Slashdot Mirror
IE9 Team Says "Our GPU Acceleration Is Better Than Yours"
An anonymous reader writes "Over on the IE blog Microsoft's Ted Johnson writes, 'With IE9, developers have a fully-hardware accelerated display pipeline that runs from their markup to the screen. Based on their blog posts, the hardware-accelerated implementations of other browsers generally accelerate one phase or the other, but not yet both. Delivering full hardware acceleration, on by default, is an architectural undertaking. When there is a desire to run across multiple platforms, developers introduce abstraction layers and inevitably make tradeoffs which ultimately impact performance and reduce the ability of a browser to achieve 'native' performance. Getting the full value of the GPU is extremely challenging and writing to intermediate layers and libraries instead of an operating system's native support makes it even harder. Windows' DirectX long legacy of powering of the most intensive 3D games has made DirectX the highest performance GPU-based rendering system available.' Some Mozillians hit back in the comments to the IE Blog post and others have written blog posts of their own. PC Mag's Michael Muchmore seems to conclude that IE9 and Firefox 4 are more or less the same (despite the title of his article) while Chrome currently lags behind."
If any program makes your GPU drivers crash, then take it up with the GPU manufacturer. If the drivers are crashing, then they're defective.
Unfortunately, Firefox's stubborn refusal to pass Acid3 legitimizes IE9's stubborn refusal to do the same. The dev team needs to swallow its pride and implement the standards.
That's misleading. IE9 gets something like 96/100 in the Acid3 test.
That's absolutely OK for most practical purposes.
That's fine, but realise that the web is about hypertext. Shockwave and flash are supposed to be on the web in the same way that movies and sounds are: as embedded elements of media. Building an entire site or app in shockwave or flash is NOT building for the web, it's only running a non-web app over http.
I tried to submit something through the feedback thing, but as far as I can tell, things written there go nowhere, so who knows.
No, we read pretty much all the feedback (through filtered and clustered searches) -- the volume is very high, and so we can't respond to individual comments, though.
We are aware of the issue with hardware acceleration on certain setups. Try updating your graphics card drivers and try again?
-- Alexander Limi, Firefox User Experience Team
If IE8 is any indication, Firefox comes a damn sight closer to passing.
Not perfectly in compliance, granted, but really rather close when compared to what it looked like in IE for me.
Firefox does 97, IE9 does 95 on Acid3.
The GPU, as it's normally on PCIe these days, has DMA capabilities. On most (all?) x86 systems DMA isn't restricted through an MMU, unlike CPU memory access. This means that by sending the correct commands to the GPU you can access any part of the system memory.
If this is possible in reality I have no idea, but that's the concept.
--sitharus
Acid3 isn't a particularly useful test for real world performance and the folks doing the coding and development were right to push it down the list. It uses deliberately broken code to see how the browser handles it. Handling broken code is a bad idea, just make sure it fails without causing a vulnerability and let the web dev fix it. Most decent web devs would rather have a consistent properly functioning target than a browser that handles other browsers broken code.
AMD x86_64 processors have an IOMMU. Intel's first x86_64 processors didn't but I don't know if this is still the case. IOMMUs are also important if you are running virtual machine software that allows some VMs access to physical hardware -- Xen lets you do this, for instance.
Pretend that something especially witty is here. Thanks.
I haven't seen a website require IE in years.
SVG Fonts is an optional part of the SVG 2.0 spec, FWIW. Frankly, I wouldn't call the other browsers' half-baked, supported-enough-to-pass-the-test-but-not-much-more support to be that much better than not supporting it at all.
IE 9 still can't pass Acid3.
So what? According to http://en.wikipedia.org/wiki/Acid3 IE9 gets a 95 and Firefox got 94. Besides the ACID test is about how well a browser handles the testing of esoteric, completely fucked up, marginally correct coding. It's also testing compliance for stuff that isn't rarely if ever used, and some stuff that's not even in the current standard (e.g. the CSS2 recommendations that were later removed in CSS2.1, reintroduced in the draft CSS3). It simply doesn't represent the real world.
In particular, have a look at http://en.wikipedia.org/wiki/Acid3#Criticism which summarizes the farce that is the Acid3 test.
Your "proper upload progress" would most likely involve Javascript or Flash, which not all people may have enabled or even installed on their computers.
The file input field bug is again one of the main reasons why lots of websites resort to using Flash or complex Javascript libraries to simulate an input field, because it's the only way to be sure it looks the same in all browsers (Chrome is a real problem here as their file input field looks totally different than the rest)
It's a pain in the ass to do workarounds and the ones hurt are the actual developers - one of the big reasons Firefox was started in the first place.
"Chrome still doesn't support basic features like saving tab state after a restart"
Factually incorrect.
"Cross-platform" means that Windows hates you, and doesn't want your damn chocolates. Or your fucking flowers. Bastard.
IE9 cheats on popular benchmarks (scroll to the bottom). And they still come second-to-last.
Flame away, but one area where MS is currently destroying the competition is on GPU acceleration. Mac is playing catch up, and unfortunately Linux is still a mess. There is a reason game companies still get away with releasing for Windows and ignoring Mac and Linux.
Fringe? It's still 60% of the browser usage:
http://www.netmarketshare.com/browser-market-share.aspx?qprid=0
Re-read OP's post. They are discussing IE9, which does not run on those things (unlike various other versions of IE). So... the 60% marketshare stat you provide is irrelevant to their premise.
In addition, read MSDN's post. It says...
(translated)"HEY!!!! We're FINALLY first with SOMETHING!!!! Let's rub it in everyone else's faces!!!!!!! Maybe they wont notice the fact that once again we wont be compliant with web standards!!!"
(in MS Marketing Speak) "We’re excited that other browsers have started to use hardware to accelerate graphics performance. With different implementations starting to become available, now’s a good time to blog about the difference between full and partial hardware acceleration."
StarTrekPhase2 - The Five Year Mission Continues!
Go to "Options."
See here for a handful of AMD boards which do support the IOMMU present in the 890FX chipset. In addition, the ASUS M4A89TD Pro/USB3 supports ECC as well, which is nice. Sadly, outside of the server chipsets, the others in the 800 series do not support the IOMMU.
You think that's bad? There was a critical bug in there for years that would completely overwrite your profile with a blank one, including your history, and bookmarks. Back when email was still integrated into the Mozilla browsers, your emails would get wiped too! The bug was caused by writing out the configuration files one line at a time, so that if the browser crashed during a configuration update, you'd be left with a partial configuration file. On the next startup, the browser would detect the error, and cheerfully overwrite your entire profile with the default profile to 'fix' it. The file contents were overwritten in-place, making disaster recovery practically impossible for most users. I won't even mention the performance hit of writing a 100KB file with 10,000 individual IO operations every time Firefox is closed, because compared to the data loss that's insignificant.
The Bugzilla forum had about 4 dupes of the bug, each with over a thousand panicked posts by users. Some of the reports when back years.
When it happened to me, it took me about an hour with Sysinternal's Filemon tools to figure out what was going on. The fix is trivial: simply write the new config file out-of-place, and then replace the original with it once it has been fully written. This is programming 101, standard practice for most Linux/Unix apps. Even Microsoft Office apps do this!
The bug went unfixed for at least 3 years after I first noticed it, despite at least a dozen posts by professional programmers who had even highlighted the source files and line numbers where the change needs to be made.
Bugzilla seems to be totally ignored by the Firefox programmers. I suspect that just like many open source programmers, they only care about the "shiny new stuff". Mundane work like fixing bugs is boring, so nobody does it unless forced to.
IOMMU is in the chipset and not in the CPU and it's been in chipsets since the very first AGP motherboards appeared on both AMD and Intel. However you don't even need IOMMU because DMA can only read/write into specifically mapped memory areas anyways, all you need to protect the OS from overwriting from a GPU or any other device for that matter is: don't map it into DMA I/O space ffs.
> I don't see Flash going anywhere for at least a decade
No one cares that Flash exists. What's important is that it be possible to develop tomorrow's web sites without having to use Flash, and that it be possible to browse the web at least somewhat reasonably without having Flash (e.g. not all sites need to work, but there should be sites in a given category that work without Flash). That's a somewhat realistic goal right now; for example very few banks require Flash (though some do).
> Silverlight won't have the install base of HTML5
The goal is to keep it that way, yes.
> Apple doesn't have enough influence to change the direction of the web.
You apparently haven't had to deal with the "if it's on a cell phone it must be Webkit" mindset of developers of "mobile" sites. See the part dealing with -webkit-text-size-adjust at http://blogs.msdn.com/b/iemobile/archive/2010/05/10/javascript-and-css-changes-in-ie-mobile-for-windows-phone-7.aspx which Microsoft was forced to take out later. Note that there have been calls for Gecko to similarly add support on mobile for some of the -webkit-* stuff Apple has been pushing people to use. Those calls have been resisted so far, but as for the future.... who knows.
Seeing that Chrome still doesn't support basic features like saving tab state after a restart
How is that different from:
Options->Basics->On Startup: Reopen the pages that were open last
Or are you making these claims without having actually used Chrome?
People should get over Acid3.
Some of the features Acid3 tests for are already obsolete (SVG fonts superseded by WOFF) while other crucial features are still buggy.
Actually it's much easier to find which AMD boards support it, as it's currently standard on all Opterons after a certain (relatively recent) generation, and supported by all chipsets which support them (at least AMD made ones, 3rd party may be different.) I believe it's not possible to disable, but it starts in all-accessable (ala no protection) for compatibility reasons.
Intel, has theoretically been out longer, but you have to hunt and hope the CPU supports it. Then you have to find a chipset which supports it. Then you have to hope that it's not disabled in the BIOS by the board maker. Note that it may be disabled later in the BIOS, as I believe one revision of an ASUS' board's firmware did. (Insofar as I know, that board wasn't advertised with the capability, it just had it early on, and a lot of people interested in it, jumped on board.) Also note that for Intel it's not limited to only Xeons, but it's risky if you don't know exactly what you want.