AMD Responds to Lack of Ryzen Mobile Driver Updates, Claims OEMs are the Issue

[Darmok N Jalad]

I suspect it’s like many Android handsets—the margins are so slim that there isn’t much room in the budget to have a team of developers working on anything but bug fixes.

 

[WikiFM]

I have an Intel 8th gen ULV chip, which last got a (IGP) driver update well over a year back. If Intel can't force OEM updates, what chance does AMD have?

Intel does supplies generic drivers for most mobile CPUs (Ivy Bridge and newer). I know this because I updated a few weeks ago my Haswell ULV with a new driver.

 

[bug]

X series had their final driver released in 2010 (Windows Vista).
HD 2/3/4 series had driver support until 2013 (Windows 8.1).
HD 5/6 series had driver support until 2015 (Windows 10).
HD 7/8 (debuted 2012) and R# series have driver support through today. GCN definitely has everything NVIDIA ever released beat in terms of driver support.

X series supported from 2005 to 2010. 5 years.
HD2/3/4 series supported from 2007 to 2013. 6 years.
HD 5/6 series supported from 2009 to 2015. 6 years.
HD 7/8 series supported since 2010. 8 years.

I have picked one card at random from Nvidia and it was supported for 11 years. Yet somehow it felt natural to you to post the above and conclude "GCN definitely has everything NVIDIA ever released beat in terms of driver support". Since I don't think this is the place to settle how arithmetic works, I think we should stop now.

 

[FordGT90Concept]

You forget that GCN cards are still being released. By the time the HD 7 series stops getting drivers, it will be >11 years.

To be fair, GeForce 6 series went legacy driver support in 2013 (R304), so 9 years. They released one driver (R309) in 2015 because of security vulnerabilities in R304.

GeForce 6/7 (Windows 8), 2013 (2015 security update) (7-9 years excluding security update)
GeForce 8/9/100/200/300 (Windows 10), 2016 (8-10 years)
GeForce 400/500 (Windows 10), 2018 (6-8 years)

NVIDIA does tend to support drivers longer than AMD/ATI does. I think this can be explained by ATI especially tending to create new architectures frequently where NVIDIA likes to refine. When AMD took to refining, the companies have similar driver lifecycles.

I've used ATI and NVIDIA cards since the 1990s. I've never once complained about inadequate driver support because cards 3+ years old are usually functionally obsolete.

 

[ValenOne]

I have an Intel 8th gen ULV chip, which last got a (IGP) driver update well over a year back. If Intel can't force OEM updates, what chance does AMD have?

From device manager, I force driver update on my Surface Pro 4, but once Intel's driver is installed, I can update drivers via the normal install method.

Intel's generic mobile driver monthly updates intervals are better than AMD's non-existent generic mobile driver updates.

On my HP Envy x360 15z bq-100-CTO with Ryzen 5 2500U, it's HP supplied web site drivers are still in Nov 20, 2017 (version 22.19.655.1 which can BSOD via Chrome). I looked into other recent HP Envy x360 models with Ryzen APU like 15z-cp000 CTO which has newer drivers version 23.20.821.2560 (Aug 2, 2018).

Using desktop APU driver version 24.xx.xxx.xxxx which seems to be higher version than 23.xx.xxx.xxxx. I used force update method for 24.xx.xxx.xxxx drivers.

Microsoft's supplied drivers for my Surface Pro 4 has better update intervals when compared to my HP Envy x360 15z bq-100-CTO.

I wouldn't recommend AMD laptop for non-IT my family members e.g. I purchased multiple Surface Pros for them.

I can install generic mobile Radeon drivers for my Radeon HD 8870M (renamed into R9-M270 series) and one wonders why the difference in AMD's mobile driver update policy between mobile APU and my laptop's Radeon HD 8870M.

Oh sure. RTG actually works on improving dGPU drivers. If that were the case i would not have moved on from FuryX. Bugs upon bugs upon bugs. Every single driver after 18.5.1 broke VR for me l, submitted tons of reports and only email i got was “noted”

No, hell freaking no. I have given RTG way too many chances. They are no longer the ATi i used to like.

I have no problems updating my R9-390X desktop drivers(for ground floor's HDTV), but AMD mobile driver updates are problematic which needs some IT level support skills.

Yes, when you include GPUs and similar generic PCI Express-based hardware. When looking specifically at APUs, that's not the case. People have tried AMD's generic driver on these machines (and E-#50 previously) and the generic drivers have lots of problems. OEM-tailored drivers are absolutely required because the hardware changes they made are extensive.

I'm actually running AMD's desktop APU driver version 18.10.1 on my HP Envy x360 15z bq-100-CTO (Ryzen 5 2500U) via device manager force update method and it works fine.

 

[FordGT90Concept]

On my HP Envy x360 15z bq-100-CTO with Ryzen 5 2500U, it's HP supplied web site drivers are still in Nov 20, 2017 (version 22.19.655.1 which can BSOD via Chrome). I looked into other recent HP Envy x360 models with Ryzen APU like 15z-cp000 CTO which has newer drivers version 23.20.821.2560 (Aug 2, 2018).

Thanks for the details. I passed them on.

 

[R0H1T]

Intel does supplies generic drivers for most mobile CPUs (Ivy Bridge and newer). I know this because I updated a few weeks ago my Haswell ULV with a new driver.

Which doesn't work if you have OEM drivers installed, at least with win10 it won't. You have to uninstall the device & delete the drivers manually.

 

[eidairaman1]

AMD should´ve learned with Bulldozer that relying on others is a big mistake.

Support the stuff you build, nobody else will care about your stuff.
Others sell it once, but on the long term your Firmwares and Drivers and Software evolves, that could aid you via OEM if you do the support by yourself.
AMD needs to better its reputation at the endconsumer on every field.

"Don´t be the Alternative, be the better Choice."
SIMPLE

990fx had great drivers. But like intel laptops drivers are nil

 

[john_]

This statement from AMD reminds me of Nvidia's statement for PhysX.

Nvidia
"We can't warranty compatibility when AMD GPU is primary, so we lock it".
AMD
"We can not warranty compatibility with every laptop, so we go to sleep"

I was saying in the past for Nvidia that they could let PhysX unlocked as a beta and with no support. AMD could do the same. Everything else is just BS.
That from someone who supports AMD but also acknowledges that AMD is in many cases shooting it's own feet.

 

[bug]

You forget that GCN cards are still being released. By the time the HD 7 series stops getting drivers, it will be >11 years.

To be fair, GeForce 6 series went legacy driver support in 2013 (R304), so 9 years. They released one driver (R309) in 2015 because of security vulnerabilities in R304.

GeForce 6/7 (Windows 8), 2013 (2015 security update) (7-9 years excluding security update)
GeForce 8/9/100/200/300 (Windows 10), 2016 (8-10 years)
GeForce 400/500 (Windows 10), 2018 (6-8 years)

NVIDIA does tend to support drivers longer than AMD/ATI does. I think this can be explained by ATI especially tending to create new architectures frequently where NVIDIA likes to refine. When AMD took to refining, the companies have similar driver lifecycles.

I've used ATI and NVIDIA cards since the 1990s. I've never once complained about inadequate driver support because cards 3+ years old are usually functionally obsolete.

Neah, ATI/AMD has quite a chequered history when it comes to drivers. I believe their shorter support cycles are due to them having to rewrite their drivers more often. But kudos to them, their drivers today are in a far better place than they were 10 years ago.

 

[eidairaman1]

The biggest problem here is that some OEMs like Lenovo block the instalation of non-signed drivers, meaning that you can install only drivers from Lenovo website, which is any way you put it, stupid. I have got a HP pavilion laptop with 6700HQ and nvidia gtx 950m and I installed all my drivers from Intel/Nvidia website, no problem.

Bios black listing

 

[ValenOne]

Searching on the internet, I get the impression not many people actually have problems with the HP Envy x360. It appears to be one person that's driving the noise. Others might be jumping on the bandwagon just because they'd like to see APUs have a driver release cadence like GPUs do but that's not going to happen because of cost.

It might be that Envy's firmware requires HP-signed drivers to work. As pointed out previously in this thread, Lenovo has done that. Apple too.

Generic AMD drivers needs INF edit to insert HP's hardware IDs, but Windows 64bit blocks unsigned modified driver installs.

 

[bug]

Generic AMD drivers needs INF edit to insert HP's hardware IDs, but Windows 64bit blocks unsigned modified driver installs.

That's true, but before Windows started enforcing drivers being signed, a modified INF was all it took to make a driver work with laptop parts. That's why I'm taking this all "mobile parts need drivers with a different magic in them" claim with a grain of salt.

 

[ValenOne]

990fx had great drivers. But like intel laptops drivers are nil

Intel's monthly mobile driver updates are fine with Surface Pro 4 with 620 IGP and Surface Pro 5 with 640 Iris Plus.

 

[Valantar]

Intel's monthly mobile driver updates are fine with Surface Pro 4 with 620 IGP and Surface Pro 5 with 640 Iris Plus.

I haven't exactly been paying attention much, but I've noticed both Intel-based laptops in my house getting iGPU driver updates through Windows update with some regularity.

 

[eidairaman1]

I haven't exactly been paying attention much, but I've noticed both Intel-based laptops in my house getting iGPU driver updates through Windows update with some regularity.

Windows update, pretty laughable

Intel's monthly mobile driver updates are fine with Surface Pro 4 with 620 IGP and Surface Pro 5 with 640 Iris Plus.

Isn't that a Microsoft product to begin with?

 

[Valantar]

Windows update, pretty laughable

Laughable? Because they use a built-in, automatic, zero effort update system? IMO, that's how it ought to work.

 

[eidairaman1]

Laughable? Because they use a built-in, automatic, zero effort update system? IMO, that's how it ought to work.

Yup, breaks the system to boot too

 

[TheinsanegamerN]

Underdogs can absolutely do wrong, but they do of course have less power to abuse. The very existence of this thread ought to tell you that nobody here is that blind. Also, RTG's dGPU driver effort in the last couple of years has been excellent. Which is a big part of why this is so baffling.

Except it isnt baffling.

The biggest reason RTG's drivers have improved so much is the architecture itself. GCN has been sold since 2011 in some form, and newer forms of GCN are very similar to GCN 1.0 in operation, hence why the whole "finewine" meme exists, AMD's 7 year old cars are still optimized for because they are the same arch. 7 years of development, and even minimum wage code monkeys can usually fix most problems, giving the illusion they are doing magically better. We saw this with vega, performance issues and driver problems all over the place for months that shouldnt have been in the final driver. We saw this with the 300 series, with black screen issues still being present. AMD's driver team hasn't magically become competent, they are just working with such old tech that even ignoramuses could support it properly.

You also see a lot of people in this thread trying to sidestep the issue, pointing fingers at intel or OEMS. The fact is, I can update my intel laptop to the latest version of their iGPU driver, regardless of what the OEM offers on their site, and it works. Occasionally you will have a machine that has issues, but 99% of the time it just works, because the chips are the same regardless of manufacturer. Similarly, most gaming laptops with nvidia chips can take the latest driver straight from nvidia's website, and the occasional machine that cant is lambasted by the community until the OEM works with nvidia to resolve the issue. AMD's previous APUs had GPU drivers independent of the machine itself, Llano, trinity, and richland were all like this. There is 0 excuse for the GPU driver for ryzen mobile being tied to a OEM driver package, that should be a separate driver.

What AMD tried to do here is the same BS that ATi pulled in the 90s, they would make a GPU and outsource the driver development to the OEM, resulting in a lack of driver updates and numerous bugs. AMD tried the same thing with ryzen mobile because they know they dont have the driver team to support it, so they try and get OEMs to do it while giving brain-dead responses to customers about how the OEM is responsible for AMD's driver, and just like the 90s it is backfiring on them. If AMD was dumb enough to sign a deal preventing them from releasing drivers for their OWN HARDWARE, they deserve to go out of business. That kind of idiocy wouldn't fly for ANY OEM other them AMD.

The only response AMD should have given was admitting they screwed up and go and fix their friggin driver package. Separating drivers for different pieces of silicon isnt hard, and the only way AMD hasn't fixed this is sheer incompetence, which many of us expect from AMD at this point.

 

[Valantar]

Except it isnt baffling.

The biggest reason RTG's drivers have improves so much is the architecture itself. GCN has been sold since 2011. 7 years of development, and even minimum wage code monkeys can usually fix most problems, giving the illusion they are doing magically better. We saw this with vega, performance issues and driver problems all over the place for months that shouldnt have been in the final driver. We saw this with the 300 series, with black screen issues still being present. AMD's driver team hasn't magically become competent, they are just working with such old tech that even ignoramuses could support it properly.

You also see a lot of people in this thread trying to sidestep the issue, pointing fingers at intel or OEMS. The fact is, I can update my intel laptop to the latest version of their iGPU driver, regardless of what the OEM offers on their site, and it works. Occasionally you will have a machine that has issues, but 99% of the time it just works, because the chips are the same regardless of manufacturer. Similarly, most gaming laptops with nvidia chips can take the latest driver straight from nvidia's website, and the occasional machine that cant is lambasted by the community until the OEM works with nvidia to resolve the issue. AMD's previous APUs had GPU drivers independent of the machine itself, Llano, trinity, and richland were all like this. There is 0 excuse for the GPU driver for ryzen mobile being tied to a OEM driver package, that should be a separate driver.

What AMD tried to do here is the same BS that ATi pulled in the 90s, they would make a GPU and outsource the driver development to the OEM, resulting in a lack of driver updates and numerous bugs. AMD tried the same thing with ryzen mobile because they know they dont have the driver team to support it, so they try and get OEMs to do it while giving brain-dead responses to customers about how the OEM is responsible for AMD's driver, and just like the 90s it is backfiring on them. If AMD was dumb enough to sign a deal preventing them from releasing drivers for their OWN HARDWARE, they deserve to go out of business. That kind of idiocy wouldn't fly for ANY OEM other them AMD.

The only response AMD should have given was admitting they screwed up and go and fix their friggin driver package. Separating drivers for different pieces of silicon isnt hard, and the only way AMD hasn't fixed this is sheer incompetence, which many of us expect from AMD at this point.

a) Why are you framing this as a response to me, when the key points (outside of smearing a development team) in what you're saying are things I've said myself earlier in this thread?
b) Pointing fingers at Intel? What? Who? Where? How? I'd like to see how that works, considering they have nothing to do with Ryzen Mobile drivers.
c) So you're denying that AMD have for periods in the last few years outpaced Nvidia in game-ready and WHQL driver releases? 'Cause they have. Have there been bugs? Sure, but nothing worse than Nvidia. This is pretty much to be expected with cutting-edge hardware. Still, matching your 5-10x bigger competitor is quite all right.
d) Your "this is so old tech anyone could work" and your examples of new hardware launches having issues contradict each other. Obviously there are big enough differences between versions of GCN to matter from a driver POV.
e) Nvidia hasn't made an iGPU since those were located in chipsets. Not applicable. Also addressed earlier in the thread.
f) You really ought to have read this thread more carefully before posting. All you're doing is repeating what's already been said, but in a less constructive and more vitriolic tone. What's the point?

There's no doubt AMD needs to fix this, and divorcing iGPU drivers from the general APU driver package (and, arguably, integrating them into regular GPU driver packages) is an obvious necessity. What contract terms they are or aren't subject to is unknown, but regardless of that, even two updates a year is weak. Rants like yours, though, don't help anyone.

 

[FordGT90Concept]

Occasionally you will have a machine that has issues, but 99% of the time it just works, because the chips are the same regardless of manufacturer.

That's also the case here. As far as I know, it's only the HP Envy x360 that doesn't permit AMD's generic driver package to be used. There's a lot of other laptop models out there with Ryzen Mobile where it installs fine.

The machines work, just not ideally. There's no legal framework to even go after HP or AMD over this.

 

[bug]

That's also the case here. As far as I know, it's only the HP Envy x360 that doesn't permit AMD's generic driver package to be used. There's a lot of other laptop models out there with Ryzen Mobile where it installs fine.

The machines work, just not ideally. There's no legal framework to even go after HP or AMD over this.

Why would you say that? Neither the original news nor AMD's statement single out HP (much less one HP model). What do you know that we don't?

 

[FordGT90Concept]

All Ryzen Mobile APUs are on a 6 month WHQL schedule and that isn't going to change (AMD doesn't do beta for APUs because of QA costs). The only Ryzen Mobile system that hasn't gotten an update in more than 6 months, as far as I know, is the HP Envy x360. Every SKU of the x360 gets it's release driver forked from AMD and then never updated thereafter. Newer x360s have fewer issues than older x360s simply because of having a more mature driver.

 

[eidairaman1]

All Ryzen Mobile APUs are on a 6 month WHQL schedule and that isn't going to change (AMD doesn't do beta for APUs because of QA costs). The only Ryzen Mobile system that hasn't gotten an update in more than 6 months, as far as I know, is the HP Envy x360. Every SKU of the x360 gets it's release driver forked from AMD and then never updated thereafter. Newer x360s have fewer issues than older x360s simply because of having a more mature driver.

To me the fault lies with Hewlett-Packard

 

[Valantar]

All Ryzen Mobile APUs are on a 6 month WHQL schedule and that isn't going to change (AMD doesn't do beta for APUs because of QA costs). The only Ryzen Mobile system that hasn't gotten an update in more than 6 months, as far as I know, is the HP Envy x360. Every SKU of the x360 gets it's release driver forked from AMD and then never updated thereafter. Newer x360s have fewer issues than older x360s simply because of having a more mature driver.

To me the fault lies with Hewlett-Packard

That sure sounds like HP is... uhm, involved in the problem, yes. Have to suppose that AMD probably could have forced through a more consistent t update policy, but I doubt they're willing to fight for that when they're barely getting back in the laptop game and have to focus on wooing OEMs for more design wins. A shame, though, as it ultimately hurts everyone involved.