS L3.32 so that is good.
I am unable to create a video about UEFI resolution stuck at 1024 x 768. For me, it may be a hardware configuration incompatibility as loading UEFI defaults still results in 1024 x 768. The only options for the UEFI HD resolution is "Auto" and "Disabled".
Odd CAS timing with geardown mode (GDM) disabled (Command rate 1T or 2T does not matter) with RAM speeds 2933 or higher cause NO POST at POST code 15. RAM tests known good at CAS latency 16 and 18, but no POST at CAS 17. I already know GDM on with odd CAS latency automatically rounds odd CAS timing up to the next even CAS timing (17 -> 18) Here is video documenting the issue
RAM Speed 2866 GDM on, tests good CAS 16, 17, 18. GDM off, tests good CAS 16, 17, 18 @ command rate 1T and 2T
RAM Speed 2933 and 3200 GDM on, tests good CAS 16, 17, 18. GDM off, tests good CAS 16, 18 @ command rate 1T and 2T. POST FAIL at CAS 17 with either command rate 1T or 2T
Loading saved UEFI settings is still an issue on BIOS L3.32. Screen goes black and computer becomes unresponsive except caps/scroll/num lock keys and ctrl + alt + del to restart. It is an intermittent issue and hard to reproduce consistently. I have made a video with my experience. Here is video link of my experience as well as steps I did to reproduce.
Load UEFI defaults and save profile 1 as "UEFI Defaults" and restart
Go back into UEFI and change settings as follow
Boot tab > Setup Prompt timeout > 10 | boot beep > On | full screen logo > Disabled | Above 4g decoding > enabled
Security tab > secure boot > enabled
h/w monitor > cpu fan 1 setting > customize >monitor m/b > numbers from top to bottom for cpu fan 1 : 50, 55, 60, 55, 60, 55, 75, 80, 80
advanced > southbridge configuration > wan radio > disabled | bt on/off > disabled
advanced > storage configuration > sata hot plug > disabled
advanced > amd cbs > DF common options > memory interleaving > channel
advanced > active page on entry > OC tweaker
Save user default > select profile 2 > name it "test"
save settings and restart
go back into UEFI > load user default > select UEFI defaults > results in black screen. sometimes it comes back, other times it stays black screen unresponsive.
also there is another issue. setting "Above 4G decoding" causes POST hang on code B2 and beep code 1, pause, 3 beeps, 3 beeps, unless secure boot is manually set to enabled as well. requires cmos reset to restore working state. see video here.
I have not had the time to test Rtt resistance settings, and software tools used to read that setting have not yet been updated to be compatible with the 3.xx series of BIOS/UEFI.
I have not yet re-tested P-state overclocks, partly due to new second generation threadripper and precision boost overdrive replacing p-state overclock for me.
CPU cooler 65% at 50C, 80% at 57C, 90% at 63C, and 100% at 70C, but when testing with OCCT and RealBench, the fan speed doesn't bump up until the CPU temp reaches 82~83C, and it is not even 100% it starts with 80%.
Now if I put 100% at 50C, the CPU fan starts running crazy so I guess the controlling system works at least with the first phase.
I thought the fan controlling system was lagging so I put the CPU temp near 60~70C using RealBench benchmark, but the cooler just remains silent, no speed bumping whatsoever.
So I guess either the HWMonitor is reading a wrong temp or the fan control system is not working, but I think the temp is right (32~34C at idle, -100mv vcore offset). This is really annoying. I tried A-tuning but it didn't work either. I'm using the latest BIOS (ver 1.2).
Any idea or solution would be appreciated. Thanks in advance :)