Ryzen - Aorus AX370 Gaming 5 Review + Linux Test | Level One Techs

We get our hands dirty with the Gaming 5!
Confirmed working ECC with Dual Rank Kingston ECC kit!

This is a companion discussion topic for the original entry at https://level1techs.com/video/ryzen-aorus-ax370-gaming-5-review-linux-test

Those puns tho :P

Nice review!


That intro music is wondeful. I like it so much better..

Hey Wendell, I could be wrong but my understanding was that the Aorus Gaming 5 didn't just support a CPU with integrated graphics, I thought it actually had its own onboard video processor.

Per Newegg's spec page for it: http://i.imgur.com/OKS3aEp.png

Also, PC Part Picker lists it as having its own video onboard, not "Depends on CPU", although I know PC Part Picker is sometimes wrong.

I'm looking strongly at this board for a new GPU Passthrough build, and I specifically picked it because of the onboard video of the motherboard thinking it would possibly alleviate some issues due to the bad IOMMU groupings since onboard video can't be passed through anyway.

Just thought I'd check and see if my understanding was wrong or what.

Great video though, I've been eagerly awaiting it ever since you said you had the Aorus 5 and the TaiChi for review!

To make it short: It doesn't.

Gigabyte is just weird on wording that on their website.

Integrated Graphics Processor:
1 x HDMI port, supporting a maximum resolution of [email protected] Hz
* Support for HDMI 1.4 version. Maximum shared memory of 2 GB
* Actual support may vary by CPU.

so yeah. I guess PC Part Picker just takes whatever is on the website, and the one on Newegg is literally copy&paste.


PS: Why do these videos always come online while I'm at work :/ Youtube and Webproxies blocked :(

Damn! I was afraid of exactly that. Well, thanks for clearing that up for me.

Once again , I'm glad that I didn't get in a hurry and order before I had all the information. Fortunately, I'm not in a huge hurry on this build, I don't plan to buy parts till after I graduate college in May, and if possible, I'd like to wait for Vega GPUs too.

The good part of that is it gives Ryzen some time time to mature, more mobos to come out with hopefully better IOMMU groupings, because I was pretty set on doing another PCIe Passthrough on this rig.

So, I think this has been asked a couple times before, but I haven't seen a definitive answer. Can bad IOMMU groupings be fixed down the line by the board manufacturer, or would that be the responsibility of AMD to do something about it? That's getting beyond my knowledge level a bit.

I guess what I'm curious about is whether or not the two X370 boards reviewed by L1Techs so far are recoverable for PCI Passthrough purposes, or if we'll just have to wait and hope that more boards come later with better groupings.

99% sure it's a software thing because it makes little sense m.2 is in group 0 with promontory unless that's really a unified x8 PCh interface.

Plus also those apparently unused bridges which there are a lot.

Suggests we may see an x8 promontory with pcie3.0 down the line that uses both sets of lanes though. How cool would that be.


I followed you up to here

but then you lost me. So, the current thought is that it should be fixable, just might take a little time? That'd be great because I still really would like to use the Aorus Gaming 5 in my new passthrough build if they can get that sorted out.
Like I said above, I'm not building for several months anyway, so that gives the hardware some time to mature and such.

I'll be sure to keep a close eye on the L1Tech channel for more news on this stuff as it progresses, as I'm sure you'll be reporting on it.Thanks for clearing that up for me!

Maybe they could fix iommu groupings with future firmware updates?
But i think that the biggest question here is, how much of an attention this item would get?

I think some people on the "inside" are blown away by the numbers on these videos about this subject.. 165k views on passthrough, and crazy number of views on the "state of iommu" video in a short amount of time.

at least thats my impression from mobo vendors. and their engineers are working really hard poking at it, sending me test bioses, etc.


Thats exally great news.
I was kinda expecting at the market niché of these particular boards,
that the vendors wouldnt really botter too much with fixing it.
But seeying that they taking these things very serious is great news.
So they are definitelly watching the video´s, thats awesome.
Of course we probablly gonne see more X370 boards in the future aswell.
Its all in a early stage atm.

I am unfamiliar with the quality of level1techs review. "Everything works" includes sensors ?

PS for those who are wondering E2500 "Killer nic" a chip from Atheros and is supported by the "alx" driver.

I'm more wondering why or which sensors wouldn't work :P

It is supported as far as I read yes, but the whole network prioritization stuff is Windows-only software as far as I know. Though in my opinion that should be done on a network level anyway, but w/e :D

I'm more wondering why or which sensors wouldn't work :P

Temperature, voltage, RPM and PWM control. Some ITE chips have no Linux support for example it8655e
https://github.com/groeck/it87/issues on ASUS Prime X370-pro

The mobos that have advanced uefi options I tend to not test the sensors as heavily in linux. The UEFI control of fans based on sensors, ramp up/down time is pretty good on later gen boards. I'll try to do better testing of sensors in future like I used to do. I do like the UEFI control of fans/temp setpoints tho.

1 Like

Thanks for the review! I really enjoy the style of Level1Techs, in depth info, no bs and calm and relaxing to watch!

You mentioned in the review that you got 2933MHz to work, I've got this board as well and even with the latest UEFI version I can't get it stable above 2133MHz or even boot above 2400MHz. Could you perhaps post what timings/voltages you used to get 2933MHz to work? Or what the profile settings are that worked?

Well I wrote my own code to control the fans, didn't get around publishing it yet. (Has some feature I like ;) )
Running sensors && find /sys/ -iname "*pwm*" on the board would answer all questions btw.
Unless it's horribly broken, always an option, the only way to know that is to actually test it.

As soon as this board is in stock, I should be ordering it. But if it's horrible broken I might as well order the ASUS board, since that is in stock and is cheaper.

EDIT: another bloody nice feature: ability to change the settings without rebooting ;)

Any socket issues with this sample?

One of our friendly forum members has/had one with a physically loose socket from what I understand.

So... Just got the chance to watch the video. I didn't even know there were temp sensors included in the package :)

About the RGBW header... It's not that white isn't white enough, but the white from regular RGB strips tends to have a slight blue-ish tint. The white on the RGBW strips tends to have a warmer colour.