Slashdot Mirror
1100 MHz 'Athlon Killer' Due From Intel in December
jeffstar writes "According to this article at The Register, Intel has an 1100 MHz 'Athlon Killer' IA32 chip coming out. Yum, that's the kind of sauce I like." Sounds great. If it comes out - and performs - as promised.
Nonsense. They are, as every chip manufacturer is, pushing as hard as they possibly can to advance the state of the art. If a G4 stomps all over a PIII in certain applications, it does not mean that Intel has secret PIIIIs or PIIIIIIIs sitting in dark closets waiting to produce, it means that Intel is no longer producing the highest performance chips. What with the Athlon work being done, it is possible that Intel is now the weakest of major chip manufacturers now that Cyrixes and WinChips aren't exactly a factor.
Not only could they not 'release this processor at any time', they still haven't released it, and there is every reason to believe they will not release it in the manner they suggest either. It will take longer or be slower. They don't have stuff waiting offstage- this is _the_ premier 'paranoid corporation', the last one in the world that would be sitting around going "Ho hum, we got this chip here, seems to go real fast. Maybe we should make some of them and sell them, or gosh, why don't we just ignore the competition and tack it up on the wall for a while instead? It's real purty-like."
Uh-uh. Sorry. There is no Intel Fairy. That mystical creature seems to be hanging around Motorola, nVidia and AMD these days...
Do you realize how much 8 megs of SRAM would cost? SRAM is the whole reason that many processors (UltraSPARC, Pentium Pro/Xeon, ect.) cost so much damned money. If AMD had an 8 meg Athlon it would cost more than 2 or 3 Xeons.
I'm a loner Dottie, a Rebel.
One fellow ("ntsucks") said, "Either Intel has a stunning ability to improve its engineering process and timelines or they were withholding better chips until we had all purchase their current chip du jour"...
:)
To which I would have to add, "...or they are talking absolute crap".
Come on, people, get real. You're being manipulated. It's crazy to take some company's random promises as accomplished facts. You sure wouldn't do it for Apple, why pretend that Intel has a crystal ball? The intelligence you're insulting is strictly your own, 'cause other people are reading your comments and going "uh-HUH. Riiiiight. Aren't people credulous? Damn."
Seriously. Take a few deep breaths.
While that may be it's on-chip frequency that isn't always its emission frequency. If it were it's emission frequency there would be no way in hell the FCC would let it be sold in the US.
I'm a loner Dottie, a Rebel.
This is exactly what intel want you to think.
Microsoft use similar tactics - spread rumours about some wonderful product you're releasing in the near future, that beats your competitor's. Foolish consumers wait for your product, while you slowly let the release date slip, and slip, as all the while your competitor is losing business.
1. Megahertz is a dead end
Allready processors are too fast for the rest of the system. This has been alleviated for the last decade by an increasingly complicated system of caches and chipsets. At worst you'll go throgh 3 levels of processor cache, main memory, disk cache and finally disk, for a total of 6 levels of memory. This could go on indefinately but will have decreasing returns, unless the architecture of the computer can catch up to be generally faster. SGI/Cray has done this well.
2. Megahertz == Marketing
Ever since the P2, it's been terribly obvious that Intel just develops to satisfy what the majority clueless consumer wants- a higher megahertz number. The P2 made it blatant by being inferior to the older P's when run at equal megahertz. The only benefit was that it would run at higher megahertz.
Efficiency
No x86 has been really efficent- in many ways. More gates, more watts, more space, more heat. The unfortunate predominance of x86 is leading to space robots being designed with pentiums because Intel can push through to get the chips certified. When multiprocessing becomes a necessity as clock speeds dead end, who will be able to afford the power and large case for cooling that 8-64 P[3-5]'s will need? It's absurd.
Start Running Better Polls
JC (at http://www.jc-news.com/pc/) made a good point on his page the other day about this, I'll quote it here...
"Register put up a very interesting bit here. It's about a surprise Willamette introduction in February of 2000 ("paper launch" in December, chip actually appearing two months later, according to the article). I passed this by despite the fact that a good ten percent (slight exaggeration, but you get the idea) of y'all emailed the URL to me. It just doesn't seem likely, considering the design, to our collective knowledge, hasn't taped out (and if it did, it was likely recently). Takes about a year from tapeout to production. You do the math. However, as I said, I wasn't going to put up a link to it, but I just realized something (thanks to Jocelyn Fournier, I think, for nudging me in this direction). The specint95 score of the P7-1100 shown at that register article is utter crap. If it is really the case that it is that slow, then Willamette will be pretty pathetic for servers, especially if you consider the 1MB on-die L2. The quoted score is 43 at 1100MHz. By my guesstimations (with the help of idiot from Ace's), an Athlon at 1100MHz would score between 50 and 55 (perhaps subtract a point or two for dropoff from linearity), depending on whether or not you optimize for prefetching. This means that Athlon pastes these alleged Willamette scores in specint. Actually, from the look of it, given Intel's Coppermine presentation at PF, it seems that Coppermine is also faster than Willamette in specint. I didn't check at all with the Winstone score, but as you can see, if Register's data is true, then it isn't really great news for Intel. I don't know about you, but I'll prefer to believe the more reasonable assumption that Willamette will come out in 2000 Q4 (or 2001 Q1) but will be totally rippin' in performance."
0.18 micron does not refer to the wafer size. It is related to transistor size. The smaller the circuitry, the faster it can go for less power. If you want to bump up the speed, you have to raise the voltage, which allows you to go for high clocks and mo heat dissipation. I think the equation is somewhere is Hennesy and Patterson's book.
Hasdi
This kind of RAM uses a narrower connection to the memory controller (typically 16bit) than tradionnal SDRAM (typically 64bit), but transmits data at a higher frequency (350MHz here), on both edges of the clock (hence PC700).
Though you get higher memory bandwidth (1.4GB/s here, versus 0.8GB/s for PC100 SDRAM) you must be aware that the memory latency is worse.
See Rambus, Inc's web site
I've seen a fair few comments about the Register being biases against Intel and MS, but I've got to say that as far as I'm concerned their coverage of future chips and stuff is pretty accurate. If they say it's doubtful, I'd tend to believe them.
As happens way too often, /. managed to lose my first response to this, while giving every appearance of having accepted it. *sigh*
;-) Basically I haven't seen much in subsequent processor designs to counter the excellent arguments H&P put forth in their book regarding How It Should Be Done.
>What you describe is essentially the initial MIPS project started by Hennessy at Stanford.
I'm glad someone noticed.
>Most delay slots are never filled
Obviously, this can vary a lot, but according to the studies I've seen a single delay slot can be filled with something besides a NOP >80% of the time for most kinds of code. The important thing is that a NOP is no worse than a stall, except that the stall usually has a lot more wasted circuitry associated with it.
Yes, dynamic scheduling can do "better" than static, but at what cost? Does the improvement make up for the additional complexity and limitation of clock rate? More importantly, are there other things that can be done with that real estate which provide even better bang for the buck?
Slashdot - News for Herds. Stuff that Splatters.
- WPI * IPC * CPS
Classical CISC architectures tried to maximize WPI, and this limited the other two factors. RISC was mostly intended to maximize CPS, intentionally sacrificing WPI to do so. Pipelining, superscalarity, and branch prediction are all targeted toward increasing IPC in different ways. VLIW and EPIC improve either WPI or ICP depending on how you look at it.WPI = Work Per Instruction
IPC = Instructions Per Cycle
CPS = Cycles Per Second
All of these approaches to improving performance tend to have characteristic challenges associated with them. In the current case, you have to deal with the fact that massively superscalar architectures require an instruction stream that keeps all the functional units fed. That means that compilers have to try to resolve data dependencies and competition for functional units, either of which would cause a stall, and also deal with branches which cause bubbles in almost any architecture. It's a very tough problem, which is why chip designers turn to second-order tricks such as speculative/predicated execution and VLIW/EPIC.
Personally, I think that's all a trap because it causes chip complexity to skyrocket and undermines the very idea of RISC. If I were designing a chip, my goal would be to crank the frequency sky-high and make the compiler (or a translating front-end processor such as Transmeta is rumored to be working on) do most of the worrying about how instructions get scheduled. In particular, I'd go for:
Slashdot - News for Herds. Stuff that Splatters.
I don't think you need a 1GHz chip for word :)
processing. I do think you need it to solve
large eigenvalue problems and run atmospheric
model simulations. A PC/Linux combo is an
excellent alternative to expensive workstations in
scientific research. It's ironic that these
super-fast chips are really overkill for 99.9%
of the population, but for the 0.1% of us who do
serious number crunching it's a great deal!
The Register article you link to is very significant, but perhaps not in the way you intended. On 15/04/99 the Register reported the following:
Intel is twisting the knife by showing OEMs performance predictions stretching out until late 2000 featuring a Willamette IA32 processor rated at 1100MHz competing with an AMD K7 at a paltry 666MHz.
No specific figures are quoted, but graphs pitting the rival chips against each other show the Willamette 1110MHz scoring around the 50 mark in Winstone98 against the K7 666MHz at 35. On SpecInt95, Willamette reaches 43 against the AMD part's 20.
The same graph shows a 666MHz Coppermine appearing in late 1999, a clear 12 months before AMD is expected to reach the magical figure.
And perhaps more worryingly for AMD, a Coppermine-based Celeron appears in early 2000 (probably at 500MHz and 100MHz FSB with Streaming SIMD) which is predicted to perform almost on a par with the K7 666 reckoned to be due 6-9 months later.
Rather than demonstrating inaccurate reporting by the Register, this report simply presents Intel's OWN predictions.
It appears from this that Intel was expecting AMD to be unable to supply 666 MHz Athlons until Q4/2000! As you can see, Intel's current production is right on target, but their predictions for AMD were way off!. AMD is over a YEAR ahead of *Intel's* schedule. There's no way for them to adjust for this misprediction quickly, so expect Intel to lose a *lot* of market share to AMD over the next year.
The problem is, it's very tough for the compiler to do a good job scheduling statically. Most delay slots are never filled. Much more information is available at run-time (in a limited window for the hardware), so it can make some better decisions than a static compiler can.
However, the compiler can look much further ahead than the processor, so it seems that some sort of hybrid solution is called for. Whether that involves profiling and feedback optimization a la FX!32 and others, new ISA or something else is still an open question, I think. IA64 has made steps in this direction.
--
- A.P.
--
"One World, one Web, one Program" - Microsoft promotional ad
"Remember when the U.S. had a drug problem, and then we declared a War On Drugs, and now you can't buy drugs anymore?"
Yes, stupid question I know, ofcourse we want it, but what I mean is should we be so eagerly anticipating something that will simply have more resources to be eaten up by sloppy code? (Let's face it, M$ still hold the biggest share in OS, despite best efforts) Shouldn't we be also pushing for tighter and better coding, which would fix probably half the resource problems we have now?
nice speed, but do i have to install it in my freezer.
* * * --they cant all be your best, that would be confusing
"Simpler to modify and read" in what sense? If you mean it's easier to read programs written in those languages, and to modify an existing program written in those languages, how much of that is due to the language and how much of it is due to its implementation being interpretive?
(Is "interpreted" a property of a language or an implementation? I think the first LISP implementations were interpreted, but LISP compilers exist; most C implementations are compiled, but I think C interpreters exist. I could imagine Perl or VB implementations that generate compiled code - I have the impression that VB code can be compiled into machine code - and if you were to translate one of this sort of language into, say, Java byte codes, and to run them in an environment with a JIT compiler, is it interpreted or compiled?)
Some of the benefits may be due to the implementation being interpretive, e.g. an interpreter might be able to do a better job of telling you where something blew up (although symbolic debuggers can, at least sometimes, do a decent job of that, at least if the code is unoptimized), but I'm curious whether a sufficiently clever non-interpretive environment could do as good a job.
I.e., speeding up "higher-level" languages might be doable by means other than throwing faster processors at them; one can debate whether they're better doable by those means, but that's a separate question.
But, yes, it's not ipso facto the case that faster processors server only to encourage sloppy code; some might debate whether software and what it can do has progressed in any truly useful fashion since the days of the Manchester Mark 1, but....
It wasn't even six months ago that people talked about AMD's "Pentium Killer". Now its the other way around. Changes fast doesn't it? Used to be everyone after Intel was trying to make the Pentium Killer. This is the first time I recall in x86 land that Intel is the one making the "Killer".
Perhaps this a true sign that AMD is a legitimate competitor to Intel; not just in the low-end but the high end too. If you didn't think that already.
VENI! VIDI! VICI!
----
Every year during my review, I just pray the words "slashdot.org" aren't mentioned.
Seriously curious...what would be the point of this for the majority of us? I'm running a single celeron 400 (soon to be dual) and it does everything I want, without problems. Kernel compilations are in the low single digits, I can play any games full speed, why does the average person here need something this fast? We don't, other than possibly being able to say "haw, my computer is faster than yours."
:P
Just some thoughts...though I wouldn't complain getting one of these things for my birthday or anything
This sounds a lot like more vaporware.
Sure, it'll come out eventually.
"Released on paper sometime in January" with the chips actually available sometime around two months after that. Now doesn't that strike you as equivalent to "The check is in the mail?"
I want one. We all want one. But announceing plans to release something drastically cooler than everything else on the market should require a definitive time frame, especially when using that many "killer" buzzwords.
Athlon killer? Who even has one yet? Where do I get a mother board for an Athlon?
I can't believe this was anything but the PR departments intentional release of memos to get noticed and to try to take sales from AMD.
Computers can only simulate determinism. ~Hermetic.
The Register didn't have any sources. Sorry, but if you haven't read the article, do so.
;)
"We know, from a highly reliable source..."
"It's also worth referring back to this piece, which also came from a highly reliable source..."
"Another reliable source tells The Register..."
"One US source says..."
Hehehe. Boy I get a lot of laughs out of this kind of journalism.
But seriously, it seems to me that at this kind of speed (if it were to be true), the processor isn't going to be the bottleneck (but that will vary depending on what you are doing, of course). The slow point for most of the things I do, is, believe it or not, my internet connection. (And I'm over cable modem.) Give me a low end pentium class machine and a blazing link, and I'll be a happy man.
However, that is all just my opinion...but I got it from a reliable source.
This sig is false.
"Revolutionary" in what sense?
The 386 was the first 32-bit x86 processor, and the first one with support for demand paging - it had a new instruction-set architecture. Not particularly revolutionary in general, but revolutionary for x86.
The Pentium implemented the same instruction set architecture (with some minor additions); it was primarily revolutionary in its implementation, in that it was the first superscalar x86 chip to ship (again, not particularly revolutionary in general, but revolutionary for x86).
The latter means that, with Pentium, they pretty much, well, "went for the speed race".
...or getting it from HP. (I have the impression a lot, perhaps most, of the ideas in IA-64 came from HP.)
The Lame Unit In I.T. does have a new instruction set, because it'll be the first IA-64 implementation; is that the kind of "better technology" you're looking for?
It won't be an "Athlon Killer" unless it is competatively priced. Assuming this report is reliable, Intel takes it from paper to silicon, and a lot of other stuff, it still won't appeal to te typical computer buyer (which ain't us anymore) unless there's not too huge a price gap. Which would mean Intel selling under cost yet again, and how long can they keep that up? Fiscally, a long time, admittedly, but I'm talking logically.
fh
But I'm doubtful. Intel doesn't have a particularly stunning record with delivering chips early and I'd rather not buy one of their step 0 chips anyway.
Let's see, AMD gets market share and major recognition with a quality product, and now suddenly Intel is claiming that it can suddenly make much faster chips RSN. Whatever.
I'm personally sick of talks of vaporware. I love new technology and reading about the future, but I don't buy my computers based on speculation from unnamed sources regarding the possible date that a chip will get put to paper. It's utterly irrelevant.
Call me when it's in silicon.
it seems to me, that adding pipelines (moderate space cost) and execution units (higher space cost) would be bring more performance than higher clock rates. that is, if the same effort to increase clock speed was put into superscalar expansion, the pay off would be greater, provided there's enough space.
also, it seems to me, multiple short pipelines would yield higher preformance than fewer high clocked, deeper pipelines.
i believe the reason intel goes the faster deeper route (compared to slower, wider) simply cuz:
1) it's cheaper to deepen pipelines and it isn't too hard to get a good enough signal to noise ratio for higher clock speeds (is that even an issue?)
2) marketing. this is the obvious one. they can say "our chips are fast! more MHz than our competitors" and the general public doesn't know any better.
ps, please correct me on anything, i'm just guessing at some of this stuff
You'll need a 1100mhz to run the OS minimum.
1. The register is a rumour mill
2. At 0.18 micron this stuff needs a supa-dupa cooling system. Maybe with sharper fab, you can get this speed
3. Needs very large cache and very wide memory bus and heavy interleaving because the last time I checked the memory is still running at 100MHz max.
If I were you, i'll either get a dual celeron bundle at $799 or a 400 PPC750 with monitor also for $999.
Actually, the alien overlords have allowed Intel to bring forth this new chip for the good of mankind. However, you might want to wear tin-foil and collanders on your head, lest the RF emmissions from this Gigahertz+ chip turn you into a mindless overclocking zombie (An unfortunate side effect discovered when the celerons were first overclocked.)
HOW DARE YOU!!!?!?!?! Are you crazy???? Of course you, as Joe Consumer, need an 100 MHz CPU for your Word and Quicken programs!!! You should be ashamed for being perfectly happy with what you have!!! You MUST feel the need to upgrade every time Intel releases a CPU!!!! You will be assimilated!!
I'm left wondering if this article is going to be any more accurate than one the Register ran earlier this year when they said that the 666MHz Coppermine would appear in late 1999, "clear 12 months before AMD is expected to reach the magical figure". Yeah, right.
HH
Yellow tigers crouched in jungles in her dark eyes.
She's just dressing, goodbye windows, tired starlings.
largely that it's very hard to parallelize code so that you can run it through separate execution units without stalling the processor. With the Pentium's two shallow integer execution units it was possible to hand-optimize your assembly to keep the two pipes filled. But breaking up code that is linear in design (i.e. most programs have a single "flow" and assume linearity of execution as their core model) into parallel chunks is a hard problem.
Continuing down the "more, simpler pipes" path is akin to explicitly parallel chips. It's a hot area of research, and there are some applications for which it might pay off (the ones where multiprocessor machines already pay off, perhaps: servers that are doing several unrelated things at once) but for doing just one thing and doing it fast, faster deeper is probably far easier a problem. Remember, Intel has had problems with the old P6 core (ppro/pII/pIII) because it's already very hard to write a compiler that doesn't stall it left and right.
With all that said, I don't see any mention in this article about the actual design of the new chip, except for some very vague (and likely wrong imho) stuff in the article about Wilamette that's referenced in this one.
If you look through the back-stories of The Register, you'll find that basically everything they've ever said has turned out to be a lie, or unproveable. I'd take this story, and the one about the 1GHz copper Athlons in January, with more than a little grain of salt. Also, those SpecINT numbers are much, much worse than even a P3 overclocked to the same speed would post. I think The Register can't even be bothered to lie convincingly.
--Conquering the Earth Since 1978.
One step closer to the frequency of a typical microwave oven (around 2.4 GHz)... Imagine, your pizza will never get cold if you just put it on top of your tower.
God, root, what is difference? - Pitr
I'm happy to see that the GHz barrier is likely to be broken before 2000, if just barely (though you have to wonder how much vaporware this is.) As for the MB RAM barrier, I guess it's always possible, but it's starting to sound like overkill (well, maybe not for W2K, but certainly for most of the uses I make of apps under Linux!)
Now all I need is a 1 TB hard drive to go with that 1 MB RAM and 1 GHz processor. In a Palm Pilot. There's nothing like misusing power to put any 1990 supercomputer to shame on playing X-mines!
"There is no surer way to ruin a good discussion than to contaminate it with the facts."
Intel didn't release ANY information about the Willamette (P7) at the Microprocessor Forum. Now, as much as people want speed, Intel also has shareholders to appease. If they knew about an impending P7 release, they'd have to make that knowledge public otherwise they'd be misleading shareholders to believe that the Coppermine is the Q4 1999 and Q1 2000 contender.
While we're on that subject. If Intel does paper release the P7 in December, they've pretty much signed the death certificate for the Coppermine and PIII line. Now Intel's a marketing genius (love them/hate them for their technology, but any company that can convince people they need a PIII for the Internet has strong marketing) so there's no way they'll throw away all those ad dollars on the PIII line quite yet.
The Register had been getting better, but this is reverting to their old self...
AMD has been trailing Intel since the 286. Intel
has had plenty of time to destory AMD.
Thanks to AMD and others, Intel is not a monopoly.
And, you can buy lots of nice machines for what
a fully loaded TRS-80 cost in 1978.
Funny - my Celeron 366 is currently running at 550, and my P3/450 is running at 600...
1.1Ghz? Intel? So is this a replacement for my
stove or central heating? Do I need one of those
big restaurant freezers or can I just move to
Nome and keep it outside?
I believe the fact that Intel has had many high cost machines to even release to the public due to cost. How many want to bet that intel has machines in there building running at even higher than 1100.. Think about it, they just release stupid ?33's, ?50's and ?66's just to have more money in their pocket. People will by whatever is the cheapest out and most cost effective.
In the past, Intel has always been able to dismiss their competition as inferior. Oh, sure, there were times companies like AMD or Cyrix had a slightly faster chip, or better price-performance on the low end, but they were always brief and/or insignificant. Undercutting prices is a common enough thing on the low end; AMD/etc simply has to reduce costs below the giant Intel.
/. example, does a few things better then Linux.
Athlon was different. Athlon challenged Intel on Intel's home turf, and won. It was the fastest high-end x86 CPU around, and is going to stay that way for at least several months, if not longer. Intel had a serious threat for the first time. AMD may still be small compared to the behemoth Intel, but David was smaller then Goliath as well.
The fact was one thing, but as we know, the spin can be another. Intel could have found some sort of flaw in Athlon, or fired up the FUD guns. In most cases, you can argue some point or other as an advantage over your competitiors. Even Windows, to use the favorite
But Intel did not do that. Intel could not find a way to counter Athlon in the trenches. Intel looked for ammo, and found none. For the first time, Intel looked at the competition, and found itself unable to immediately compete!
Now Intel is scrabling to catch up, to try and build a counter-weapon to use against Athlon. The fact that they feel the need to "kill" Athlon is very telling. It is one thing to know you have a threat. It is quite another to classify it as the threat.
By accepting AMD's challenge, by admitting that the Athlon is strong enough that they need to target an entire product series at it as a "Athlon killer", Intel admits that they have lost a battle. That AMD has stole ground away from them. That Intel is wounded enough for it to hurt.
It may be only in pride, or in market perception, that Intel feels pain. Their sales are still large enough that AMD is no immediate financial threat.
But suddenly, the small fry that they paid little mind to before, has woken up and bit them hard. Hard enough for Intel to step back, shake itself, and wonder what to do about this new threat.
I imagine the British felt a similar feeling when their American colonies fought to break lose -- and started to win.
It will be very interesting to watch this war as it unfolds.
dragonhawk@iname.microsoft.com
I do not like Microsoft. Remove them from my email address.