As a reference point, in Overview I have Radeon Software Version: 17.7, Radeon Software Edition: Crimson ReLive.

In Windows Device Manager\Display Adapters I see one entry, namely "AMD Radeon(TM) RX Vega 11 Graphics" and its driver properties are:

  Driver Provider: AMD

  Driver Date: 06/02/2018

  Driver Version: 23.20.827.0

Under Events I also see the "requires further installation" entry though, so I suspect it is of no consequence. Have you tried re-booting and reinstalling the driver.

I'm seeing a different build of Windows. I've got 1709 build 16299.248. I downloaded the latest ISO on 13th Feb while waiting for my APU to arrive and made an installation USB stick. Since installing there have been a number of updates, so it might be worth forcing a check.

Can you confirm the version, please? I'm seeing 17.7 and suspect that's what you meant. The other version is perhaps for some other AMD GPU - maybe as a result of misidentification.

Anyway, glad it's all good now.

Everything under the Software tab matches mine exactly. We just differ on the Radeon software version on the Overview tab. I'm seeing 17.7 while you report 7.7 - so is that a discrepancy, a typo or do you need to visit Specsavers  ?

Now you're reporting 17.7, which is the same as mine. Last time you said 7.7. Go back and check 

That's right. Don't get angry. Look back a few posts. You said there you had version 7.7 when, it turns out, you meant 17.7. It's no big deal but you confused us both.

I wouldn't have thought it was my ram, worked fine in with my 1700 in Asrock Z370 Killer and Taichi. But I guess I could test it, happens a lot during any downloading, games drivers, the WIFI fluctuates and then it just locks up. Same thing happened when I tested the 2200G on my X370 Killer.

JinxB1rd wrote: Same. I didnt thought that my ram could be bad, cos i used it since august without any noticable trouble. With Ryzen 1500x. But as extended vga memory it has to be perfect, or u got freezing screens. Every vga card cause freezing, if the ram is bad. Better, if you test it. It can be explain everything. /sry for my english/

JohnM wrote: +1 on the advice about testing the RAM. If the RAM proves to be good, are you using WiFi to download when the problem happens. Try using the cabled network instead. There are other threads here about WiFi problems. Personally, I always used cabled Ethernet whenever I can as it's hugely more reliable.

Just started memtest86 at stock clocks... I'll check its progress a little bit later. Wifi was never an issue with my 1700 with my Killer or Taichi, Wifi has only been an issues with the 2200G. Does the same thing on both boards. If the wifi doesn't work the boards are kinda useless to me. I disabled the wifi on the Killer and it stopped BSOD'ing while using a wired connection. Didn't bother to test it on the B350 Itx yet.  Will update later with the ram situation.

My 4th Asrock board with wifi.. so its a little frustrating the same issues are happening only with the 2200G.

You'll eventually want to run MemTest86 with the same clocks as you plan to use Windows, but starting with stock is a good place to begin your troubleshooting.

Like I said, I avoid WiFi whenever possible. It's a convenience but no substitute for cabled Ethernet. I haven't even plugged my board's aerials in. You could try re-seating the WiFi card in its socket. It's just a PCIe 2.0 device and not even connected to the AM4 socket. It hangs off the Promontory chip so I don't see how changing from a Summit Ridge to a Raven Ridge processor could affect it.

Yeah I get what you are saying... But this particular machine is going to need wifi. If it was my primary, I would be wired.

JinxB1rd wrote: Dont forget the AIDA64 stability test.

I'll mess with that a little bit later. Just finished an hour on memtest no issues.

Also, just a suspicion that there aren't really "chipset" drivers for the 2400G.  The package we're downloading appears to be for the GPU portion.  The SMBus and GPIO are much older drivers in the same package (from last summer it seems) so I don't thing any of that is fundamentally changed.  Only the "Display Adapter" driver is new, at least on my installation.

The only other oddball behavior I noticed is an ugly memory leak on the Windows Audio Isolation Graph, which never happened before I installed the 2400G.   Not sure if it's related or just an weird coincidence.  If it happens again, it might be an error/bug with the HDMI Audio subsystem in the new GPU driver.  I remember there used to be an ugly collision between early AMD audio drivers and Realtek HDMI...

Isn't that what you'd expect when Cool 'n' Quiet is enabled?

but i cant find a option to set dram size for gpu...

Can anyone help?

+1.

But if you still want to do it: http://forum.asrock.com/forum_posts.asp?TID=7770&title=2400g-where-can-i-set-igpu-vram-size-in-bios

(Sorry, making it into a real hyperlink doesn't work as the "?" gets translated.)

Also, on my other machines, setting power to High Performance sets CPU to full and disables all other forms of throttling.

This is exactly what's happening to me on the AB350 Gaming-ITX/ac with a 2400G, BIOS 4.43. It's as you describe while downloading using the wireless will freeze and / or throw the DPC_Watchdog_Violation errorr Curious to know if you figured anything out yet.

I was having the same issue, So far it seems fixed after downloading the newest driver for the wireless from intel.

That was the very first thing I did was updated wireless driver. Currently running 19.51.11.2.

I had to temp the fates, After moving the computer to its home location away from the router I just had another BCD_Watchdog BSOD from another small download. It almost looks like the driver is crashing if it drops enough packets, and takes everything with it.

Only you can answer that one but given that you "love" it and the fact that there will obviously be both a new BIOS and new Radeon driver coming along soon I'd hang in there.

I have an issue where my screen will turn a solid color with any sound continuing to play, and the computer will be able to respond to keystrokes like mute audio or something. It is easy to reproduce the solid color when playing FurMark, but it happens elsewhere. Hitting my physical power button on the PC will shut down the computer, not before the solid color randomly changes to another solid color however.

I have also have an issue where a game will crash and say something like "unable to access 0000 0108 or FFFF FFFF" of a local addressable space, not sure if that is related to the cpu/iGPU managing its memory share or something that causes conflicts, but I never had that problem on my other computers playing the same game for hours.

These are all in addition to:

- various blue screen errors... commonly being the VIDEO TDR issue

- Instant crash when trying to overclock the iGPU even the slightest (+50MHz) with AMD Master software

- frequent frame drops and stutters in games

My best solutions (based on my results) to reducing the frequency of occurrences, beyond the common recommended setup instructions, is to down clock RAM, turn of SMT in bios (good for certain games), and slightly down clock iGPU.

2400g

ASRock mini ITX fatality Mobo

16gb Corsair Vengeance DDR4 3200 (stock)

Yes, the only way I can seem to get past 2133 is with XMP enabled, I turn off XMP and my ram goes down to 2133.

My exact ram Module is shown on my mobo Memory QVL (Raven Ridge):

http://www.asrock.com/mb/AMD/Fatal1ty%20AB350%20Gaming-ITXac/index.asp#MemoryRR" rel="nofollow - http://www.asrock.com/mb/AMD/Fatal1ty%20AB350%20Gaming-ITXac/index.asp#MemoryRR

as:

So I figured it would be all fine and dandy. I have noticed that it does say that anything over 2667 with XMP could be unstable depending on the CPU compatibilities.

-------

I just tried running FUR Mark with XMP off, 2133, and I got error code with blue screen, I got code Thread_Stuck_In_Device_Driver and Video_TDR separate occasions immediately after the slightest iGPU overclock. Also I see no off for XMP so I figured "Auto" is off.

I think it is also worth mentioning that at one point I could not get my PC to even post, It would be stuck on an endless boot loop. I finally solved it buy unplugging and plugging the ram back in. at the time my ram was at 3200 XMP.

I have also ran Memtest86+ with XMP 3200 for about 10 Horus and nothing showed up. Additionally Prime 95 doesn't seem to cause the issue when I use the option most taxing on the ram.

-------

 I have the b350 mini itx fatality board.

I have been out of the custom PC game since around 2012, So I am not familiar with XMP.

Is there a way to even get 3200 without XMP? I will check my bios rn

Let me tell you a little about my situation. I'm not interested in overclocking my 2400G. I'm happy with the CPU and GPU at stock. The only thing I have done is slightly overclock my Corsair Vengeance RAM from its rated 2666 MHz to 2933 MHz. That enables the rest of the processor (the "uncore") to run at its optimum speed - anything more would be an overclock of the processor.  Note that the APUs have a higher optimum uncore speed than 1000-series Ryzens (2666 MHz). I ought to have bought faster RAM but I originally planned to get an R5 1500X and changed my mind and got the 2400G instead. But I was lucky and was able to run my RAM at the slightly higher speed simply by selecting XMP and then 2933 from the drop down menu. I didn't have to change anything else to achieve this.

Right now it's very stable, otherwise I would have reverted to 2666. The only BSODs I've had in the last few days were due to a bug in GPU-Z (fixed in the latest version) and from testing the WiFi card - I don't normally use this because cabled Ethernet is better in every respect but I tried it in order to help someone else.

In case anyone is interested here is the link for the sale till Thursday:

https://www.newegg.com/Product/Product.aspx?Item=N82E16820232530

Apologies if this has already been answered! I've looked high and low and haven't been able to find it.

Is this able to be set in the BIOS? Where is that setting?

Different board but the BIOS menu structure should be similar - does this help: http://forum.asrock.com/forum_posts.asp?TID=7805&title=how-can-i-oc-my-2400g-vga-from-ab350mhdv-bios

Here's some ideas for those having 2400G/2200G problems.

Next, for those getting a BSOD with things like CPUZ, Cleanmypc, or other utilities, these things generally load a low level filter driver at windows runtime (startup).  AIDA has already had a couple of small updates that stabilized theirs.  Not sure about how fast all the freeware software people can update.   If you get a BSOD with Cleanmypc, then simply don't use it.  Those things are mostly useless (they may make you feel good, but they don't do much that is really useful)

Next, the video problems that seem driver software related and are causing crashes appear to be related to power management.  There is something definitely borked with the current Ryzen power plans, as well as the way the 2400G/2200G are handling low power states, sleep, and screen sleep.  The windows power plans are not working correctly with these CPUs.  They are doing something, but even with Cool and Quiet off (this is mostly unnecessary at this point anyway), and the CPU set to 100% in the power plan, the cores continue to throttle all over the place.

If you overclock, or hard set the CPU multi from the BIOS, and Cool and Quiet are on, then you will get mountains of firmware errors in the event log, because Windows tries to throttle the CPU, but the BIOS fights it and overrides the throttling.

I'm running the same build. Can you do me a huge favor and link to the drivers you used? This thing is driving me insane.