Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Chromium blacklists Nouveau

lists.freedesktop.org

Post reply on HN
Print view View original

381–390 of 411 posts

Re: Chromium blacklists Nouveau

#381

Post by frabbit »

dekhn wrote: post #365
frabbit wrote: post #359

Earlier quoted context omitted.

Is that science? A bunch of computing units running secret code?! Reproducibility in computational science is recognized as a huge problem: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3383002/ A lab or institution claiming a publishable result based on discounted hardware from the manufacturer running code that does who knows what for their TensorFlow/whatever calculations is engaged in something other than science.…

let's step back from the brink. Plenty of good science is done on proprietary foundations. It would be unrealistic to do otherwise. Reproducibility doesn't necessarily mean every component of the system is completely open and inspectable.

   Reproducibility doesn't necessarily mean every component of the system is completely open and inspectable.
If it's not "open and inspectable" then it cannot be reproduced by another group/researcher. It's just unsupported claims building on a wonky methodology.

I wonder what percentage of "Machine Learning" results are based on such shaky foundations and when they'll receive the same amount of skepticism and scorn we see heaped(deservedly) on the field of psychology.

Re: Chromium blacklists Nouveau

#382

Post by drb91 »

geofft wrote: post #99
drb91 wrote: post #86

Earlier quoted context omitted.

> I think it makes sense for the browser to attempt to deliver the best experience it can to users Are you talking about chrome, the least user friendly of all browsers? This is just corporate talk for “we don’t care about making our software work with nouveau”.

I think Chrome is empirically quite user friendly - it's not installed by default anywhere and people choose to use it. It is true that Chrome isn't particularly power -user friendly for certain use cases, but that's different. Also, if you look at the bug report, it's quite clear they spent a lot of time trying to make nouveau work.

Well, people have always used bad software; popularity has never been a good metric of quality. Nothing about Chrome protects users from Google. This is a fundamental role of the browser.

Re: Chromium blacklists Nouveau

#383

Post by drb91 »

PhasmaFelis wrote: post #143
jlarocco wrote: post #125

Earlier quoted context omitted.

There's more to being user friendly than having a dumbed down UI and being technically correct. Helping corporations spy on users is definitely not user friendly, and Chrome is by far the worst offender there. Saying, "nobody cares," because they're ignorant that it's going on is no excuse, either.

User-friendliness is usually understood as being about usability and interface design. Spyware-riddled software may be shitty, and that should be called out, but it's orthogonal to user-friendliness. Edit: You seem to disagree. Google "user-friendly" and realize that I'm not defending Chrome.

Just because corporate language doesn’t include any way of expressing representing consumer and user interests outside of what can be represented in a transaction doesn’t mean it isn’t worth expressing.

Re: Chromium blacklists Nouveau

#384

Post by staktrace »

shmerl wrote: post #379
staktrace wrote: post #377

Earlier quoted context omitted.

Please don't encourage people to turn on WebRender without at least warning them that it is experimental, still under development, can cause your computer to blow up, etc. Also the environment variables you specified will force it on even on non-Nightly builds which is very undesirable since those builds are not getting fixes uplifted. (I work on WebRender for Firefox)

I've been using WebRender on beta builds on Linux for several months already - it works quite well (AMD / radeonsi). WebRender is more raw indeed, but regular layers acceleration should surely be re-evaluated instead sweeping blacklisting like it happens now.

Always remember the YEINU principle: https://blog.mozilla.org/nnethercote/2014/01/02/yeinu-your-e...

Re: Chromium blacklists Nouveau

#385

Post by alacombe »

shmerl wrote: post #185

I'd say, if you care about using open source driver, you'd avoid Nvidia to begin with, and use AMD or Intel instead. Here is a very good post from one of the key Nouveau developers, about why you should do exactly that: https://www.phoronix.com/forums/forum/linux-graphics-x-org-d...

AMD cards for anything advanced are crap, they couldn't recogize one of my monitor (>2 monitor setup), and they're not playing fair with their AMF extension on Linux (works great on Windows though). Even on Windows, I cannot get the card to properly get my minitor resolutions (it's sitting beyond a multiplexer, though there no problem with nvidia).

Re: Chromium blacklists Nouveau

#386

Post by alacombe »

altfredd wrote: post #376
atemerev wrote: post #321

Earlier quoted context omitted.

AMD has their own set of quirks, they are not open sourcing their main driver either. Nouveau was always a stopgap measure, to boot the system and make it somewhat usable until you install the official binary driver; so is amdgpu. Even if you are open source enthusiast who wants nothing but FOSS on your system (and this is a very minor use case), these drivers do no better than your integrated Intel graphics, which i…

"Nouveau and amdgpu were always barely working" — nah, you are spreading misleading information and lumping together completely unrelated projects at different stages of development. Noveau has been almost close to usable at some point... Until Nvidia started the whole signed firmware nonsence. Then it quickly became intolerable due to Nvidia, supplying no signed firmware or heavily crippled firmware. These days you…

I gave AMD a try with a couple different card, couldn't get them working properly. One was on a dual GPU setup, AMD failed, nvidia worked. Then I needed hardware accelerated HEVC encoding, bought a rv560, works great on windows, but AMD isn't playing fair with AMF on linux, once again failed by AMD, and I didn't want to install the proprietary nvidia drivers to use VDPAU, so that's a stalemate. Then I tried the AMD card on a split virtual monitor setup, once again AMD failed me, couldn't use both dvi & display port at the same time. Now I run a 1060 with nouveau and some quirk, best experience so far. Oh, and on windows, my rv560 can't properly get the resolution of the monitor behind an hdmi demux, the previous nvidia card worked like a charm.

AMD 0 - Nvidia 3

Re: Chromium blacklists Nouveau

#387

Post by MaulingMonkey »

frabbit wrote: post #381
dekhn wrote: post #365

Earlier quoted context omitted.

let's step back from the brink. Plenty of good science is done on proprietary foundations. It would be unrealistic to do otherwise. Reproducibility doesn't necessarily mean every component of the system is completely open and inspectable.

Reproducibility doesn't necessarily mean every component of the system is completely open and inspectable. If it's not "open and inspectable" then it cannot be reproduced by another group/researcher. It's just unsupported claims building on a wonky methodology. I wonder what percentage of "Machine Learning" results are based on such shaky foundations and when they'll receive the same amount of skepticism and scorn we…

> If it's not "open and inspectable" then it cannot be reproduced by another group/researcher.

That doesn't follow. They can (attempt to) reproduce with the same hardware, firmware, and drivers. They can reproduce with similar hardware from another vendor (AMD). If you're using OpenCL, you can reproduce on CPUs, albeit much slower and probably only for partial results. Maybe you want to break out some pen and paper to verify a few calculations of the CPU for even more partial results because you don't trust Intel's firmware blobs, even though you've presumably cross checked against AMD hardware, but you've gotten quite silly at that point.

Maybe some of these changes won't give you bitwise identical results - but then again, neither does most science. You check if the broad results and patterns are statistically similar. Maybe they are, and the results survive a large variety of similar but different test setups. Maybe they don't, and you discover an uncontrolled variable, like the exact chemical composition of the surface of your glass beakers being contaminated by impurities, or your AI being sensitive to the exact floating point rounding behavior of your GPU.

If anything, you want to see if your results survive variations in the uncontrolled variables, to make sure you've properly determined what factors are controlling the results.

Re: Chromium blacklists Nouveau

#388

Post by alacombe »

imtringued wrote: post #349
atemerev wrote: post #321

Earlier quoted context omitted.

AMD has their own set of quirks, they are not open sourcing their main driver either. Nouveau was always a stopgap measure, to boot the system and make it somewhat usable until you install the official binary driver; so is amdgpu. Even if you are open source enthusiast who wants nothing but FOSS on your system (and this is a very minor use case), these drivers do no better than your integrated Intel graphics, which i…

amdgpu is not a stopgap driver. It's mature and reliable enough driver to play videogames including a lot of windows games via proton. I don't see a single reason to use a proprietary OS or driver anymore.

As long as you don't have a funky setup, sure, but I've never been able to get AMD cards properly working for what I wanted to do (dual GPU setup, AMF on Linux, and multi-virtual monitors, ie 2 simultanious display on 1 monitor, or even getting the proper resolution from monitor behind an hdmi demux) , both with the opensource `amdgpu` AND the proprietary Linux drivers.

Re: Chromium blacklists Nouveau

#389

Post by alacombe »

cheald wrote: post #12
bubblethink wrote: post #10

This sucks. Chromium doesn't do accelerated video decode for intel on linux either. And now this. What's the AMD scene like ?

The amdgpu drivers in Linux are fantastic these days. Works out of the box with most hardware. No proprietary binaries needed, other than the firmware blobs. AMD is doing a really solid job of supporting their Linux user base.

No multi-GPU support (at least 2 years ago, using the binary drivers), no AMF support on Linux, and couldn't get multi-display support on a split screen. AMD is fail for me...

Re: Chromium blacklists Nouveau

#390

Post by alacombe »

zanny wrote: post #170
bubblethink wrote: post #10

This sucks. Chromium doesn't do accelerated video decode for intel on linux either. And now this. What's the AMD scene like ?

https://wiki.archlinux.org/index.php/chromium#Hardware_video... You can get hardware video on Intel and AMD.

Only decode, there is no va-api support for AMF, say to get hardware HEVC encoding.
Post reply on HN