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."
IE 9 still can't pass Acid3.
http://CryoLANparty.com/ A lan I'm staff on!
What good is having GPU acceleration that only works on one platform? The -entire- point of the trend of doing things in-browser is to make cross-platform compatibility a reality. If I wanted a game to work just on Windows, why wouldn't I just make an application that did that?
Taxation is legalized theft, no more, no less.
...is that thanks to the lack of an IOMMU on consumer x86 computers, JavaScript exploits in the browser can now give you access to all the computer's memory, and along with it, ring 0. I can't wait to see the first whitepaper on the subject :)
So now IE 9 can make my GPU drivers crash. Instead of simply locking up and making me kill the process.
Free software web browser projects should reply by saying that they have better privacy, give away less personal / identifying information, help users avoid being mislead into clicking on ads, etc. etc.
I've never noticed whether my browser has fast, or slow, or any GPL acceleration.
Expert in software patents or patent law? Contribute to the ESP wiki!
I'm just glad there's some real competition in the browser market. Kind of nice when companies actually work on improving their products in order to get market share rather than resorting to becoming litigation machines.
Although the next question should be.. "so? pretty much everything else sucks AND you have managed to build yourselves A Bad Name in the Internet." , I'd say at least that much they can say about being better than others.
So yeah, even if other browsers (i.e. FF4) might come close, give credit where its due. They got it right. At least they got that ONE thing right ;)
Personally, and this is my opinion of course, even if they got the entire browser right I don't think i would use it having alternatives like FF and Chrome.
I just don't feel right "supporting" (if you could say "supporting) what still represents and carries the name of "Internet Explorer": ridiculously anticuated, slow, bad software which has brought countless hours of headaches to many web designers, programmers, users, etc.. and is/was generally seen as a necessary evil which you had to support because everyone used it. I find it just disgusting. IMHO
*cough*
The issue is video at the moment. Without accelerated video the browsers are not giving full capability to HTML5.
Stuff like http://www.thewildernessdowntown.com/ are fun, preview of what comes, but unfortunately very browser specific.
Waiting for a video standard....
---
Franck
http://www.avonsys.com
Franck Martin
Avonsys
I use Lynx. A thousand times zero is still zero.
I find it ridiculous how browsers battle over something like this when they can't fix very old and stupid bugs, and fully support some older standards such as CSS 1 and CSS 2.
For example, Firefox crashes when a user loads a 2-3 MB GIF file, because each frame is kept decoded in memory and the browser goes over the 2 GB memory barrier (for 32 bit applications). https://bugzilla.mozilla.org/show_bug.cgi?id=523950
Or, another example, the file input box ignores any css color rules simply because the html specs doesn't specify any rule so for several years nobody is able to decide something. It's actually since 2000 ffs: https://bugzilla.mozilla.org/show_bug.cgi?id=52500
Or, for several years now, when uploading a file using a form, the progress is stuck somewhere around 50% and it's discussed over and over but nobody can actually do even a temporary simple fix. Since 2004: https://bugzilla.mozilla.org/show_bug.cgi?id=249338
It's actually surprising they're able to code something as complex as gpu acceleration when they can't fix small bugs and at the same time it's unfortunate that basic things are forever and ever skipped in the hunt to get the latest "features" (sometimes just to check something on a feature list) instead of actually getting some things working properly.
That's misleading. IE9 gets something like 96/100 in the Acid3 test.
That's absolutely OK for most practical purposes.
Is IE really making the claim that they can incorrectly display your website faster than the competition?
Hmm...
wake me up when we don't have to waste an extra 20% fixing apps for IE
Their accelerator being better may or may not be true.. But it's still built into Internet Explorer.
The biggest pile of shit we've ever had to deal with. Over and over for every crappy version.
How many BILLIONS of hours have been wasted directly because ie is a pile of shit?
Will they license/open this technology? Will they have to charge?
They better warm it up, we all are waiting.
Is this the fabled Scotch mist?
Color me surprised!
An crash accelerator
What I'm reading is "Our browser is so fat we have to tap the GPU to make it appear fast." Frankly, bloat is nothing to be proud of.
...is that thanks to the lack of an IOMMU on consumer x86 computers, JavaScript exploits in the browser can now give you access to all the computer's memory, and along with it, ring 0.
You are claiming that the graphics drivers / host kernel give user processes the ability to read any location in the computers physical memory? Or is a separate vulnerability in the graphics driver required?
I don't care if Albert Einstein rises from the dead and announces on Colbert that he has proven that Internet Explorer's display technology is fastest that the laws of physics allow.
I still will not use any browser controlled by Microsoft.
... brings all the boys to the yard ...
Damn right! It's better than yours.
We could teach you,
but we'd have to charge
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
Great thing that they put an emphasis on the combined performance of all browser subsystems and picked the next one, seriously focussing on Graphics perfomance. That's a good new direction after all the hype over Sunspider and JS benchmarks.
Do people even give a crap about how one browser renders a page 0.0001 seconds faster than another? I mean, we still have a huge honking bottleneck here (internet speeds), so it's not like render speeds is that much of a factor at all when viewing a page.
> Will they license/open this technology?
Which technology? Their methods of ignoring established design principles in favor of quick & dirty programming? Patented.
Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
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.
OK, so the display portion that takes milliseconds at most now takes 4-5 milliseconds less time. Meanwhile the browser's taking 10-30 seconds choking on bloated Flash, over-large images and hundreds of K of insanely-convoluted nested Javascript files. Somehow I don't think graphics acceleration will help speed up Web sites significantly.
It's kind of like cars: sure the McLaren F1 may be faster than my Ford Focus, but it's not the car that's setting the 75mph speed limit.
Well that'll teach me for not reading the minutiae of hardware revisions :P
--sitharus
I haven't seen a website require IE in years.
I've never gone to a website and cared about how fast it rendered. What I do care about is how secure I am and if the browser is able to deal all the pop ups, pop unders and other junk.
The IE dev team are just lacking any other decent USP to sell the merits of IE over other browsers. Firefox hasn't really made all that many big improvements for some time. So there's not much for IE to copy.
http://gs.statcounter.com/#browser-ww-daily-20100911-20100912-bar
The real war was against IE6 and ActiveX and that has been won.
Microsoft is no longer holding back web development, the new problem is going to be XP hold-outs running IE8 and 7.
You're looking for This one. W7 and Vista have together less than 30%, and that's the only operating systems IE9 will run on. So if they get 100% of those, which seems unlikely, their max upside today is 30% of the total browser market. Since as you note they only get 60% share even though Windows is over 90%, it's a 20% upside potential for IE9 today - probably less since early adopters are also the people most likely to choose a different browser. Fringe. Not enough to dominate the developers.
XP has a very long tail. It's still selling in the market and will be installed through downgrade rights for the entire life of W7. XP will likely still be over 50% three years from now. IE9 doesn't run on XP.
Help stamp out iliturcy.
What for? Object tags have been around for years, and can embed ANY type of content in a webpage. I was doing video streaming (along with custom play/pause buttons etc.) with them and the older embed tag back in 2000. I believe object has been around since '96, and embed longer. AND, object can support any multimedia content, including video, audio, flash, panoramas, applets, etc. I'm not sure how advanced/generic DOM scripting is for them, but that should be relatively easy to define as well as anything for the video tag can be defined.
The only thing special that I know of about the video tag is a standardised video format, but mpeg4 and now x264 are pretty much de facto anyway. WebM is nicer, if it supports the multiple angles and subtitles and everything that matroska supports. Apart from that, the open standard is nice, but I don't see much to get excited about.
Great, we now have GPU-accelerated HTML in IE9. Note to the IE9 team: howsabout implementing the one thing MOST LIKELY to need that acceleration - the canvas tag. Nobody is going to switch to VML or whatever proprietary garbage you're pushing.
There wouldn't be much to license; they'll simply be pushing high-level graphics calls down the API/driver stack to the graphics layer. The open equivalent would be for firefox/webkit to have high-level graphics API calls added to the X rendering libraries (cairo or whatever) and call those directly when running on systems that have the necessary libraries. The X-window graphics stack would then do its part, by providing high-level graphics primitives and high-level API functions implemented with fast, low-level code that's tailored to your hardware. The most obvious candidate for achieving that is Gallium3D.
And their rendering engine is still worse than everybody else's. So IE9 can render broken pages faster than any other browser. Whoo!
Audioscrobbler
I think they're saying "My Javascript brings the boys to the yard."
Help stamp out iliturcy.
I don't care about GPU acceleration. I just want it to support CSS like every other browser. Is that so fucking hard? It must be harder for a web browser to support web standards than it is to harness the power of a GPU.
WebM is nicer, if it supports the multiple angles and subtitles and everything that matroska supports.
It doesn't. The WebM container supports almost nothing Matroska does. No chapters, no subtitles, not even tags, no multiple audio tracks, no multiple angles, no segment linking,...
What WebM supports is one VP8 video track, with one Vorbis audio track and that's it.
Does their GPU Acceleration bring all the boys to the yard?
http://en.wikipedia.org/wiki/First_Amendment_to_the_United_States_Constitution
Hahha, I see. Thanks for that :)
Does it matter? for Microsuck users, your games play faster while you are pwned.
Can someone please explain why Mozilla don't simply use OpenGL on all platforms?
Are Microsod planning to kill WebGL or something?
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.
...and it might actually matter when you can actually find a motherboard with a chipset that also supports the IOMMU on the CPU. At the moment, that means an X58 chipset (socket 1366) for Intel, and for AMD, you're pretty much out of luck.
Game! - Where the stick is mightier than the sword!
IE9 cheats on popular benchmarks (scroll to the bottom). And they still come second-to-last.
You're ignoring the fact that a large percentage of websites are no longer pages, but instead are really javascript driven applications running in a browser. For large web applications, this type of performance improvement is far more important than the bottleneck of downloading the app.
but I'm testing ff4b5 and it is really glitchy. I can't even use the gpu accel in it because that's so glitchy.
Most of the time, ff4b5 doesn't render pages properly or scripts fail to run properly in it. Way too glitchy for a 5th beta!
They're using their grammar skills there.
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.
When viewing scaled video, it's a huge factor. And when using web applications (as opposed to reading the news) it's a significant factor. Oh, and when scrolling, not like anyone ever does that with webpages.
And we're not talking tenths of a millisecond here. If each scroll operation takes you 200ms (easy to run into without hardware acceleration on some sites out there that are sticking video or large translucent images over fixed-position backgrounds), you just lose.
Yeah, well, if everybody else could use Microsoft's secret undocumented API calls, their GPU acceleration would be even faster. And safer. And not have daily security patches. And crash less. And have smaller code bases. And not suck.
/satire
MS is big into bragging lately. Braggadocio may be their #1 product now.
http://lkml.org/lkml/2005/8/20/95
FF4b5's GPU acceleration makes all of the text fuzzy and hard on the eyes, in Win7 on an nvidia card. Is there a solution to this?
I wish I could introduce you to the hell that is HP's partner portal, their Learning Center, the portal for support. It's a carnival of the obscene. As someone who understands web design I have to hope there's a special level of hell devoted to eternally tormenting these web developers.
Not only do these sites require specific versions of IE, but then you come to a certain point where they don't even work with those, so you have to migrate the session to other browsers through trial and error until you find the one that works with it. It's sick. It's like an online skill test that requires four nines of web proficiency in order to download a freaking driver update or read the product alerts.
In a perfect world some auditor would have these web developers separated from their skin slowly, under a saltwater and lemon juice shower while rats ate their organs, with a blaring Phil Collins soundtrack.
Help stamp out iliturcy.
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.
Firefox is still the best (if IE9 is better, it isnt out yet)
I'm just going to sit here and giggle and point.
Opera works fine for reading HuffPo.
That is at worst a lie and at best misleading. No consumer processors have a IOMMU that you can use for this. I have had much interest in the subject (mutiple virtual OSes accessing the GPU at the same time through Xen would be awesome). The problem is that it just isn't there yet.
GPU acceleration isn't that important for most browser rendering. The major benefit is found in image compositing and HTML canvas, both of which can be accelerated without major architectural upheaval for the legacy bits of the browser. Not rapidly leveraging hardware for every conceivable operation isn't going to doom the portable browsers.
It takes more effort to optimize portable software. The correct abstractions must be found, by design at first, and then through rework as reality is discovered. The slight performance deficit the portable browsers suffer will eventually vanish, and long before it matters to anyone. Other factors will continue to dominate browser preference.
Microsoft has lost the search war, the marketing war, the content distribution war, the standards war (HTML, PDF, etc) and the mobile war. Silverlight remains irrelevant and IE gives up more eyeballs every day. Accelerating IE9, a browser which will probably land with as big a thud as will Windows Phone 7, isn't going to fix all of this. Microsoft is a legacy business with legacy products, howling for the attention it no longer merits.
Lurking at the bottom of the gravity well, getting old
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.
So... I've been wondering how much of an attack vector we are creating by sending content more directly into the GPU libraries and the security limitations of such libraries which are so geared for performance they must be skimping on security? are GPUs limited in their DMA access like say... firewire/1394 was? (eventually it was)
With VMs allowing access to hardware for speed boosts one begins to wonder how secure the VMs are and as they get complex... just how much of a VM they are --> leading more towards a lower layer kernel system... But then the lack of solid OS + good design + standards is why we are stuck with another layer of abstraction which openly ADMITS that current OS are too immature.
Democracy Now! - uncensored, anti-establishment news
I can piss further than you can!
(Just don't look because I'm standing my mountain of money and my minuscule limp dick is actually normal - my marketing dept says it meets or exceeds the standard one!)
Many (all?) modern GPUs have virtual addressing with page protection. Switching is handled by kernel mode driver (we are talking Vista+) and is not controllable from user space. Now we are talking KMD attack which is not different in principle from any Ring 0 driver code attack. Good luck!
IE9 has zero (0) users. That means it is fucking irrelevant.
Don't get me wrong ... it's great that they're cloning WebKit. It's great that they will finally have a real Web browser.
But ship something, Microsoft! Sheesh.
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.
And why do I care? Nobody has cared about MS in 5 years and coming out with a web browser that is finally half-decent is not going to change anything at that POS company.
You mean Windows Vista and Windows Vista 2.0?
I hold very few opinions. I hold information based on observation and fact. If you wish to disagree, please use facts.
What is this win-dose? You keep talking about.
- To understand recursion, we must first understand recursion -
Yet ANOTHER reason to kill off the x86 legacy crap for good. The Burroughs B5000, S/360, AS/400, MULTICS, and LISP machines all had built-in hardware support for this...back in the 1960s/1970s. The x86 doesn't even have anything close, not even their VT Technology(TM).
When Intel introduces Some New Cool Marketing Name(TM) Technology, it's either crap (like Pentium 4/NetBurst), or copied from someone else (x86-64 from AMD, FMA from PowerPC and practically every other RISC architecture with a floating point unit, CMOV from Motorola 68k, AES instructions from VIA PadLock). Oh, wait...Intel x86 doesn't even HAVE half of those things yet, they're just FUD proposals to make customers think Intel is delivering and to go along with them instead of AMD (see also: the Shitanium fiasco and the FMA3/FMA4 debacle). And their new instruction mnemonics like "PCLMULLQLQDQ"? "PHMINPOSUW"? "VFMADD132SD"??? WTF? Who came up with these names? Does Intel think that making 50 similar special-purpose instructions with nonsensical 12-character mnemonics that differ by a single letter is a good way to get people to use them?
Another thing: it's an insult to the CISC name. When people think CISC architectures are ugly, it's because of the x86. The ugly Intel architecture "implied operands" are a legacy of it's Z80-style accumulator architecture. SAHF? LAHF? JCXZ? And EAX, EBX, ECX, EDX are the only FOUR so-called "general-purpose" data registers (ESI, EDI, ESP, and EBP are all address registers for pointers). The x86 is an overgrown Z80, a UISC (Ugly Instruction Set Computing). There's no reason why CISC should be conflated with ugly encodings or two-operand arithmetic or terrible addressing schemes (x86 "segments" are junk compared to real virtual memory segments), or pointless restrictions on operands (why do shifts need the shift count in CL? why do divides/multiplies use EAX?.
Now maybe I'm a little too old-school, and a little too logical, but it seems to me like the functionality they're boasting should be implemented at the GUI level, not in the individual apps. Flash video is an unfortunate hack, that came about because no one did what should have been done: provide a cross-platform way for a browser to tap into the GUI's existing video abilities. Windows has DXVA and the legacy VFW architecture, Linux has XvMC and VDPAU, Mac has Core Video.
SVG ? Vector primitives. Windows provides that in GDI+, Linux has Cairo, and Mac has Quartz 2D.
Or, if we wanted to be pompous elitist bastards, we could say that all of this crap could have been handled by OpenGL and OpenCL, if only those big dumb wealthy corporations would get off their pedestals and recognize the value of open standards. But we're not going to do that, now, are we ?
-Billco, Fnarg.com
My laptop cost about £400 about two years ago. It is *not* top-of-the-range. I favoured hard disk space, a large screen, and lots of ports over any other factor. Still, it's dual-core. It runs XP, so there is no GPU acceleration of the desktop etc.
I load opera. I open my saved sessions that contains 26 of the websites that I check every day. Before I can blink, 26 tabs open and my download light starts going crazy. By the time the first link has loaded, they've *all* loaded. I can't read faster than my computer can display them and lots of them are quite heavy and they are all loading simultaneously over bog-standard 54Mbps wireless with a bog-standard home broadband connection. I even have all caching disabled (except for a tiny RAM cache in Opera) because, basically, it's a waste of my precious disk space and RAM because I can't notice the difference between that and just loading each time. Firefox has a MUCH longer loading time but performs pretty much the same. IE doesn't seem much different but I have an ancient version and probably wouldn't be allowed to install IE9 on XP anyway.
Flash games work. iPlayer works. I can have dozens of windows open and loading simultaneously and the only bottleneck is the Internet connection. However, even in work with a squid cache that prioritises caching all those sites (I'm the IT manager) and two 24Mbit lines, I still can't click on the first tab before it's loaded its content (and the first tab is all that matters, because the rest have to wait until I've read the first one).
GPU acceleration is worthless here. I'm really *not* going to notice and the future is smaller and smaller devices that load at these same speeds. My phone's HTML rendering is lousy-slow but then the connection is lousy-slow and my phone is ancient. Anything half-modern renders the first text in less than a second and can scroll smoothly. That's all that counts.
In an age where average people think it's acceptable for a computer to take 5-10 minutes to boot, to take 20-30 seconds to show something actually loading after they double-click, where machines are bogged down with AV and anti-spyware and firewall and toolbars and everything else, GPU acceleration won't do shit. If people go for it, it'll be because it was enforced, or because of the "HDTV/Bluray/3D" gotta-have-it phenomenon.
Firefox's handling of gifs is awful in general.
Whilst the gif spec may not specify a minimum wait time between frames, that doesn't mean you shouldn't implement one. The vast majority viewers implement a minimum pause because it has the handy feature of not having a gif that uses up 10-20x the CPU it actually needs to and doesn't render frames at 5-10x what the monitor or program could actually display.
Also: numerical font-weights. I'm using numerical values for a reason, don't just approximate it to one of 4 thicknesses (although Firefox isn't the only browser guilty of this).
Mod Up,
Actually 'Tridge' said this 10 years ago, about the same time Sony was smart talked, then TVIO's were hacked. Pop the IOMMU into special diagnostic mode, and VM's can have their own tee'd virtual scratchpad. Ring 0 is good, but to set diagnostics on is another level. These chip companies are still holding back, but CERT has seen a few.
"We're excited that other browsers are starting to optimize for the Windows platform .. Today, IE9 is the first and only browser to deliver full hardware acceleration of all HTML5 content"
"Microsoft marketing is making noises about IE9 having a monopoly on "full hardware acceleration". They're wrong; Firefox 4 has all the three levels of acceleration they describe. It's surprising they got this wrong, since Paul Rouget published a great overview on hacks.mozilla.org a few days ago (and our plans, source code, and nightly builds have been public since we started working on this stuff many months ago) link
"Firefox's hardware acceleration interacts with a machine's graphics hardware via DirectX or OpenGL, depending on platform" link
Does the testdrive site serve up different code depending on the browser?
Yes, because in previous versions the graphics card was completely unused. You realize every PCI Express card has an IOMMU, right? I realize you probably wanted the type that allows hardware virtualization on the CPU bus side, but neither suddenly makes exploits impossible, and lack of one does not magically give exploits free reign.
Think before you post. Browsers now use more features of the GPU, but in both cases there was memory mapping going on and data being sent to the GPU. Now they use other APIs, and neither is magically more or less secure. Both are only as secure as the underlying API. None of the implementations are likely manipulating the memory mapping directly but are using the driver stack and system API, which is what how things were rendered before.
If it was so easy to read or write arbitrary memory using *ANY* common PCI or PCI-Express card, don't you think we'd be doing it to get system level exploits from user mode already? It has and is constantly being tried, with much less hindrance than a browser in between. Please find me *one* published exploit using any common PCI or PCI-E card to read/write arbitrary memory. It doesn't happen on Windows, on Mac, or in Linux. It doesn't happen at the app level. Lack of IOMMU and hardware acceleration is not some magic combination that has never been pushed for exploits before, and will suddenly, by adding a browser, reach critical mass and doom us all.
Are you claiming that Chrome and Mozilla are inherently insecure as well since they turned on the magic acceleration bit? That all OpenGL and DirectX apps have this major security vulnerability just because they use accelerated hardware? How about accelerated audio processing? Oh no - where does it end?
By the way, I am a professional security researcher, and am quite aware of IOMMU attacks and issues, SMM mode attacks, and the possibility of such attacks. But your spreading of FUD and the resulting 5 Interesting really shows the lack of knowledge about these things slashtards on average exhibit.
In short - there is no reason this should be more or less secure. One could argue the old GDI methods are more insecure since they are crufty and not designed with modern secure coding practices in mind. One could argue the perhaps the GPU cannot access random memory like some magic genie without the driver telling it to, and perhaps the driver interface is well checked since it is smaller than GDI.
Some people try any tactic to attack their favorite boogeyman.
Correct me if I'm wrong, but isn't DirectX also an "abstraction layers [that] inevitably make[s] tradeoffs which ultimately impact performance" for various brands of GPUs ?
What a depressingly stupid machine.
You can put lipstick on a pig, but it's still a pig. I'm not talking about IE, I'm talking about Javascript. Browser vendors need to stop obsessing about the pretty pig competition, and get to work implementing modern web standards.
Basically MS have lost market share, mindshare and respect for their browser and
now when they see the vista of browsers competing effectively against them, they've
sat down and said 'gosh darn it, how can we leverage our Windows(tm) advantage to recover
and totally dominate the Web once more?' What do the other guys not have that they just
can't emulate on our platform ?
And they all agreed that 'leveraging' DirextX would be the best.
So they've made up this pointless metric of displaying shit faster on the screen, like the web's
a video game or something and said - 'see, this is really important and we're the best at it! Now buy
Windows 7 and download IE9 so we can totally crush the competition already!'
I'm pretty sure they've also instructed every salesman and evangelist to trumpet IE9's hardware acceleration
to try and crowd out the competition.
MS really can't brook any competition in their playpen.
don't be a spelling loser
I feel obligated to nitpick some of your points:
1) CMOV was not copied from the Motorola 68k, which doesn't have such an instruction - it has a conditional set instruction, which is not as general-purpose as CMOV.
2) The x86's ugly architecture is a legacy of Intel's desire to be source-code compatible with the Intel 8080, not the Zilog Z80, which is itself a derivative of the Intel 8080.
3) While there is no theoretical reason why a CISC architecture should be limited to two-operand arithmetic, I'm not aware of any CISC architecture with three or more operand arithmetic.
I do like your 'UISC' acronym.
In the future no one's computer will be able to render a web page, but they will all boast about how great it performs!
Honestly, it takes a $500 graphics card and an OS 35% of PC gamers use in order to access DirectX11, plus a desire to not use PhysX. I wouldn't think "we made the internet require real hardware!" would be something to boast about. I bet the hardware vendors are happy, though; they might finally convince people to stop buying netbooks that are completely sufficient for their needs and go back to selling hardware that's major overkill with excellent margins.
"The typical state for most contracts is some wishy washy thoughts about what would be nice that then turn out to have been a hallucination one of the managers had the previous night after too much LSD" - by zuperduperman (1206922) on Sunday September 12, @07:02PM (#33556372)
No doubt as of the results of too many "Acid3 tests"