Notice we have only 1 advert? We are supported almost entirely by Patreon and Donations!
Subscribe to GOL Premium
A Bad AMD Catalyst Driver Is Bad For Everyone
Posted , 17 June 2014 at 6:47 am UTC / 8684 views
Attention grabbing headline? Check. It seems the ex-Valve engineer Rich Geldreich has noted a recent Phoronix benchmark using apitest on how badly AMD's Catalyst driver performed.

RichThis excellent GL 4.x micro-benchmark that has been making waves recently is really interesting. Now that it's on Phoronix it's about as mainstream as it's going to get: NVIDIA Slaughters AMD Catalyst On Linux In OpenGL 4.x Micro-Benchmarks

At first glance the results sound great for NV: "The AMD Catalyst driver gets absolutely annihilated for these GL4 micro-benchmarks." But unfortunately it's bad news for everyone working in GL because it clearly demonstrates just how fractured and inconsistent the GL driver landscape actually is when the rubber hits the road.
Source

The first thing to note is that the apitest utility may be Nvidia biased, and in addition to that there are AMD-related bugs in apitest itself. So, looking over the test and peoples comments it does seem like there are bugs in apitest itself that could be causing such low performance.

I personally think AMD's future on Linux is with the open source driver which seems pretty decent and even comparable to AMD's Catalyst driver in some tests. It seems their Catalyst driver just constantly gets outclassed. The only problem with this is that the open source driver always plays catch-up with new hardware releases.

I hope that AMD won't just rely on the open source driver though, as a lot of people still look to Catalyst. With AMD's Mantle API how much effort are they really putting in? Who knows, but tests like this make it look bad.

I am the owner of GamingOnLinux. A fan of anything techy, and not just Linux stuff.

You can follow my personal blog here.

Camoceltic commented on 17 June 2014 at 7:27 am UTC

Catalyst right now is a joke, and a not-very-funny one at that. Not even kidding, my computer is much more stable when I'm running without Catalyst. Just using the stuff that's in a brand new OS install, I can actually play most of my games without worrying about a whole system freeze.

Catalyst right now is a joke, and a not-very-funny one at that. Not even kidding, my computer is much more stable when I'm running without Catalyst. Just using the stuff that's in a brand new OS install, I can actually play most of my games without worrying about a whole system freeze.

Bad Penguin commented on 17 June 2014 at 8:21 am UTC

Here we go again. AMD support is almost non existing for years. But that is ok, I'm not supporting AMD anymore. My computers are an AMD free zone meanwhile.

Here we go again. AMD support is almost non existing for years. But that is ok, I'm not supporting AMD anymore. My computers are an AMD free zone meanwhile.

Anonymous commented on 17 June 2014 at 9:09 am UTC

apitest is not useful as a driver benchmark, only an opengl api usage bechmark. Doing otherwise is just silly, and very untrustworthy.

apitest is not useful as a driver benchmark, only an opengl api usage bechmark. Doing otherwise is just silly, and very untrustworthy.

Guest commented on 17 June 2014 at 10:02 am UTC

When Wayland becomes default display protocol we will see who will be laughing. AMD is already working on Wayland support while nVidia stated that they won't support it because then their Linux drivers will look completely different from their Windows drivers.

When Wayland becomes default display protocol we will see who will be laughing. AMD is already working on Wayland support while nVidia stated that they won't support it because then their Linux drivers will look completely different from their Windows drivers.

Anonymous commented on 17 June 2014 at 11:22 am UTC

paupavWhen Wayland becomes default display protocol we will see who will be laughing. AMD is already working on Wayland support while nVidia stated that they won't support it because then their Linux drivers will look completely different from their Windows drivers.

At least back up a proof that nVidia actually said that idiot...

nVidia claim that they are working on Wayland

http://www.phoronix.com/scan.php?page=news_item&px=MTY0Njc

[quote=paupav]When Wayland becomes default display protocol we will see who will be laughing. AMD is already working on Wayland support while nVidia stated that they won't support it because then their Linux drivers will look completely different from their Windows drivers.[/quote] At least back up a proof that nVidia actually said that idiot... nVidia claim that they are working on Wayland http://www.phoronix.com/scan.php?page=news_item&px=MTY0Njc

FutureSuture commented on 17 June 2014 at 12:40 pm UTC
  • Contributing Editor
  • GOL Supporter

Bad for everyone? Hardly. If AMD's proprietary driver is so poor that it forces people to use AMD's increasingly better open source driver, that's good for everyone. AMD is likely to throw even more developers at its open source driver as well then.

Bad for everyone? Hardly. If AMD's proprietary driver is so poor that it forces people to use AMD's increasingly better open source driver, that's good for everyone. AMD is likely to throw even more developers at its open source driver as well then.

Samsai commented on 17 June 2014 at 4:13 pm UTC

Anonymous
At least back up a proof that nVidia actually said that idiot...

nVidia claim that they are working on Wayland

http://www.phoronix.com/scan.php?page=news_item&px=MTY0Njc
You can get your point across without without having to use the word "idiot". We are trying to be a friendly community here.

[quote=Anonymous] At least back up a proof that nVidia actually said that idiot... nVidia claim that they are working on Wayland [url=http://www.phoronix.com]http://www.phoronix.com[/url]/scan.php?page=news_item&px=MTY0Njc[/quote] You can get your point across without without having to use the word "idiot". We are trying to be a friendly community here.

Anonymous commented on 17 June 2014 at 4:34 pm UTC

"Add new solution, TextureArrayMultiDrawBuffer, to work around possible bug in AMD linux drivers with glMultiDrawIndirect"

this is not something tampered to nvidia. multi_draw_indirect is basic requirement for zero-overhead. i'd say this is workaround so amd doesn't look as bad. in perfect world one shouldn't need it

but, as far as saying it is bad... no, it isn't. Rich simply has some beef with GL as it seems and takes every chance to take a potshot asap. i stopped taking him seriously when it was obvious that his whining is personal nature way more that technical. he did some really good things, but lately it seems like all he is doing is refreshing search "opengl bad"

you don't approach problem with whining that world will collapse like he is doing, you approach it by exposing and then fixing it. and benchmark is doing exactly that

one has to realize that apitest didn't exist not so long ago and as far as i get it, it covers needs for implementing zero-overhead which was not present until not so long ago. right now, these are actually first real benchmarks for those features useful not only to nvidia, but to others as well

it is a call for amd to get their sh1t together and fix it properly. and i kinda don't doubt they will do that very soon, they simply can't afford bad publicity like that or sales will drop in favour of nvidia. just imagine how it looks like when you read it "mantle is 25% faster than others, nvidia is 10 times faster than amd in latest gl features". what does this say about mantle?

"Add new solution, TextureArrayMultiDrawBuffer, to work around possible bug in AMD linux drivers with glMultiDrawIndirect" this is not something tampered to nvidia. multi_draw_indirect is basic requirement for zero-overhead. i'd say this is workaround so amd doesn't look as bad. in perfect world one shouldn't need it but, as far as saying it is bad... no, it isn't. Rich simply has some beef with GL as it seems and takes every chance to take a potshot asap. i stopped taking him seriously when it was obvious that his whining is personal nature way more that technical. he did some really good things, but lately it seems like all he is doing is refreshing search "opengl bad" you don't approach problem with whining that world will collapse like he is doing, you approach it by exposing and then fixing it. and benchmark is doing exactly that one has to realize that apitest didn't exist not so long ago and as far as i get it, it covers needs for implementing zero-overhead which was not present until not so long ago. right now, these are actually first real benchmarks for those features useful not only to nvidia, but to others as well it is a call for amd to get their sh1t together and fix it properly. and i kinda don't doubt they will do that very soon, they simply can't afford bad publicity like that or sales will drop in favour of nvidia. just imagine how it looks like when you read it "mantle is 25% faster than others, nvidia is 10 times faster than amd in latest gl features". what does this say about mantle?

STiAT commented on 17 June 2014 at 5:51 pm UTC
  • GOL Supporter

paupavWhen Wayland becomes default display protocol we will see who will be laughing. AMD is already working on Wayland support while nVidia stated that they won't support it because then their Linux drivers will look completely different from their Windows drivers.

Will it? The platform specific code of the NVidia driver is ~20 %. Since they feature their own OGL implementation which interfaces with the kernel, that part won't be hit, so it's just about interfacing the wayland protocol. I think a company as NVidia will be able to do that in a decent timeline. Even though, for driver vendors it's coming faster, since distributions think about running Wayland and XWayland on top for the desktop already (Fedora in example).

More questionable will be: What about Mir? Looks like not even one driver vendor is looking into that. The current Mir runs with Intel drivers, and custom patches by Canonical. Does not sound very promising to me (and I hope Mir won't get adopted any time soon - it's so stack breaking besides all the technology discussion).

[quote=paupav]When Wayland becomes default display protocol we will see who will be laughing. AMD is already working on Wayland support while nVidia stated that they won't support it because then their Linux drivers will look completely different from their Windows drivers.[/quote] Will it? The platform specific code of the NVidia driver is ~20 %. Since they feature their own OGL implementation which interfaces with the kernel, that part won't be hit, so it's just about interfacing the wayland protocol. I think a company as NVidia will be able to do that in a decent timeline. Even though, for driver vendors it's coming faster, since distributions think about running Wayland and XWayland on top for the desktop already (Fedora in example). More questionable will be: What about Mir? Looks like not even one driver vendor is looking into that. The current Mir runs with Intel drivers, and custom patches by Canonical. Does not sound very promising to me (and I hope Mir won't get adopted any time soon - it's so stack breaking besides all the technology discussion).

  Go to:

Due to spam you need to Register and Login to comment.


Or login with...

Popular this week
View by Category
Contact
Twitter
Latest Forum Posts
Facebook
Who's Online
There are 180 in total online.
Users online: codyaj1995, coryrj19951,
Misc