Slashdot Mirror
Globe covers Possible Pentium III Flop
DrPsycho writes "Canada's national newspaper
The
Globe and Mail has in its Report on
Business section today an
article on everyone's
favorite overhyped chip, the PIII. Talks plainly
about the competition Intel faces from AMD and
Cyrix, as well as the problems they're having
with PR and a general lack of consumer confidence.
" Maybe its just me, but wasn't the 486, Pentium, PPro
and P2 each predicted to be a flop? The P3 is definitely looking grim
though. No major performance game. Many times more expensive.
Hmm
WTF is a "performance game"
You'll buy it. Because Intel will stop making PII's.
I think Intel will really have something to bring to the table when they put the cache back on the chip in the PIII like they are planning on doing here soon. Get a Celeron until then...
It won't flop - with a PIII/450 only being $20 or so more than a PII/450, there's no reason for anyone not to buy the PIII, and only the PIII will be available in the higher megahertz ranges, at least for the time being (500, 550, etc.)
I think I'll say PIII some more. PIII PIII PIII.
Linux get more popularity than MS would want. Originally, they used Linux to say they had no monopoly even though they really thought Linux wouldn't hurt them... but low and behold, already it's getting attention. That demonstration didn't get the refunds but it got massive attention. Linux and the trial are why M$'s stock is hurting lately. Also, AMD and Cyrix will get more publicity and will be accepted as equally good chips, and if they can handle it, a good bargain. 1999=bad year for WinTel, I think. Keep it up d00dz! Riots==good for media attention... see europe and kurds.
Last time there was no K6-2, K6-3 and the sub $500 computers. Let's see how someone creates a sub $500 computer with a $600 chip. How many people want to buy a $2000 computer today ?
Also, there are no current use for 3D in home or biz apps. Maybe games but the K6-2 and a 3D graphics card should beat the expensive PIII.
If that not enough there is the Tracking number issue. How many people want to go naked on the Internet ?
Well, every generation comes to an end. Bye, Bye PIV.
Maybe I'm wrong, but I don't think the people who
benchmark the new chips use software optimized for the new one, they
use software optimized for the old ones. If you use
software optimized for a PII on a PIII, you shouldn't expect it to
be much faster. This is in contrast to the PowerPC, in which
the chip (the G3) is optimized to work with software, rather
than the software having to be optimized to the hardware.
I was hoping that the PIII would gigaflop.
Apple PIII? Boston Creme PIII?
Don't you realize what this is ??? Another company pushing out a product with a nice little name to satisfy customers, and sell the previous product... Now that the PIII is out Intel will begin selling more PII's then ever as well as the PIII. Hopefully this will drive down the price of 486's so i can get that beowolf cluster ive been dreaming of :)
It sounds like this chip is a little bit better than the P-II but not really much different. The only thing wrong is the price. They'll gouge those with deep pockets who have to have the latest thing, but after a while the price will come down. Intel has made boatloads of money from every processor they've released, and I doubt this will be an exception.
The best thing about the P-III is that Intel will have to cut prices on the P-II and Celerons.
I don't think Intel will enjoy the huge sales that it has in the past with its previous microprocessors. The PIII is extremely pricey and dosen't really provide the same bang for the buck.
Consumers (in my experience) want something that is fairly cheap. AMD and Cyrix will most likely enjoy an enlarged market share soon. AMD should really win with the K7.
I don't have any hard data to back this up. It's just a gut instinct, but I am certain that Intel is going to be treading dangerous waters. Intel will have to sacrifice x86 compatibility to maintain its position in the chip market and x86 is where "it's at" for most consumers.
Geeks. My last PC was around 2,500. And it was not a brandname, I hate giving out money for "support" I never use and their "marketing" efforts. So, for this price I had heavy duty machine to play and work with.
Wake up, there will ALWAYS be cheap CPU for basic usage and really powerful machines. Some people will always buy fastest (most expensive) chip available.
Every new Intel product has had a significant price premium attached to it: that's just the way it works. Three months from now, if not sooner, the price cuts will begin.
Do you think that Intel marketing hasn't figured out *exactly* how many PIII's they're going to sell at their target price?? They're not selling big volumes to end users: I'd bet that >90% of their capacity is already allocated to OEM's who placed their orders a long time ago. Intel doesn't have to wait and see if the PIII will succeed or not: they already know! If there were a problem, they would have adjusted their prices prior to the rollout.
The PIII has the same core as the PII, so it should be faster by roughly the ratio of their clock speeds, all else being equal. It remains to be seen how much of a premium KNI is worth: the software isn't there yet, so today's performance comparisons are not useful.
As for the K7, it can't start eating Intel's lunch until it ships in quantity. Which could be never, based on AMD's uninspiring track record. What if it is late, doesn't meet clock rate targets with acceptable yields, or has process problems? All of these have been the rule rather than the exception for AMD. Maybe they'll pull a rabbit out of their hat on this one, but maybe not.
I will NEVER own a CPU with a tracking number. It is WAY too easy to turn it back on. Your average Joe-Dumbass may not know the difference or even care, but I like the tiny fraction of privacy I have left.
at the time I worked in a lab, which got all PPro's. 25 all total. They were nothing special. I got a Cyrix-PR2/233+ for my home system, which kicked the crap out of it.
PPro's were garbage.
Also makes pizza taste better.
Intel's marketing people have their heads squarely up their collective asses! If the PIII is a flop they should start firing some of these clowns.
Why all the secrecy about KNI, and what about all the tools that were supposed to be ready when KNI/SSE is released. Haven't seem anything from MS, Intel, Watcom, et al.
It's as if they think techies can't work on ideas and concepts without having hardware for a couple of months. Hey if you can't give us hardware, a simulator or too would be useful. But no only Intel's special developers get advanced information and none of the rest of us can get any help. Hope the FTC is paying attention!
"The Pentium III is the power option," he said. "It's the difference between driving a Chevrolet and a Ferrari."
Uh, no, it is more like the difference between a Ferrari and a Ferrari with a PIII sticker on it.
Save your $$$ and buy a 300A Celeron while you still can.
I wish people would get over the FUD on this.
It's not like Intel was the first to come up
with the whole idea. In fact right now there
several companies already shipping machines
with unique identifiers. All the new macs have
ethernet cards and serial numbers in their ROMS.
Sun, SGI, HP, DEC, etc. all ship machines with
ethernet as a standard component and they don't
even make efforts to protect that information.
All "brand-name" hardware vendors are now
shipping ethernet as a standard component as
well.
Duh.
The P3 won't flop because vendors like Compaq, Dell, Gateway, HP, etc. will put it in their new models. Most people will naturally go for the biggest and best chip on the market. Besides, Intel has plans to stop manufacturing the P2 in a few months which will leave you will no choice (except for AMD, Cyrix etc).
Some supporting information would be nice.
How about mentioning some benefits of the K7.
Does it fit Socket7 or SlotX?
Floating point performance?
Power/voltage?
Cost?
And the number one question - Will it run Linux?
(Stupid question!)
I don't think they will take PII's out of the market. I reckon that they will sell PII and PIII for some time. At least until PIII's prices come down.
Intel isn't running an de facto monopoly anymore, competition is getting though. They will have to sell cheap chips to mantain market share. They will also have to sell expensive ones to make money.
This is why I say that they will go on selling PII and PIII together for a long time.
Regards.
Yeah, the MAC address. I've never seen a card
that supported firmware level MAC changes. It
completely breaks the idea of uniqueness which
was a big priority for D-I-X during the dev
of the standard.
Regardless of how it's touted (which was quite
dumb actually, because spoofing it will be
trivial for a creative hacker) the idea of a
unique identifier associated with a particular
computer is old hat. The ppl at privacy.org
and such must have some kind of "secret" agenda
to pursue. Probably just bashing the big dog.
IP masquerading and it's more common industry
incarnation (nat) require tables on the nat
machine to route incoming packets. Connections
that are initiated on the "inside" cause the
gateway/nat machine to create a table entry
mapping your regged address and a port number
to the internal address and port number.
Connections that are initiated outside require
setup on the nat box specifying where the
connection should be routed internally.
As far as MAC addresses being used on the
network, your local ethernet address is dropped
from the headers of packets at the first router.
(Which is not to say that you can't put your MAC
address into the data portion of a packet and
send it out.)
And since all the software we're running is 'optimized' to run on a 386, who cares what new-unused-crap-of-the-month they stuff into the CPU. The Intel/AMD/whatever chips are just glorified overclocked 386 emulators anyway.
Has anyone gotten an information on what will be eventually published as the recommended hardware specs for running Win2k?
MicroSoft definately has an impact on making/breaking the success of hardware. Hard drives over 2 Gigs didn't come down in price until Win95 OSR2 came out. USB and hard drives over
8 Gigs started to become popular with the release of Win98. Maybe MS can pull another save for Intel. (Wasn't there between beta releases of NT4 a release that forced non-intel CPUs to run alot of additional debugging code making Intel processors look like a much better choice for NT4?)
So, whether or not there is any performance gain, we'll all buy PIIIs so that we can be tracked continuously. Whether the PIIs dry up or not, whether there are AMD or Alpha or Sparc or Cyrix or PowerPC or anything else cheaper or better, we'll want the tracking feature.
The fact of the matter is, Intel has advanced the state of the art of x86 compatible chips. The delta speed increase or the initial cost are not the themes one should be concerned with.
What people should bring away from this announcement is not, oh-my-god, that Intel rakes in ample dollars and not that Intel's marketing department has gone mainstream; but rather that the spectacular pace of PC innovation marches on, and that the immediate future (with new memory technologies and
So, just feel good about that and stop misleading poor slashdot readers while playing armchair pundit.
I've heard this bullshit all the time too, that the consumer market is pushed by games, however I would like to see one shrewd of evidence that backs up this claim. Why would anyone want to spend thousands of dollars so they can have headaches trying to get games to work, only to find out that they don't have the latest and greatest 3d card so they have to shell out more money. Instead one could by a console for less than $150 and not have to worry for years about upgrades an dplay games by just popping them in. (Yes I know high end PC's are better than consoles, however is it really worth the extra money?). If someone just wants to play games, they will get a console not a PC.
Well. Like my grandfather used to always say, "Pussy just plain tastes better when it stinks."
Seems like most of the people who post here are gamers, for some reason... "Forget the PIII, buy a 300a and two voodoo 2 cards!"... Somehow I doubt having two voodoo cards will help my build times in visual studio or shave time off a large-scale ocr project... If all you do is game, stick with the low end and stop complaining that the piii barely makes a difference in your quake framerates. It only makes you look like children.
Last july AMD had a prototype K7 running at 500 Mhz. And I heard recently Intel had a 650 Mhz Pentium 3 in their office. Does it mean they will release it. NO. They'll only release it when AMD and Cyrix have a chip threating their current flagship.
At the time of the PPro release, most people were still running 16 bit stuff, including the games by which you were doubtless judging performance. The PPro was never intended for games, 16 bit stuff, or any consumer version of Windows.
The mere fact that you believed Cyrix's lies abut their performance is proof that you do not use your computer for real work. Cyrix has never made a chip that lived up to its billing, and certainly never actually ran the advertised speed.
In other words, they love to LIE (not much different that Intel's approach to comparing the PII with PPro).
get an SGI not a pc
KNI isn't going to do jack for your compiler- it's for massive uniform processes, like audio and video (and possibly OCR and voice). And you're better off with a cluster of cheap MIIs or something unless you need real-time output, thus the emphasis on gaming and other high-throughput entertainment.
Well I don't know. I once worked on a
large ocr project for the IRS and I was
trying to figure out how to slow it down.
The K6-3 is much better at games. In fact it beats the Pentium II at equivalent clock speeds when 3DNOW! is being used.
You can today buy a well equiped gaming PC for under $500. That to me is quite worth it.
With a console, you're limited to ONLY games.. and each game costs $60-$70 (or more!) to buy. After that, you're limited to multiplayer via one screen, which imo takes all the fun out of it. (Deathmatch on one screen sucks.. oh gee, where is my opponent? look at his section of the screen!)
Not only that, but with consoles, every time you buy that "$150 console" you need to buy ALL NEW games. Add $300 (for ~5 games) to the cost of a console, and a PC isn't so expensive. Generally, with a PC, you can just reinstall all your old games, and they will work.
On top of that, you can get upgrades for your PC games (I *DON'T* mean bugfixes here..) such as new maps, levels, player models, sounds.. whatnot.
Usually your 3D hardware will last for a couple years.. and it costs under $100 to buy a good card nowadays. My 3 year old Voodoo still works great for all the games out there. Granted, it's not top of the line, but it's not bad, either.. that $150 I spent back then was to me worth it. (And I already *HAD* a PC at the time.. so total cost was $150 to me)
And to finish it off, you can do an almost unlimited amount of other things with your PC. Read newsgroups, surf.. email, IRC, do your taxes.. buy things at near wholesale prices.. you get the idea.
So for $550 for a console, or $500 for a PC, I'll take a PC any day.
Maybe in 16-bit software but that's the only area where the PPro was weak (and I hardly consider it a weakness since that is not what this chip was designed to run). Why do you suppose that there are no other chips used in the biggest x86 boxes except for the Xeon? Besides it's unmatched scalability (no other chip out there is as easy to do SMP on) it is also the best x86 out there in terms of raw performance per clock speed. A 200mhz PPro is faster than a 200mhz P2 (except for MMX-enabled apps). It's got a faster cache and that's what makes the difference. Don't beleive me? Just over-clock a PPro 200 to 233 then run an CPU-intensive benchmark against it. As other industry experts have said, the Pro eats the P2 alive .. leaves it in the dust. It's second only to the Xeon in the x86 world. In fact it is a Xeon but without the 100mhz front-side bus and MMX instructions.
Flop? Sure on toy-boy boxes made by Packard-Bell but not on servers and workstations! A dual PPro 200 runs as fast as a dual PII 266 (due to it's full-speed cache). And until the Xeon came out, it was the only processor out there that was widely available for four-way SMP or greater.
is that Merced is a complete fucking disaster, and Intel have had to go back to HP for PA-RISC for Merced's successor.
Now, Intel are gods, and HP 'also rans' in the hype department, so HP arent allowed to talk about this, and 'McKinley' [is that a river, or a president, or both?] is actually a HP chip marketed by Intel, cos Merced sux majorly dood.
or so I'm told.
Vryl
That's why I'm using a K6. (actually the effect is that I've been using AMD chips for 6 years now, 486DX40, 486DX4-120, K6-200, K6-300).
This is an often quoted myth. The truth is that for some time the PentiumPro was the fastest CPU for integer operations commercially available on the planet, using sound benchmarks, i.e. SPECint. I'm remember this because my professor of microprocessor architecture, a strong RISC propnent, was quite annoyed about that. Of course, the new generation Alpha, MIPS or UltraSparc was sold/benchmarked shortly after that, and Alpha has regained the lead now. But it is quite amusing that PC press consider the PentiumPro as a big failure from Intel (because of 16 bit code), while in truth, never an x86 Intel processor has been more technically successful. Hell, NT stations could run some code faster than any workstation back in that days (Linux was could nearly too :-).
OS for MMX is a totaly different ball game.
MMX uses the FP registers, which are saved during a context switch. (The processor mode will also be saved, so everything will be restored -- KNI on the other hand, uses separate registers, which wont be saved during a context switch, unless the CPU itself (from P6 upwards) does this).
This means that if you use two threads which
use KNI, and the OS doesn't perform context switching correctly, then incorrect results will
start popping up (when one process sticks something into a register, the other thread puts something else in, and then then first gets control back...well sortof)
- SpecInt results 09/1995 (before PPro)
- SpecInt results 12/1995 (PPro)
- SpecInt results 1Q 1996
- SpecInt results 2Q 1996
.I'd really really like to see your benchmarks. (Would be Bogomips, methinks).
> Good one. Hehehe. An "Internet chip." Now
> that's f'ing funny. Like a CPU has *anything* to > do with the Internet.
Then its just a coincidence that ~99.9% (or roundabout) of Internet attached devices use a
CPU? (p.s. I am counting routers and switches)
More seriously, efficient streaming needs chips
that can handle the sorts of processor load
generated -- this allows the compression/streaming
software designers to assume that the decompression phase doesn't have to be 'easy' for
the CPU.
The Pro has full-speed cache in sizes from 256k to 1mb. The PII has 512k half-speed cache. The PII has MMX. The Pro doesn't (except the new PII overdrive). They are the same core however.
In fact the Asus P2B already HAS a 133mhz setting (with 33mhz PCI) - if your RAM is good enuff it'll work.
Sadly since Intel chips are clock-locked nowadays this is only useful with an old Klamath chip which is only multiplyer-limited. Or MAYBE a 350 which MIGHT be able to reach 465 (but I doubt it)
Wouldn't a Celeron 400 be cool on a 133mhz FSB * 3.0 multiplier? sigh...
Remember the P2, Xeon, Celeron, and P"III" are all based off of Pentium Pro technology.
So even if the first version 'flopped' - it's still a good chip anyway, and it's decendants are definitely NOT flops by any stretch of the imagination.
I don't think that, with today's hardware capabilities, anyone with more than a Pentium 200MMX (a.k.a. 604/150) needs a special set of instructions to stream Internet multimedia. Since, as far as I know, the PIII's target, being quite costly, is Professionals who do not want special Internet Multimedia instructions, I don't think all of this is necessary. The consumer is still on shipping 56K [and slower] connections and can't take advantage of the new instructions anyhow.
I say that if you must go Wintel, go Celeron and clock the bastard. Run Linux of you've got the balls.
We're all 10 years ahead...using Macs.
And I quote -
"designed to take advantage of the internets capabilities for full motion video"
Since WHEN has the internet piped enough data to warrant a processor DESIGNED to deal with it???!!!
Or is this intel just admitting the, comparativly, poor performance of the previous pentiums?
- 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?"
Yeah, and it took 9100 of the little fuckers to best machines by SGI that used about 1/10th as many processors. Real powerful stuff.
- 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?"
- 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?"
"to drive sales of coming products that need more "horsepower" to operate efficiently"
-----
Only 28% of slashdot readers use Linux or *nix, while 54% of them use Windows. How ironic.
If you can read this message, your threshold is too low.
Of course system vendors will still put PIII
systems high on their price sheets and push
them whenever possible. Home users may not
but PIII's but you can bet businesses will
-- it is unheard of not to by the latest/greatest
hardware in a corporate environment.
"Intel is positioning the Pentium III as the "Internet chip," the first microprocessor designed to take advantage of the Internet's potential to deliver full-motion video and other advanced multimedia applications."
Huh? Wouldn't bigger pipes do a better job? So now clueless newbies will have the internet in their new 'puters. It says so right on the chip. wowzza!
"shop smart:shop s-mart" ash
The price issue will probably go away soon. Intel just wants to grab all they can from the rabid early adopters. When that initial sales spike starts coming down, so will the price. I do suspect that will happen faster than they anticipate, however.
The PIII probably says more about troubles ahead for Intel. The less than exciting improvements in the PIII vs. PII shows that the chip family is probably nearing the end of it's life. There's just not much more to squeeze out of that archetecture.
AMD could present some big problems for Intel. The K7 looks like a win. Meanwhile, Transmeta may or may not be developing the Intel killer right now. Time will tell.
The P3 doesn't thrill me one bit, and I won't recommend or buy 'em. The P2s are good enough and cheap enough - 450's or even Celerons.
Ever hear of a Sun/SGI workstation? :)
Sorry, the statements "I do real work at home"
and "both ran win95" do not belong together.
As for your benchmarks... Win95 is a 16 bit OS,
PPros run slowly on it. NT or Linux run much faster on PPro than Cyrix shite. Try it.
http://rareformnewmedia.com/
Looking on pricewatch, a PIII 450 can be found for
cheaper than 550$, whereas the lowest prices for
the PII 450 are around 500$. The price difference isn't really as significant as the article makes it out to be. Of course, this is all kind of silly considering a Celeron 400 can easily be had for 150$ (and a K6-2 400 can be had for the same price).
--
Kevin Doherty
kdoherty+slashdot@jurai.net
Kevin Doherty
kdoherty+slashdot@jurai.net
Each generation of CPU Intel has come out with has been just a whee bit better and first, and then suddenly it became a whole lot better.
:)
Going from the 386DX33 to 486DX25 was no big deal. Going from a 486DX2/66 to a P-60 was no big deal, in fact the P60 overheated so much people generally didn't recommend it. By the time the P75/P90 came out we had 486DX4/100 which was only slightly slower.
It wasn't until the P200 came out that things really sped up. And then the first PII was a 233 Mhz, which was quite a bit faster, but still not as awe inspiring as the PII-450.
The Celeron 233 and 300 at first were called flops, then the 300A and 333A came out and suddenly their fantastic CPUs for a low price?
So the PIII-500 isn't much faster... Intel's anticipating 800 Mhz available later this year.
It'll sell. Maybe not the first version, but later revisions will become the standard, as always.
But first we need to find the CPU bug and fix it. Don't know what the bug is, but the chip has got to have one in there somewhere, which of course will generate all sorts of media hype!
Actually the big thing right now is frankly that the system bus, memory, harddrives and other I/O cannot keep up with the CPU.
You have to remember these benchmarks like WinBench 99 are measuring complete system performance. And frankly to be honest, most people don't need a PII-450 much less a PIII-800. However there are some applications out there which do heavy compuational analysis, and for them a PIII may very well be quite a big thing.
FYI, the company i work for does some dsp stuff, so we got a couple katmai (PIII) chips from intel to write for. Our algorithms run about 130% faster on the PIII than on the PII.
The comercials for this chip seem to point at it being better for web-based multimedia...
Eh? wasnt microsoft going to push some web based multimedia technology, called 'Chrome' - and then dropped it due to lack of interest? the p3`s only saving grace will come Later this year when the 820 mb chipset arrives and pushes the bus speed upto 133mhz and allows for r-dimms... until then, its a piece of junk.. a p2 'overdrive' for all intents and purposes... or maybe closer to a 166mhz pentium MMX, as compared to a straight 166mhz pentium. It is Dumb, in other words, for intel to be releasing and hyping this chip without the supporting technologies to increase a systems performance... they're releasing the p3 early, apparently because of amd, but without the 820 chipset al they are doing is embarasing themselves...
man is machine
It was like this when Pentium came, it was like this when PII came, why would it be any different now? In a few months Intel will provide faster PIII's and then all the powermad will buy them.
What I never could understand is how the last 10% is worth doubling the price for so many..
--
Pirkka
Assuming 100% accuracy of whatever frequency measurement equipment you used. Just because the computer itself says it's running at a particular speed doesn't necessarily make it so. All it has to work with are local oscillators and real-time clock chips. My cheap digital watch keeps time better than any PC I've ever worked with. Come to think of it, the 8088's did better than anything since.
I see even classic Slashdot is now pretty much unusable on dial up anymore.
My system: Celeron 463a, TNT-agp, 128meg PC100, abit bx6.
System built for friend: K6-2 350@400, TNT-agp, 64meg PC100, Asus P5A.
Both systems using nvidia detonator drivers & DX6.1 running win98 ( basically for games only).
Viper racing ( d3d game) internal benchmark @ 1024x768 on my celeron 463a(4.5x103) 50.3fps, K6-2 400 (4x100) 57fps. I don't play quake, so no nuumbers there. Unreal d3d is also a little faster on the K6-2 despite the slower clock rate.
Kindness is the language which the deaf can hear and the blind can see. - Mark Twain
Well, Pentiums are interesting chips (on the ix86 market), but the reason why you have to try hard to find a Pentium box to buy is that Intel stopped production.
So, if you only want Intel x86 chips (yes, some do), and Intel'd stop producing PII/Cel, what would you buy? yupp. A company with a "good name"'s way to sell their latest chips: Just stop selling the older ones.
And yes, this will all change while AMD/whatever's reputation raises. But, for example, I haven't seen even a single AMD tv spot here in Germany. Intel does. And geeks are a minority in PC buyers.
.sig: SEGV
The PPRO lived as long as it did because of it's scalability; not because of the performance of a single chip config. I don't see it as a failure. I see it as a real "workhorse" chip that held it's own long after faster single processors were available due to it's scalability in SMP configurations. Xeon changes that. Until Xeon, the fastest SMP machines were PPRO.
-Steve
Folks,
.25 micron process. Once it reaches the .18 micron process, expect speeds to go to 600, 700 or even 800 MHz--and it could happen by the end of 1999. And those 800 MHz PIII machines will be running Rambus DRAM, not your current PC100 SDRAM. Any operating system that can take advantage of that speed will really be hopping along. (^_^)
I remember when the Pentium II first came out--the speeds were only 233 and 266 MHz available commonly, with the 300 MHz part at over $1,000 per CPU! =:-O
Nowadays, you can get a Celeron 400 MHz CPU that runs rings around those early PII's for about $135.
Remember, the Pentium III CPU will arrive with the speeds of 450, 500 and 500 MHz--and that's using the
It'll be _very_ interesting to see what the AltiVec-enhanced G3's will be running at by the end of 1999--800 MHz? MacOS will be REALLY flying along at those speeds.
Raymond in Mountain View, CA
However, the massive bandwidth requirements of the Pentium III when it goes past 600 MHz and the speed of "Slot A" on the K7 _will_ require a lot faster RAM than today's PC100 SDRAM DIMM's. :-/
Hopefully, Rambus will have stamped out its problems by the middle of this year, so those really fast machines will have the RAM necessary to run them in the first place.
Raymond in Mountain View, CA
Lets all upgrade to the newest 95% hype products: Windows 98 and Pentium III's! My computer will "work better" and be oh so much more expensive and special - right.
Check out Tom's Hardware Page for benchmarks. The geometry stage of the pipeline is still handled by the CPU, which means that floating-point is important, which means that the K6-2 loses. The best solution for gaming seems to be an overclocked Celeron (though Intel isn't going to like that any more than it likes the K6-2).
In theory, KNI should be very useful for 3D games. However, It will rapidly become a non-issue as geometry acceleration moves to consumer graphics cards this summer.
And I was under the impression that 3D gaming was one of the primary drivers of the consumer computing market nowadays. You don't need a PII-450 to run spreadsheets.
Intel has almost certainly known this for a while, but couldn't switch to a new architecture without losing their market share. Now, however, they've found a solution that works in their grand tradition of throwing silicon at problems.
The Merced chip has been in development for a while. It has very significant problems, but it is designed to be a completely new architecture with a somewhat more sane design. Intel has decided to avoid breaking compatibility with older chips by giving it all of the old x86 operating modes, in which the chip cripples itself to act like it has the x86 register set, and emulates the x86 instruction set (not a big step, as most x86 compatible machines break x86 instructions down into RISCian micro-instructions already for easier scheduling). The cost for doing this is extra silicon (higher price and lower initial chip yields), but that's never stopped Intel before.
A cleanly designed x86 clone will still do a much better job of running x86 code, and a cleanly designed RISC processor without x86 support cruft will still do a better job of running software in general, but Intel should be able to pull off this switch pretty effectively.
New code will run more quickly than it could on an x86 machine, and old code will still run, so IMO people will probably spring for the Merced.
the Merced will most likely be more extensible than the x86, will perform better than_the_x86_,
Sorry, I should have included the caveat:
A Merced chip running software written for the Merced will perform a task more effectively than an x86 running software written for the x86. A Merced running software written for the x86 will run more slowly than an x86 running software for the x86, as stated in my other message. The point is that RISCian programs can be written for the Merced, while old programs can still run.
In the future, try reading my message before replying.
Tom of Tom's Hardware Page _did_ carefully benchmark several games on Celerons, K6-2s, and PIIs. The K6-2 loses by a fair margin. Go to http://www.tomshardware.com for the actual figures.
5-10%? That's nonesense. I don't know where you get your numbers, but an FP-intensive app that was optimized for KNI will get TONS more of a performance boost.
Current apps however will only have a marginal improvement, because they were not written specifically for KNI.
Isn't it getting a little silly to increase clock speeds by 50mhz? I mean we've got 300, 350, 400, 450, 500, 550... The percentage increase gets less and less each time. I hope they soon start jumping by 100. 600, 700, 800... That is enough of a difference to charge a premium for the higher end!
Maybe it's just me, but the Pentium Pro _was_ a flop. :)
According to my analyses, all chips in the Pentium family have performance that scales almost perfectly with clock speed. The "great Intel performance lie" is my name for any claim that a new Pentium has substantially better performance than an older one on the basis of architectural advances.
The PIII is not even the first Pentium to have lower performance per clock rate than its predecessor (if in fact it does). Readers may remember reports that the Pentium Pro ran 16-bit code slower than a Pentium, and all code slower than a Pentium MMX (clock rates being equal).
I maintain that this is the reason Intel generally doesn't make versions their older chips in the same clock rates as the new ones.
you couldn't be more correct. i've been saying this for years. i'm all for it, though, since it keeps costs down for the rest of us. ;)
http://www.sharkyextreme.com/hardware/weekly_cpu/
---- Dave
This will change with the release of G4/AltiVec.
AFAIK, AltiVec includes knew instructions that software will have to call specifically in order to take advantage of them. However, AltiVec also includes a whole new register that will be optimized for specific applications.
The evolution of the PowerPC has another thing going for it though... with the AIM alliance Apple will be optimizing MacOS (MacOS 8.x/9.x and MacOS X) from the ground up to take advantage of AltiVec.
Microsoft can't do this doesn't do this... and I'm not sure whether the Linux/Unix community has access to the Intel specs in order to do something similar.
Right, (dugh, for me :) ) of course. It makes sense that Intel would release those specs so that developers would use them.
I disagree on the optimization of MacOS though. At WWDCC 98 and at MacWorld SF this January, Apple made it clear that it would make changes to the internal workings (networking, display rendering, 3D etc.) of the MacOS (from an API and low-level standpoint) so that they (a) take advantage of the new instructions available and (b) take full advantage of the full 128bit AltiVec register and it's interaction with the FP and Integer registers.
I'm sure they'll be alot more clear on what this all means at WWDC 99 in May. Which should be within 6-10 months of the release of AltiVec based PowerMacs.
Just as a note...
As of MacOS 8.5 networking with TCP/IP (don't know 'bout AppleTalk) it can nearly saturate a 100Mbs connection in a mixed LAN.
I've tested it in our class LAN vs. MacOS 8.1, Windows 98, Windows NT, and FreeBSD.
One a 100Mbs connection (transferring a 40MB file) it is faster than MacOS 8.1, and Windows 98 (by *alot*)
It is about as fast as NT (sometimes slower, sometimes faster)
FreeBSD still beats it easily.
>If that not enough there is the Tracking number issue. How many people want to go naked on the Internet ?
The irony of this kind of cracked me up. Go to any search engine, type 'naked', and hit the button. The only useful metaphor I could draw from that is "look how many stupid people there are!"
Cygnus
"I feel like a quote out of context."
Just raise the taxes on crack.
The have Celeron for the low end. The PII's days are numbered. We might see a PII 500 (though I doubt it), but after than the clock bumps will all go into the PIII
Intel sells lower priced chips too. The Celeron's will go into the low end, these will go into the higher end and in time I bet this core finds its way into the eleron.
Whatever.
The PPro was kept alive far to long, primarily to protect various investments in the PPro server platform. Clock speeds stayed constant after the first few months and the costs of going over 2 processors were quite high. Increacing cache sizes did little to help.
The PII made a reasonable tradeoff. 1/2 speed cache + 2x the size let them ramp clock speeds at lower cost. The fact that Intel crippled the chip by limiting it to 2 way wasn't that big a deal for a lot of people since 4 Way PPro MBs carried a huge premium anyway.
Going further. The PPro was slower than offering by sun, MIPS, IBM & DEC when it was released, definitely at the end of the pack. Intel used to to build the ASCI red because a) they still wanted to be in the super computer business and b) they didn't have a faster chip to use. (The customer selected ASCI-RED despite that fact that it used a slower chip because the PPros low cost meant they could use more of them, which helped provide a competitive memory bandwidth)
As for your last point. At this late date the PPro makes a poor basis for an SMP system. Sure, you can go to 4-way, but it won't be any faster than the fastest 2-way PII (or PIII) for most code and it isn't likely to be any cheaper.
By Intels own admission, the Merceds x86 performance will not match a dedicated implementation.
Well, you could just get a Celeron instead, and spend the extra money on a VooDoo2. Or Two of them, hee hee. Of course it won't be faster, but the offset gained by the second VooDoo2 should be more than enough, for games. Or if you already have the 3d accelerator of your choice, get a 17" monitor. Or a 19" monitor. Gee, this is almost hedonistic.
When I see PIII I want to see Pill; Abortion PIII. Suicide PIII. PIIIbug. Hee hee, I'm giddy
AS
AS
-AS
*Pikachu*
And get paid to come up with stuff like "it's the difference between a Chevy and a Ferrari". Jeez...
all that I have to say is this AMD K7!!!!
PIII K7
100MHZ bus 200MHZ bus
64KB L1 128KB L1
512KB L2 up to 8MB
(Sorry not going (yes that say MB
passed 512) not KB)
Need I really go on
Intel is going to be in some serious
and I can't fucking wait!!!
I mean even try to overclock a Pentuim II or III for that matter. HA
Technology's a battle between companies producing more idiot-proof systems and nature producing bigger and better idiots
I work in sales for a company that stocks there computers with intel chips. So I will have the opportunity to go to and Intel train even next week. I will get to listen to hours of Intel telling the group how there product will do everything that you ever wanted it to and more for that matter. Now why might would I be going even though I am definantly not an Intel Fan. For on reason and one reason only. Most of the people that attend will be getting a PIII 500 on and an Intel board. And you know what I will do with this "Prize" if I win it. I will sell it to the first stupid asshole who wants to brag to his friends about being one of the first to have the PIII. And believe me people would pay quite the pretty penny for that. So I will sell it and continue on with my 400 K6 2 with two voodoo2 and a 32MB AGP card!!!
Technology's a battle between companies producing more idiot-proof systems and nature producing bigger and better idiots
Any one check out motorola's AltiVec Technolahy???
Must I say more?
Exactly!! Because most of the games out now don't take advantage of the super hardware... and who wants to go out and pay a crapload of money for the latest and greatest.. ( Ive done this before )
and go.. wow!! this runs fast on my machine... when games come out that support my machine... Ill be cool.. but by the time the games do come out... you spent WAY TO MUCH MONEY on your system.. where IF YOU WOULD HAVE WAITED... you could have PAID A LOT LESS>>> hehe
if you have a good video card.. TNT 16 or 32 megs... I think the k62 would kill the celeron.. in fact.. when I played games with my friends who had PII with all the bells and whistles.. game play was about the same... maybe a little faster load time.. but thats about it.. cmon... unless you are going to edit video... who neeeds that much power?? unless its cheap.. and then ill jump all over it...