Slashdot Mirror
Linus on Intel's 64 bit Extensions
ceswiedler writes "KernelTrap is running a thread on the Linux-Kernel mailing list about Intel's new IA-32e 64-bit chip. Linus complains 'what I found so irritating is that _hours_ after the Intel announcement,
people were _still_ confused about whether the new intel chip was actually compatible with AMD's chips.' It is, of course, but you have to do a thorough comparison of Intel's reference manuals to discover that-- they don't mention the fact that their new chip is instruction-set compatible with AMD's x86-64 chip." See the previous story for background. So it looks like the reason Intel was vague about their announcement is that they didn't want the WORLD TO KNOW THAT THEY WERE COPYING AND FOLLOWING AMD rather than developing some new thing on their own. Slashdot is proud to help Intel in this quest; wouldn't want the public to know that INTEL WAS SIMPLY FOLLOWING IN AMD'S FOOTSTEPS. Hope this helps.
Mod me as you will, and I do realize it was meant to be funny, but I expect a bit more from /. than portions of news postings in all caps. Leave it to the reader to decide what's important and what's not. All caps automatically annoy me, and have done since I can remember.
Thanks for your time.
In any business, when you are getting your arse kicked, you look at your competitors to see what they are doing. Why reinvent the wheel and all that....
[ Don't reply to this ]
Intel has to be very careful right now, for years they have been seen as the innovator in processors. Now AMD got the jump on them and they don't dare not respond, but they have to respond in a way that seems like it was thier choic.
-- Slashdot, making the Left look conservative since 1997.
Why would Intel be embarassed or whatever to "follow in AMD's footsteps"? I mean, sure Intel's bigger and badder than AMD, but can't you learn something from the little guy sometimes? Don't things like this happen all the time in the car industry with various technologies?
Isn't it interesting how you come to recognize posters based solely on their sigs???
Now Intel is making AMD compatible chips ?!!
but micheal beat me to the punch. I'm not sure whether Torvalds was complaining about Intel not coming out with a ready admission "We had to follow AMD because they got there first" or complaining about programmers missing the (in hindsight, at least) obvious conclusion that Intel would make a Howitzer-size hole in their clean-room booties by not going with the AMD flow.
Good for Intel. You may think that the important thing is that they are "following in AMDs footsteps," but I think the important thing is that people won't have to write for two architectures now.
Good thing AMD never copied the Intel instruction set. Or is it a good thing when AMD trys for market wide compatiblity but a bad thing when Intel does the same?
"Have you ever thought about just turning off the TV, sitting down with your kids, and hitting them?"
biased article summary much?
..but are still accurate.
Intel will never, ever, ever put anything out that their name isn't all over it.
Example: Firewire. An industry standard. Does Intel put it in their motherboard chipsets? I remember old Intel comments stating their 'commitment to IEEE-1394' but it was all a load of crap. The PII and PIII chipsets could (and should) have had it on board.
Here, finally, Intel has decided to take someone else's tech. But even now, they won't admit it's someone else's tech.
What a bunch of arrogance..
It's the big dogs that copy the little dogs. Both in this case (Intel, AMD) and in the Microsoft, Apple case. It is so satisfying when this happens.
Now, go troll somewhere else.
-- Fighting mediocrity one bad post at a time.
This is how technology goes mainstream and becomes inexpensive enough for the everyday consumer: following.
Being a trailblazer may get you bragging rights, but you risk fragmenting the industry and the market you feed. For the longest time in the 90's AMD and Cyrix went on a follow-quest, and breached the low-cost PC market. Not only did they enhance choice and lower prices, they kept the number of standards down to a minimum. Just imagine what would occur if AMD, in the 1990's, came up with something completely different, but can run exactly the same thing Intel chips can at the same price: the market gets fragmented, prices remain high and stagnant, and no one is the winner until one of the two gets clobbered, eliminating competition in the market and raising prices even further.
It's not characteristic for Intel to follow AMD, but IMO, it's the smart thing to do to be competitive.
Oh yes, just because they are following, doesn't mean they can't do it better. AMD did in the 90's and today.
------
Amadaeus
The last bastion of Mathie-ism
So we are ridiculing Intel for supporting common architecture and collaborating with competitor? I guess the implication is that the proper thing Intel should have done is develop its own set of 64-bit extensions, making it absolutely incompatible with AMD's offering. The world would be a much better place then, right?
And Intel doesn't really have to advertise the fact that it's AMD-compatible, it's not like AMD owns more than 80% of the market, and Intel is below 20%. To hyperbolize, you don't expect Microsoft to announce the next version of Office to be compatible with Joe's Software Shop's software.
AMD needs to wait until Intel is completely involved in x86 32-64 and then launch a complete advertising compaign to the General Public about this. Show benchmarks. Help manufacturers freindly to their product push their wares. And really nail them hard. No Blue Man Group silliness. No stupid ads. Just plain facts and examples. But they must do so in mainstream media. Telling a bunch of geeks about their products doesn't work, they already get all the information themselves. And they buy based on reasearch anyways, not on advertising.
I don't like the fact that the Slashdot headline takes Intel to task for doing the right thing. It would have been bad for everyone (Intel, AMD, and all of their developers and users) for Intel to adopt a "not-invented-here" approach and conjure up yet another 64-bit instruction set out of thin air. The fact that they didn't is a good thing.
Crappy journalism on Slashdot's part.
The whole 64-bit thing is hype. Until these processors extend their native word size from 32-bits to 64-bits, they're really not useful for most people.
64-bit memory extensions... whooptie doo.
Next time they decide to take a bullet and promote compatibility, they'll know better!
TEACH EM A LESSON, SLASHDOT!
REM Old programmers don't die. They just GOSUB without RETURN.
I thought Tom's Hardware was a pro-Intel shop? Their reviews and commentaries over the past few years have read as such.
Posting anonymous due to affiliations: Does AMD "give credit" to Intel for the original 32-bit x86 instruction set? I would hope not, this isn't a Grammy acceptance speech. It's strictly business, nothing to see here, move on.
Why? Intel could just come and take the high road, and claim that they are keeping their chips instruction-set compatible with AMD64 to "preserve the customers' investment" or some such marketing-speak. Good marketers never let the facts or the truth get in the way of a good spin.
The caps were illustrative and fit with the objective. There was nothing wrong with this poster's use in this instance. Get over it.
Intel aren't copying AMD, they've just stated that their 64-bit instructions are compatible with AMDs, how it's implemented is anyone's guess.
It makes sense really. Heck, if they were different, we'd be up in arms about that too... *cough*Itanium*cough*
While I found the submission entertaining, I still have to give props to Intel. OK, yes, they are making a chip that's compatible with AMD's instruction set, but this can only be a good thing. Instead of running out and introducing a new 64 bit instruction set to the market to directly compete with AMD, and thus create market confusion and compatibility problems, they've decided to do the best thing for us, the consumers and programmers - embrace an existing standard to avoid market fragmentation.
Yes, it might have pissed Linus off that they weren't very forthcoming about it, but just think how ticked he'd be if they introduced something completely different and he had support two competing 64-bit architectures.
Maybe Intel is taking a lesson from IBM. Just because you are the big boy on the block doesn't mean you can make your own rules. Anyone remember Microchannel Architecture?
-R
I really don't understand why Intel is handling this so poorly. Someone in the higher ups must have thought this to be a horrible end of the world type of thing. In my mind it isn't. They have ended up making Intel look worse if they would have just quitely said, we are supporting x86-64...which is compatible wiht AMD, at the end of the sentence.
AMD is an x86 processor. Something Intel invented. Becuase of the agreement between Intel and AMD over the use x86, Intel can use the new extensions without paying royalties.
A) The only people that might loose faith in Intel are some techies, most of who are already AMDFanBoys ( and girls ) anyways. The average consumer ain't going to care who created the 64 bit extensions.
B) AMD DID THE WORK. No need waste time designing the specs out.
C) MS has an OS ready to go out the door, no time waiting for you apps to be deployed.
D) AMD has spent a lot of time marketing the technology, all you have to say is we do it with more GHZ ( please don't let the GHZ thing spin off into another thread ).
E) You've got something to help ease the pain between your Xeon and Itanium lines.
This is a good thing for Intel. Sure you are copying AMD's instruction set, but lets face it, compared to the man hours needed to actually implement the instruction set in trannies, an instruction set is pretty simple. Intel saves money, says hey look we are not a monopoly anymore don't hate us, and has a good product.
Intel made a bad PR decision, they should just admit it and move on.
Actually I'm glad that Intel has done at least as much as making it compatible. Shouldn't we at least be glad for this? I mean if they put a bit of marketing spin on what they have done, fine it will go away.
Imagine the mess though, if they decided, "ok we're going to make our instruction set just a little differnt and then use our dominance in the market to win over AMD." It would mean more work for hardware designers (I know PCI bus should take care, but you still need to test), kernel developers, window's driver's writer's, distributors and you and me, because we'll have even a harder time shopping for hardware.
I'm pretty certain that MBAs have been considering the above option. This is a compromise and people have to learn live in a world that is not ideal and thus full of compromises.
Mod me as flamebait I'm sure but those of you who think Slashdot is a news site need to get some wires uncrossed. It's no different than any other blog out there.
Why is it that we're always seeing people expect quality editing and other journalistic qualities like lack of bias? Even worse who are these idiots who mod such comments up?
It's just a bunch of guys posting links to other news sites! This is a community FOR NERDS not a replacement for your newspaper.
Onto the actual news this is great. You could say it's the final proof that intel can no longer really dictate things completely in the x86 market. The trend was fairly obvious for a while but this has a nice way of finalizing that fact.
then how come there are production versions of Windows xp-64 itanic and win 2003-64 itanic, when there is still only a developer release of Win 2003 amd64?
Wintel, anyone?
Let's not forget that Intel did release a 64-bit processor before AMD. Who can blame them for trying to break away from an instruction set that we've been using since cavemen were carving computers out of stone. It's not their fault that everybody wanted to stick with x86. So now to appease the general public they are forced to beat that dead horse one more time. In fact Intel should be lauded for being compatible and not trying to monopolize the 64-bit market.
One thing to like about Linus is his anti-FUD, full discosure style, evident here just IMHO.
C|N>K
Whats the story with Linus these days, he is getting fiesty about all sorts of things... this is very unlike him. What happened to the mild mannered Finn of formers years? He must be still peeved about that SCO bollocks. While its all very amusing to see him lifting Intel out of it, I sort of miss the zen like dryness/sarcasm of yesteryear.
"Thats right buddy, the large print giveth, and the small print taketh away."
Imagine, Intel conforming to an existing standard!!!!WOOHOO! Way to go Intel!
If we could get more co-operation like this in the industry, software and hardware would work much better, and crash less often!
ttyl
Farrell
CAN-CON 2019 - Ottawa's only book oriented Science Fiction Convention! October 18-20, Sheraton Hotel, Ottawa, Canada h
It's not about credit, it's about "Developers! Developers! Devlopers!" The general public doesn't give a toss, and few people are going to be snickering behind closed doors over this sensible business decision. But anyone writing low level software - such as Linus - needs to know staight up if it's compatible or not. Not mentioning x86-64 means they screwed around the people that make their chips worth anything.
Forget thrust, drag, lift and weight. Airplanes fly because of money.
Just a passing thought, but perhaps the reason Intel wasn't forthcoming about AMD x86-64 compatibility was because they didn't want people running out and buying Athlon64's, knowing that any compatibility issues between Intel/AMD are now moot.
Who doesn't like free music?
Well, in this case it was good. I'm sure if it weren't for the 800-lb gorilla Microsoft refusing to support more than one 64-bit X86 architecture, Intel would have annoyingly forked yet another extension incompatible with AMD's.
This would have significantly raised the costs of software packaging for everybody for years to come. In fact, the extra hassle would probably make for a significant decrease in the number of programs that even bothered to release 64-bit versions at all.
People like to think that AMD is a bunch of guys working out of someone's basement. In reality, AMD is a hulking monster of a corporation. This is a company with tens of thousands of employees and 2.7 billion (US) in revenue in 2002. So, yay, one monstrous corpororation is better than another!
Too bad Intel doesn't have the self confidence and class to cheer for the competition, and *then* turn around and get back to the business of leaving them in the dust.
What you say is true now but AMD could do some real damage though by launching a series of commercials with some catch phrase (think "Intel inside") that plays up the fact that Intel chips are based on a standard developed by AMD. Something to effect of "Why pay more for a processor based on AMD's standard when you could be paying a lot less and using the real McCoy?"
Not exactly phrased that way, but you get the idea. It could be a real plus for AMD if they could find the right way to market it. Hence Intel doing the smart thing and trying to burry it for as long as they can.
This is just one small symptom of many. Intel is having extremely serious management problems now. Intel hasn't been very humanistic in the past 15 years, and now the company is failing in many hidden and not-so-hidden ways.
Which is more misleading, engineering chips with an impressive clock speed but a mediocre instruction-per-cycle ratio, or to engineer chips with an impressive instruction-per-cycle ratio and market them as comparable to the industry leader's equivalent processors?
I'm not crazy about AMD's processor ratings, but I understand the necessity. Joe Blow doesn't know squat about processors, all he knows is that higher clock numbers are (supposedly) better. At least AMD keeps their marketing department out of the engineering meetings.
Whats the big deal with this? I quickly went through some AMD's Athlon PDF's, and while all of them mentioned SIMD/SSE and MMX support, I didn't spot any mentioning that those are actually technologies originally developed by Intel. And why is this? Maybe because Intel and AMD have the cross-whatever license on their technologies so they can leave all that out. Come on, who would voluntarily put their competitors name in their product sheets?
After all, its not either of them copying anything from each other, but just making their own product compatible with a certain set on instructions, while still using their very own under-the-hood implementation.
I don't think anyone is surprised by this -- Intel would be nuts to mention AMD in any press release about anything unless it's incredibly negative toward AMD (which this definitely is not), and even then it would be ill-advised from a amrketing perspective.
AMD could send out the press release:
AMD Develops Intel Instruction Set
Sunnyvale, CA -- February 24, 2004 --Intel Corperation announce last week it would be using an instruction set pioneered by long time rival AMD (NYSE:AMD) in it's new 64-bit processor....
What's bad, though understandable, is that they don't mention at all that this is the case.
TCP: Why the Internet is full of SYN.
Who says it was HTML? Did you see a DTD statement?
Was that really necessary?
We got the point from Linus the first time.
He was posting his opinion in a mailing list. You're posting juvenile caps on the front page of a very highly-visited, corporate-owned tech news site. It just lowers the image of Slashdot all the more, and no matter how many times Taco professes that Slashdot is just a "hobby," it is viewed as the pinnacle of Linux community opinion and tech news by everyone else.
I know you guys already hit the bottom with the "Microsoft Violates Human Rights in China" article, but at least show a little maturity in the process of editorializing something. Despite typos and endless dupes, at least Taco only writes one- or two-line remarks. He would have said something like, "Most people know AMD was first, so it seems silly for Intel to behave that way."
I used to wave the mighty intel flag and argue ferociously with anybody supporting AMD....until I actually tried an Athlon, then my song changed to fuck intel. What I'm trying to say is nobody really cares what the brand name is, its the performance that counts. An Athlon 800Mhz out-pacing an Intel 1.5Ghz is what won me over.
Wishing I was a millionaire since 1969.
It's not so much that AMD is "still following in Intel's footsteps", it's that AMD chose to remain x86-compatible. If that's following in Intel's footsteps, then Intel is following in Intel's footsteps too, I guess, because Intel sells lots of x86-compatible chips.
The continuation of adding on register extensions is great for backwards compatiblity but it makes the instruction set a mess.
But -- who cares? Modern CPU chips translate instructions into RISC-like micro-ops, and feed the micro-ops into multiple execution units. AMD chips can do a whole bunch of stuff in a single clock cycle, which is why they are much faster per clock cycle than an Intel chip. The pain of a wacky instruction set is isolated in the translation part of the chip, and doesn't significantly hold back the chip in other ways.
RISC fans predicted years ago that CISC would die, because RISC is so much better. But CISC chips contain RISC cores these days, and meanwhile architectures that were originally "RISC" have all kinds of special instructions for working with video data and such (doesn't seem so "reduced" to me). What really happened is that RISC and CISC kind of met in the middle.
And the old idea that RISC instructions would win because they are easier to decode didn't pan out. CISC instructions get decoded to RISC-like micro-ops, as I said, and it turns out not to be a huge deal. Meanwhile, those CISC instructions are denser than RISC instructions, so you fit more of them into your limited cache space, which helps speed.
In short, modern chips do all kinds of clever stuff, and the instruction set architecture is not really holding them back.
The sad thing is that a new cpu could have a compatibility layer that had a slight performance hit but with a lack of software supporting new 64 implementations people wouldn't buy it because the pretty little bar graphs that the sales drones produce.
If you want me to feel sad, you need to back this up with some facts. Show me why you feel the Athlon64 would be faster if it were not backward-compatible with x86.
As it is, the Athlon64 is already a sweet chip in 32-bit x86 mode (you know, "following in Intel's footsteps"). Then it gets better when you run 64-bit software (mainly due to the extra registers). Good in 32-bit, better in 64-bit... why am I supposed to be sad again?
steveha
lf(1): it's like ls(1) but sorts filenames by extension, tersely
Does anyone out there think that Intel has been planning this for a long time? Or at least this sort of thing? Why else would Intel try to secure IP rights to AMD patents royalty free, years ago?
Not that I really feel bad for AMD; lets not forget people, AMD started business doing exactly this sort of thing to Intel. Personally, I can't wait for nanotech to take off to the point where at least home fab facilities are affordable. Then we can design our own cores and forget this whole debate altogether. For those curious, there are already people doing that in FPGAs; its too bad that affordable FPGAs are an order of magnitude or two slower than best-of-breed processors these days.
On the flip side if you find that you suddenly need a 256-bit bus or 8 pixel pipelines, an FPGA can reconfigure itself on the fly for that. It'd be great if every program carried with it a set of core designs for the various subsystems, and could reconfigure them on the fly.
I would love to have a computer with 4-8 FPGAs on a PCIX card, a GPU consisting of 2-4 high speed FPGAs, and a nice big high speed FPGA for the main processor. Need hardware SSL? No problem. Hardware MPEG2 to DivX transcoding? No problem. Highly optimized pixel pipelines? Just send me the bits baby...
Its a nice dream anyways...
For those interested in trying something like that out on their own, I highly recommend http://www.opencores.org/
I am disrespectful to dirt! Can you see that I am serious?!
How the hell do you mark an article -1 Toll?
It's nice to see an editor with a backbone for a change.
Trolls dont like to be Flamebait, because they burn so well. Protect our Troll heritage!
I would do it slightly differently: I am one of Windows' driver writers.
:)
See, the problem is that "driver" can't be an adjective-- it's not in adjective form. I see "driver" as the object noun in this clause, i.e. the drivers are created by the writers. So "writers" is the subject of this clause and Windows is a posessive identifying whom that subject belongs to. Really, you're referring to "Windows' writers" (who happen to write drivers).
Following the same convention, it would be UNIX's driver writers, Linux's application developers, etc. I put the platform in the posessive since-- in this sense, at least-- the developers belong to the platform. I think it makes more sense than the alternative, where driver is an adjective.
In fact, I'm really only an apprentice driver writer for Windows, but I would still describe myself that way
carrying compatibility baggage going all the way back to early DOS.(P> All these years I thought x86 was backwards compatible with the Intel 4004, and now you tell me it's actually backwards compatible with an old operating system! Well, I guess you learn new something every day...
"Freedom means freedom for everybody" -- Dick Cheney
"They laid down the law and said there would be only ONE 64-bit version of Windows XP"
Not quite. Their statement was more akin to "We will only support one x86 64-bit extention technology". AMD64 beat Intel to the punch, and since AMD64 was already established, Intel had no choice but to keep compatibility.
If they don't test it they can't guarantee it.
They wouldn't test compatibility, they'd just test that software runs on it.
And once Intel sells 10 times as many as AMD could hope to sell, who's compatible will be a question AMD will have to answer.