A couple of problems with Asrock X570 Taichi |
Post Reply | Page 123 7> |
Author | |
shmerl
Groupie Joined: 23 Oct 2017 Status: Offline Points: 612 |
Post Options
Thanks(0)
Posted: 03 Sep 2019 at 2:42am |
I just finished my build (Asrock X570 Taichi / Ryzen 9 3900X), and I noticed a few issues:
1. After waking up from suspend, the motherboard debug indicator LEDs and on-board LEDs on the power/reset buttons don't turn off and continue glowing. The debug indicator shows 00 (and stays that way during normal system operation). 2. The SB chipset always runs at around 60C or so, no matter what fan setting is enabled. That's a bit strange. I'm also using an NVMe SSD in the top slot. That happens even with latest firmware/bios 2.0 |
|
Xaltar
Moderator Group Joined: 16 May 2015 Location: Europe Status: Offline Points: 25028 |
Post Options
Thanks(0)
|
That is odd, I will check it out on my X570 Taichi as soon as I have some free
time, probably Wednesday. I don't have these issues with 1.80 on my board however. When my system enters sleep it reboots to wake, not exactly normal but then again, I have not done a fresh OS install on it yet so I need to take care of that first. If you have not performed a clean install on this system (used the OS drive from a different system) then that would be a solid first step. Sleep was problematic with both first and second gen Ryzen for quite a while after they launched. I suspect the issues are 2 fold, the AGESA and the OS. I had endless issues with sleep on my X370 Taichi and 1600x, I tried countless BIOS versions and in the end, the issue just disappeared, seemingly on it's own. Checking my logs I discovered Windows had updated itself while I was out and that was what eventually fixed it. As for the chipset temps, good exhaust airflow is essential with these X570 boards, you want to pump the heat out of the case as quickly as possible. I typically run my systems with positive pressure (more intake than exhaust) and noticed my temps were pretty high. Balancing the pressure a bit better dropped my chipset temp by about 5c. If 60c is a load temp however, that is pretty decent for an X570 it seems. I have seen reports of temps up to the low 80s under load and apparently, up to 85c is within "safe" operating temperature. Still no official word on max operating temps for the chipset but I would aim for sub 80c, a spike or 2 over that in an hour long stress test is fine but the average should be under 80c, if for no other reason than to limit heat saturation of surrounding components as well as dumping hot air into your case. Edited by Xaltar - 03 Sep 2019 at 3:40am |
|
|
|
shmerl
Groupie Joined: 23 Oct 2017 Status: Offline Points: 612 |
Post Options
Thanks(0)
|
I'm using Linux (freshly installed on this system on NVMe), but it would be strange if suspend issues would be OS specific. Also, suspend works fine for me on X370 Taichi/Ryzen 2700X (also using Linux). This LED / resume is unique to X570.
Also, I noticed this in older firmware dmesg used to give this error:
Which indicated that Asrock didn't handle Linux reporting to firmware OS name for ACPI purposes. In 2.0 this error is gone, so I assume that's good! (Some details on related issues here: https://wiki.archlinux.org/index.php/DSDT ). I can report this LED annoyance to Asrock directly, unless they can see this thread on the forum. Regarding positive pressure - interesting point. I also run my case with positive pressure, to keep dust out, since I'm using good quality dust filters from Demciflex. I get, how it can affect thermals negatively. Nvme runs at around 45C on idle, while in my older case whith X370 Taichi it was around 32+C. Not sure when it starts throttling (Samsuing 970 Evo Plus). It reaches 60C with heavy I/O. If it's OK for chipset to run at 60C may be I shouldn't worry. I tried loading the NVMe and GPU, the temperature didn't really rise for the chipset. But may be lm-sensors aren't reporting it correctly? Though when inside the UEFI UI, it also shows it around 60+C, so I assume it's correct (for me the sensor is called SMBUSMASTER 1. |
|
shmerl
Groupie Joined: 23 Oct 2017 Status: Offline Points: 612 |
Post Options
Thanks(0)
|
I figured an interesting thing. In my case GPU was affecting the setup. I have Sapphire Pulse RX 5700XT, and basiclly by default, it has very passive fans policy on low load. GPU temps are quite high, and that's what's negatively affecting the heatsink and everything under it (nvme and the chipset).
I just tried radeon-profile: https://github.com/marazmista/radeon-profile and set more active fan curve. GPU temps dropped down, as as well as chipset went from 60C to 56C! Nvme went from 45C to 41C! So now I know what needs adjustment. |
|
shmerl
Groupie Joined: 23 Oct 2017 Status: Offline Points: 612 |
Post Options
Thanks(0)
|
Also, too bad Asrock made only first PCIe slot connected directly to the CPU. Putting the GPU there positions it right above NVMe and the chipset :( Not a very good thermal design. I suppose putting the GPU in the second PCIe slot would make ventilation above the chipset better.
|
|
shmerl
Groupie Joined: 23 Oct 2017 Status: Offline Points: 612 |
Post Options
Thanks(0)
|
I wonder if putting some insulating material between the long edge of the GPU and the heatsink could improve things further.
|
|
kschendel
Newbie Joined: 21 Aug 2018 Status: Offline Points: 86 |
Post Options
Thanks(0)
|
GPU's generally aren't PCIe bandwidth limited, so you probably wouldn't lose any real performance by putting it into the next slot. Might be worth experimenting with -- if you do, please report back!
|
|
shmerl
Groupie Joined: 23 Oct 2017 Status: Offline Points: 612 |
Post Options
Thanks(0)
|
Yeah, good point, I'll give it a test. How much bandwidth does a GPU like RX 5700XT saturate? I guess it depends on the screen resolution too? Given that both board and card are PICe 4, I suppose even x8 bandwidth will be like x16 of PCIe 3?
|
|
shmerl
Groupie Joined: 23 Oct 2017 Status: Offline Points: 612 |
Post Options
Thanks(0)
|
Also, are there any materials that are non conductive and are not prone for static electricity, while also having low thermal conductivity (thermal insulators)? Placing something like that between the GPU and the heatsink should help a bit I think. But styrofoam probably is a bad option, due to static electricity.
|
|
Xaltar
Moderator Group Joined: 16 May 2015 Location: Europe Status: Offline Points: 25028 |
Post Options
Thanks(0)
|
I would just drop the GPU into a lower slot. PCIe 4.0 x8 is more or less the same
as PCIe 3.0 x16, you are correct. The 2080ti doesn't saturate PCIe 3.0 x8 let alone x16 so you should be fine with the 5700xt at PCIe 4.0 x8. If you are a curious sort like myself, you can have a bit of fun benchmarking the GPU in each slot to check for differences yourself. I strongly suspect you won't see more than a margin of error difference even with the x4 slot |
|
|
|
Post Reply | Page 123 7> |
Tweet
|
Forum Jump | Forum Permissions You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You cannot vote in polls in this forum |