Slashdot Mirror
Intel Announces Pentium 4
A friend of mine pointed me at this press release telling us about Intel's brand-new, shockingly original name (and logo) for the series of processors formerly code-named 'Willamette.' Meanwhile, I'll sit back and wait to see the logo parodies. Thanks to David Hageman.
Heh. I'm personally waiting for the Super Pro Turbo Pentium Hyper Alpha Tournament Winner's Platinum Edition. :-)
Even though the P4(Willamette) uses the same 0.18 micron technology as the P3's, it will feature separate internal arithmetic logic units that run at 3GHz, and a 400 MHz data bus and a 20 stage pipeline. IBM, also using the 0.18 micron technology for their experimental chip-the Interlocked Pipelined CMOS-have pushed it up to 4.5 GHz. In their design, they have several locally placed clocks that allow certain sections to run faster instead of waiting for the slower sections. This goes to show that the design is just as important as what and how much we put on it. Want to know more? Check out http://www.research.ibm. com/news/detail/fast_circuits.html
Standard I/O Error. Incompetent/Operator.
"Pentium 5" is kind of redundant in some weird way. I'm just wondering how much longer it's going to take us to get back to P6.
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
80486DX - Internal FPU
At this point, Intel became involved with several lawsuits because they didn't want AMD and Cyrix to relabel and resell their chips under the x86 names anymore, so they switched to:
Pentium (P5-4) - On-board cache
Pentium MMX (P5-5C) - MMX
Pentium Pro (P6) - On-die cache
Pentium II (P6; Deschutes, Mendocino) - On-card cache
Celeron (P6; Mendocino) - No L2 cache
Celeron-A (P6; Mendocino) - On-die cache
Pentium !!! (P6; Katmai) - On-card cache, KNI/SSE
Pentium !!! (P6; Coppermine) - On-die cache
Celeron II (P6; Coppermine) - On-die cache
Pentium 4 (P7; Williamette)
The primary differences between the original, deschutes, mendocino, and coppermine cores are:
1) Size of L1 cache
2) Size, speed, and location of L2 cache
3) Die layout
4) Packaging
5) x86 enhancements (MMX, SSE)
These changes ultimately resulted in:
1) Higher attainable clock speeds
2) Higher per-clock performance
Traditionally, a chip attains a new architecture identifier (ie, 486, 586) when the actual processing path changes. The Athlon was considered 786 material simply because they made massive improvements to the floating point unit, and because it utilized a completely new bus protocol (EV6 vs. GTL+). All of Intel's processors starting with the Pentium Pro up through the Pentium III Coppermine are considered 'P6' or '686' by many simply because it hasn't changed.
Take a Pentium Pro 200 and a Coppermine and do the following:
1) Downclock to 200, 66MHz FSB
2) Disable the L1 and L2 caches
3) Disable the x86 enchancements (MMX and SSE)
And although I am no engineer and I do not work for Intel, I can almost guarantee that both processors will give you the same performance.
If you try the same for any scenario, 386 vs. 486, Pentium II vs. Williamette (P-4), whatever, you will probably achieve entirely different performance marks. The Williamette, from what I've seen, is a completely revamped x86 architecture.
On the other hand, many people prefer to separate generations by per-clock performance, including cache changes and x86 extensions. The then you would have Pentium = P5, Pentium MMX/Pentium Pro/Pentium II/Pentium III (Katmai) = P6, Pentium III (Coppermine) = P7, and Pentium 4 = P8. The problem with this method is that it is open for interpretation. It's obvious to me that the Coppermine cannot be grouped with the original Pentium II, but Joe Q. Techhead may not agree with me.
Or we could take Intel's word for it (which is what they obviously want us to do) and believe that the Pentium, Pentium II, Pentium III, and Pentium 4 processors each have their own bevy of industry-dominating performance.
Enjoy the flamebait.
Alakaboo
IIRC, Intel named the 586 architecture the pentium because they wanted to differentiate their product from the 486 clones from cyrix and AMD. Unfortunately it isnt possible to trademark numbers, so Intel paid alot of money to a marketing firm to come up with "Pentium", which was easily trademarked as it was a completely new word. However, when intel was ready to launch the PII, they found that the names hexium,septium, octium, nonium(sp?),etc had all being copyrighted by a bunch of sly folk, in the hopes that intel would by the name from them. Unfortunately Intel took the easy way and decided to cop out with the pentium II,III and now 4 processors.
Dodge One
Dodge Two
Dodge Three
Dodge Delta
Dodge Delta 2
Dodge Delta 3
Dodge Delta Delta
Which gets me thinking.. "Pentium 5" is kind of redundant in some weird way. Maybe they'll call it Pentium Squared?
"Beware he who would deny you access to information, for in his heart he deems himself your master."
G4e will add 2 Alti-vec and 1 more FPU unit, and an extended pipeline. The single Alti-vec already whoop the vector processing capability of the Athlon, not to mention Pentiums...at 700Mhz, those things should easily beat out the 1Ghz K7/PIII in overall performance. I haven't read up all that much 'bout the P4s, but it is still IA-32! 'course the next question is when will the G4e hit the market...all the mac loyalists are hoping that it won't be too little too late...
livegoatpornium? hm.
seeing those words set off a random train of thought that involved goatse.cx, was heavily enhanced by large amounts of sleep deprevation, and ended with me idly asking myself the question: is pentium.cx taken?
I immediately realized it was a dumb question; why would anyone want pentium.cx and what would they _do_ with it?
I would have immediately forgotten it, except i suddenly realized something that made me wonder if maybe it isn't such a dumb question after all. Alright, think about this for a minute:
It's pretty clear that intel's naming scheme no longer directly relates to reality. Moreover, the scism is getting more extreme over time. From the pentium to the pentium 2 to the pentium 3 to the pentium 4, the differences in the chip have become less apparent in usefulness and much more arbitrary. From my totally ignorant perspective, it would appear that the pentium 3 was more a marketing construct than it was anything else; just a desperate attempt to stuff a _lot_ more not-very-useful-or-realistic complexity into the pentium instruction set so they could claim "look, we did something!" and have an excuse to run a lot of commercials, just to keep intel in everyone's mind as being cutting-edge, or something. I mean, look, it has something to do with the internet, doesn't it?? and they have cool 3d graphics and a looney toons character in the ad! It must be really advanced!!! d00d L337 1 \V1LL HAVE MY DAD R3PL4C3 MY 0V3RCL0CK3D AMD R16GHT 4W4Y with a PENTIUM 3 and it will MAKE THE INTERNET M0R3 FUN 4ND 1 \V1LL H4XOR BETTER!!!
The pentium 4 seems pure desperation, some extremely vague advancement just to pump the number up one more, just to release a lot of press releases and get people to buy stuff. Just to say, ok, the pentium 3 failed to change the world, but we're still here, and we're still alive and vital and moving!!
If i were going to be paranoid, i'd say the point of the pentium 4 is so intel can make a lot of noise about it to distract us from the IA-64. What about the IA-64, you ask? Say, that's what i want to know exactly.
But whether i'm being too harsh, and whether i know jack shit about microchips, i think i can say with some certainty that with each new "version" of the x86, the real _meaningful_ difference between each processor is getting a good deal smaller with each iteration, and the marketing aspects of a new product launch are overshadowing the technological aspects in intel's mind with each iteration. And i can DEFINATELY say the amount of time between "releases" is getting faster and faster.
My prediction:
From this point, with each "new" chip intel releases, the fluff value of the release will increase exponentially, the time between releases will decrease exponentially, and the justification for changing numbers will decrease exponentially. Eventually, intel will get to the point where they assign a different Pentium Number for each different clockrate assignment.
Thus about two years from now, Intel will have reached the point of the Pentium 110-- which they will name the Pentium CX [roman numerals!] -- and register the domain pentium.cx for it, to commemorate the Pentium CX's simultaneous release with "Windows ME harder"!
But who knows how much the tech industry will have changed by then? Hell, by that point, there may even be multiple-core G4s on the market.
(Score:0, Gibberish)
Irritable, left-wing and possibly humorous bumper stickers and t-shirts
I was doing some "field research" on firewall security when I stumbled across this from an Intel engineer. It got chopped off at the end in the process, but there's still some juicy stuff here.
---------------------------
To: Head Cheese of Intel
From: P4 project manager
Re: P4 chip layout
Message:
The die layout of the latest revision of the prototype P4 is in the attached JPEG. Here's the rundown of the labelled sections:
1. 11,290,491 transistors: CISC-to-RISC conversion. Handles a bunch of 1978 legacy bullshit.
2. I know this looks like just a dot on the chip, but it turned out the MMX stuff could be done with only eight transistors.
3. 12 Superpipelined floating point units: We designed them so that hopefully the next version of Excel will run tolerably, but we're not getting our hopes up.
4. Instruction scheduler: Arranges instructions in the fastest manner possible for those too wussy to use assembler.
5. We etched in a proposal from you to your secretary like you asked.
6. 453 transistors, 35 varicaps, four inductors and an op amp: I think this was put in while someone was drunk, but everything stops working if we take it out.
Without their cache, these modern chips amount to nothing. They are no good unless you have something filling the pipeline with the next instruction. :)
Take your k-spiff pentium III 950 and turn off cache. Boot windows. Have fun in your misery. THis is how I slow games enough to play them on my 450. Gabriel knight works just fine without cache.
Lowmag.net
Well, the problem is that Intel cannot go up from Pentium because they'll get Sextium... on the other hand, maybe it will make for sexier computers, as in: I named my computer Helga and she has a very fast sextium.
~~~Please pass the salt, I hate unsalted MD5s
how many pentium processors does it take to screw in a lightbulb?
exactly 1.8342210020334565635623561
but 2 is close enough.
all hail the thunderbird!
FluX
After 16 years, MTV has finally completed its deevolution into the shiny things network
"It is seldom that liberty of any kind is lost all at once." -David Hume
Any idea if Intel is EVER going to change the name? I doubt it. "Pentium" is such a household name, and whenever it is mentioned, the average person recognizes it, and associates it with a fast computer. Intel has succeeded in making sure EVERYONE knows its product name..
/. is a commercial entity. goto slashdot.com
Now I will finally be able to afford a Pentium II!
Washington, DC: It's like Hollywood for ugly people.
Well, they couldn't go for the old one, because Pentium IV sounds too much like they're selling silicon crack :)
Si
Coming soon - pyrogyra
BTW the correct link is her e.
"But actually trying to use m4 as a general-purpose langage would be deeply perverse" --ESR
Wow, looks like Intel finally released the P4 BBQ model in time for fourth of July. Put them ribs on the heatsink, baby, slather them in head-conductive grease, and we're cookin', baby!
If they sell enough of these we'll start hearing stories about "Intel brownouts"...
just my blog and pix
I have C comment blocks longer (and more informative) then that press release.
Do you have Linux and a DotPal? Click here now!
"Oh, I hope he doesn't give us halyatchkies," said Heinrich.
Intel is on the decline just like Microsoft started to a few years back. Though Intel is nowhere near the state MS is in, it seems these huge product releases are made out of desperation. AMD has taken a huge chunk of Intel's market, but who knows, maybe they might be able to make a comeback?
Buying a Dell computer is equivalent to dropping the soap in a prison shower.
Actually, the Pentium 4 is a very drastic change from the P3. It has an entirely new core, with some interesting new ideas (trace cache, "double pumped" ALUs). However, whether or not it is any good remains to be seen. Your comment that it is just a marketing exercise may very well prove to be true anyway.
You see, the P4 has a 20 stage pipeline. Now that's a lot compared to most chips, meaning that it will take a huge penalty for a branch misprediction. What's the advantage of a 20 stage pipeline? Clockspeed. The P4 was designed first and foremost for clockspeed, because that's all the clueless average computer user looks at. There's a very good possibility that the P4 will perform worse at the same clockspeed than the P3, but it will reach some insane clockspeeds. This is especially true in floating-point operations. The P4 only has one FPU while even the P3 has two (the Athlon has three). That just doesn't cut the mustard these days.
Of course, intel is banking on SSE2 to make up for their pathetic FPU, but that has to be specifically supported in the application. Anything FPU intensive and without SSE2 support will perform much better on a P3 than a P4 of the same clockspeed. Of course, intel will base all their bencharks on the miniscule number of applications that will support SSE2 (and pressure third-party benchmark makers to include SSE2 support as well), so they'll have a bunch or pretty graphs showing that the P4 is super-duper fast.
I am not an idiot. Please use my name to email me.
"That's right, I'm quoting myself."
-Upsilon
don't forget, the only reason we got pentium was that intel added 486+100 and got 585.99999999993646872
9) Whupamdassium
8) Fnordium
7) Pengium (Thus really cementing the break with Microsoft)
6) Really obscenely fast processor
5) Notcreativeenoughium
4) 886
3) Just another damn IA32 chip
2) Killappleium
And the number one other considered name for the Pentium 4:
1) Livegoatpornium
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
Did they learn nothing from the Rocky series?
-={(.Y.)}=-
Thank you for reading One Man's Opinion. No participation necessary. Offer void where deemed by law or PATRIOT Act.
How much did Intel pay the brand consultants to come up with "Pentium 4"? Those guys literally get paid millions to come up with things like the Lucent coffee stain logo, the AT&T Death Star, the name Agilent (sounds like a washing machine company,) etc.
Meldroc, Waster of Electrons
Remember years ago when we said 'oh, someday we'll have some really fast Intel 986 chip with tons of memory and stuff...' ... well, that day is about to happen. If Intel had continued its old naming (numbering) convention, this would have been the 986.
--
Tired of FB/Google censorship? Visit UNCENSORED!
"Around the world, PC users associate the Pentium brand with the highest PC performance, compatibility and quality available."
Oh yeah, my Pentium 120 just screams.
/* The beatings will continue until morale improves. */
Still, I'm glad they called it that. It beats Celeron, Athelon, and Duron.
Rumors that Intel might be working on the Intel Oxgen, the Urnium, or the Lutonium processors were vigorously denied by the company.
"Intel will also be offering it's new, jointly developed, industrial strength power supply for ISPs using large scale P4 systems"
"Earlier today, seven intel engineers were incinertated in a systems-test accident. The incident occurred on power-up of the test-bed of Intel's new quad-P4 board, codenamed Phoenix. Witnesses describe the cause as 'spontaneous combustion'. 'The damn thing just blew', said one engineer, 'and then everything was just a ball of fire! I'm sure glad I got out of there alive... Sources report that Intel management has reportedly been talking to executives at Frigidaire regarding the incident."
just my blog and pix
> I guess the can't call it "Sextium", although
> it might boost sales...
Sorry, you're mixing Latin and Greek. The
successor to the "Pentium" would logically be
the "Hexium"--which admittedly isn't as funny but
does have its own possiblities. The "Sextium"
would have to be the successor to the "Quintium".
Chris Mattern
Most people not only associate "Pentium" with "fast processor", they associate "Pentium" with "any processor".
True story: I was in my local Drat Shack (which also has a section devoted to amateur radio gear and other "stuff" on consignment) and this kid (who, for the record, seemed to believe "soap" was a four letter word) comes in. He starts looking at this old Pentium 75 that is stuck into a block of plain old styrofoam (can you say ESD? I thought you could). "Ouuuuh, would this work in my computer? I have a pentium-486." As you can tell, it went downhill from there. Fast.
Intel has poured gigadollars into making this the case: people ask what kind of Pentium is in their Macs, or in my Indy, or whatnot. Intel's execs would sooner chew off their own testicles than change that name.
www.eFax.com are spammers
Well, its entirely obvious to me that Intel has adopted the Street Fighter numbering system. In three or four years, we can look forward to Pentium Ex Alpha Plus.
--
It also stands a good chance of kicking the Alpha 21264's A**! .18 later this year and should top 1 ghz. That will ensure a comfortable fp edge over everyone. On the integer front it should be close, but I suspect it'll be the alpha's win here too. The williamette core is too deeply pipelined and has too much of a branch mispredict penalty.
Kicking it at what? Alphas are moving
--Shoeboy
Does no one pay attention to uname(1) anymore? :)
The Pentium was the 586; The Pentium Pro was the 686; The Pentium II and III, being merely modifications of the PPro, are also 686. Originally, the Merced was to be the 786, but it looks like they'll be going a whole new route with that, and Willamette (the Pentium 4), being a new chip, will take the 786 label.