PSA: AMD's Graphics Driver will Eat One CPU Core when No Radeon Installed

[Caring1]

No compatible hardware = End Task.
How hard is that?

 

[rutra80]

@windwhirl okay but driver is not the place for that and it should be optionally installable

 

[Xzibit]

No compatible hardware = End Task.
How hard is that?

As hard as uninstalling software for hardware you don't use apparently

 

[Caring1]

As hard as uninstalling software for hardware you don't use apparently

The point is, you shouldn't have to uninstall if the software worked correctly.
That would make swapping components a lot simpler and faster.

 

[Xzibit]

The point is, you shouldn't have to uninstall if the software worked correctly.
That would make swapping components a lot simpler and faster.

Well he did the mistake on a SSD test system. If hes doing that on a test systems i.e. not cleaning it while switching hardware and drivers. Its not far fetched what else has happened we havent had a PSA for.

 

[john_]

FPS depends a lot on the game location, or did he play the benchmark? I use actual gameplay. Cities vs outside huge difference. AC:O is also difficult to repro run-to-run because dynamic weather and other random events

He run the game's benchmark. Ultra settings in game, stock settings for the card and Ryzen, 50 fps at 1440p. Your testing reported 39.6 fps.

 

[windwhirl]

The point is, you shouldn't have to uninstall if the software worked correctly.
That would make swapping components a lot simpler and faster.

I have to agree with this, mostly because of a comment I saw elsewhere:

"Imagine if you had to uninstall drivers every time you plug in a new USB device or it would peg your CPU."

 

[rutra80]

Good find, but what are we expecting from a "driver" that has integrated chromium, phone connectivity, streaming app, animated GIFs writer and other steaming-pile-of-horseshit-bloat??

Also imagine how many vulnerabilities must be in all that crap.

 

[Emu]

"RadeonSettings.exe" is Not a driver and it is the cached process. When a user starts the "RadeonSettings.exe" manually to change settings it speeds up loading of UI part of the software.

>>Looks like AMD is doing things differently and just pre-loads Radeon Settings in the background every time your system is
>>booted and a user logs in, no matter if AMD graphics hardware is installed or not. It would be trivial to add a check
>>"If no AMD hardware found, then exit immediately", but ok. Also, do we really need six entries in Task Scheduler?

It is a well known issue! AMD is one of the companies that secretly (!) creates an entry in Windows Task Scheduler to load "RadeonSettings.exe" every time a Windows OS starts. It actually violates your privacy.

A solution is very simple: Disable it in Windows Task Scheduler or delete. Personally, I disable it and check regularly that it is Not enabled again.

A quick look at my Task Scheduler shows 10 entries put there by Nvidia, 4 by Microsoft, 2 by Google (their update service for Chrome), 1 from the developer of the "Control Station" software for my computer and 1 from Adobe (I didn't even realise that I had Adobe Acrobat installed!?). Pretty sure that Task Scheduler is the new way of starting programs on startup given that I do not even have a "Startup" folder in my Start menu anymore and haven't had one for quite a few years now. I actually quite like it given that I don't have to drill down into the registry to the "run" key in RegEdit and I can disable the tasks pretty easily if I want to without having to delete anything.

 

[john_]

About Chrome. How about Software Reporter tool ?

 

[R-T-B]

Possibly a little dramatic IMHO.

I consider this completely newsworthy and not at all dramatic. It makes me wonder what else could be going on in their code, frankly. No one here seems to be comprehending how amateurish this is.

Also imagine how many vulnerabilities must be in all that crap.

There's already one that was known for 6 months they didn't even patch bother to patch until the 2020 super waterfall release. I don't imagine they are very strong in that department.

 

[mouacyk]

I consider this completely newsworthy and not at all dramatic. It makes me wonder what else could be going on in their code, frankly. No one here seems to be comprehending how amateurish this is.

Even though the work-around is simple (uninstall, ddu, etc.), the point of code quality is quite valid considering how many issues they've had with their drivers earlier this year.

 

[R-T-B]

Even though the work-around is simple (uninstall, ddu, etc.), the point of code quality is quite valid considering how many issues they've had with their drivers earlier this year.

I used an AMD card recently and noticed huge singlethreaded CPU overhead vs NVIDIA. It really hurt me in Kerbal Space Program, a single threaded game that really needs all that ST performance.

It does make me wonder WTF is going on in there.

As hard as uninstalling software for hardware you don't use apparently

You are missing the point.

 

[Xzibit]

You are missing the point.

They point that he was on a test system and didn't clean it properly. He caught the mistake and investigated it.

He can't correct the software side only inform but if hes making similar lapses on a bigger scale well it call into question a lot more things.

 

[R-T-B]

They point that he was on a test system and didn't clean it properly. He caught the mistake and investigated it.

No the point is this code mistake is so rookie level that it means the drivers likely are chock full of god knows what in terms of code quality.

Or at least, that's what some of us are walking away with.

He can't correct the software side only inform but if hes making similar lapses on a bigger scale well it call into question a lot more things.

lol, no dude. This one ain't flippable.

 

[Nephilim666]

As someone upgrading from Vega 64 to RTX 3090 (AMD paper launch antics talked me out of 6900XT) I really appreciate this post, I'm terrible at cleaning up unused drivers/apps but I might do a clean win install after seeing this.

 

[TheGoddessInari]

It's great that multi-GPU bugs are left in GPU-Z to arm wave about this...? /s

But seriously, it's the settings application getting confused. This has nothing to do with the driver, which won't even load if the hardware's not attached. There's nothing stealthy or secret about this. Intel and Nvidia have similar processes that start up.

I like AMD, and they haven't gotten everything right, but at least they support open source, and don't push WBINVD instructions like Nvidia does. Putting all of this hate on them inappropriately is absurd. I really miss when tech journalism was journalism. Investigative, not just someone's personal opinion blog.

TPU is a nice site, but this post is pretty silly. An application that goes with hardware you've uninstalled wasn't removed, and doesn't know the hardware was removed. Big shock. It affects no one in the real world.

 

[TheoneandonlyMrK]

I consider this completely newsworthy and not at all dramatic. It makes me wonder what else could be going on in their code, frankly. No one here seems to be comprehending how amateurish this is.



There's already one that was known for 6 months they didn't even patch bother to patch until the 2020 super waterfall release. I don't imagine they are very strong in that department.

A few are hammering the message out, it'll get there.

But still.

A few have indeed commented how amateurish it is leaving the installed suite on is too, and yet here we are, last story for the whole weekend and certain to stay on the front page all weekend, interesting time to post it, seems clickbaity and a cry for clicks ,or heated debate but many like that sort of thing eh.

 

[R-T-B]

A few are hammering the message out, it'll get there.

But still.

A few have indeed commented how amateurish it is leaving the installed suite on is too, and yet here we are, last story for the whole weekend and certain to stay on the front page all weekend, interesting time to post it, seems clickbaity and a cry for clicks ,or heated debate but many like that sort of thing eh.

Maybe. It is interesting (shocking, rather) to the programmer in me, which may bias me.

But I'd argue that bias is based in knowledge of the issue, which is... not a good sign?

But seriously, it's the settings application getting confused. This has nothing to do with the driver

Except they are literally shipped together in the driver installer, which doesn't even provide an option to install them seperately. That's just trying to play semantics now. To the end user, it's the same darn thing.

Big shock. It affects no one in the real world.

This might not. I'm sure based on this finding though there are probably other uglies that really, really do waiting to be discovered. Makes me glad I jumped.

 

[ScaLibBDP]

[ W1zzard ] Wrote:
>>...I wondered why the machine's performance was SO sluggish...

Be careful when doing performance evaluations because such issues could skew results. Please take a look at a Video Technical Report dedicated to performance evaluations and how to do it correctly:

The Art of Processing Power measurements of a Computer System

Best regards

 

[srsbsns]

Here is the problem with this whole thing and the logic involved here. If you don't have an AMD card installed and you try to install the drivers it will stop and not install because there is no compatible device. By installing the drivers using an AMD card then swapping to an Nvidia card you are literally working around what AMD has clearly told you not to do by it's installer giving you a hard stop. Nitpick about code all you want but nothing changes the fact it said "dont do this" and you did it anyway.


You cant fix stupid and this article is just that.

 

[r9]

You better explain yourself what you doing with dual core!

 

[wiak]

this reminds me of that time a friend of mine switched from intel 4 core to amd ryzen 1700X 8-core and notice that he didn't get all the performance, it was due to the intel chipset driver had hardcoded the cpu cores to 4 core in msconfig> boot > advanced

 

[windwhirl]

this reminds me of that time a friend of mine switched from intel 4 core to amd ryzen 1700X 8-core and notice that he didn't get all the performance, it was due to the intel chipset driver had hardcoded the cpu cores to 4 core in msconfig> boot > advanced

Interesting. I thought the norm was to do a clean install when you switched CPU and motherboard...

 

[R-T-B]

[ W1zzard ] Wrote:
>>...I wondered why the machine's performance was SO sluggish...

Be careful when doing performance evaluations because such issues could skew results. Please take a look at a Video Technical Report dedicated to performance evaluations and how to do it correctly:

The Art of Processing Power measurements of a Computer System

Best regards

I'm pretty sure W1zzard, who has been doing reviews since before youtube was a thing, is aware.

Here is the problem with this whole thing and the logic involved here. If you don't have an AMD card installed and you try to install the drivers it will stop and not install because there is no compatible device. By installing the drivers using an AMD card then swapping to an Nvidia card you are literally working around what AMD has clearly told you not to do by it's installer giving you a hard stop. Nitpick about code all you want but nothing changes the fact it said "dont do this" and you did it anyway.

So what your saying is: AMD literally tells it's users "never switch brands, it's not supported?"

lol, no they don't and I hope they never do.

You better explain yourself what you doing with dual core!

Irritating irritable people, aparently?

Interesting. I thought the norm was to do a clean install when you switched CPU and motherboard...

I've also used near all intel cpus and chipset combos and can't recall any doing what he said. If true, it sounds like a vendor modification.