Slashdot Mirror
Intel's Tualatin P3
DavoKid writes: "Intel rolls out the .13 micron Pentium III processor based on the Tualatin core at 1.2GHz. This chip really shines and overclocks to 1.47GHz. The benchmarks are fairly impressive too!
Reviews at: HotHardware, Anandtech, and Tech Report." Also given plenty of attention is Intel's new D815EEA2 motherboard, since that appears to be about the only choice for the new chip. The consensus seems to be that this chip is at least intended to be "the new Celeron," but marketed also as a power-frugal chip to impress server-farms with electrical savings.
I just tooke a long hard look at the HUGE heatsink that runs up the middle of my G4 Cube. I also noticed that this particular heatsink did not have a fan glued to it.
Oh, and if you want hard numbers, go read up the distributed.net stats. Granted, they are a mathmatically narrow way of looking at it, but they certainly show how low MHz PPC chips do very will against high MHz x86 chips.
Also, congratulations on your purchase of a Dell computer. You get what you pay for.
Don't worry, just give Windows a few more iterations and it'll require a 10GHz chip.
-M
Intel names their processors after rivers. The ones designed in the Hillsboro, Oregon facility are named after Oregon Rivers; Hence McKenzie, Willamette, and Tualitin. Too-All-i-tin. Ironically, the Tualitin is known as a particularly slow river.
The other day, I noticed that Dell had started offering a 1.1GHz PIII option on their Precision 4100. Is this one of the Tualatins?
(BTW, my box at home is still a 550 MHz PIII Katmai, the last one produced on 0.25 micron. I have major process-shrink envy.)
Crap... I've been referring to it as the 'Guffaw'. No wonder my Mac-loving friends have been so snippy lately.
I always assumed higher clockspeeds would eventually give way to lower power and slower speeds in future chips. First to go was SMP. Now it's clockspeed.
Obviously at 1.2 Ghz Intel has stepped back from the speed deamon days of 1.4Ghz P4's and started retooling for lower power consumption and portability. Those of us who want to crunch numbers will have to spend big bucks for mainframish chips with weird names and companies with 3 letter names.
what's your favorite CPU-intensive activity?
(with all due respect to raster and mandrake)
enlightenment and gtk+ pixmap themes....
If I don't put anything here, will anyone recognize me anymore?
Vector graphics for 32bit-600dpi colour displays, fewer compromises in voice recognition, some of those cumbersome comp-sci alorithms to deal with stuff like process starvation, richer GUI programming environments, condensing server clusters into single machines, real-time dvd-quality video compression, games games and more games...
Most importantly perhaps for the short term, lower energy comsumption in Notebooks, fanless silent desktops, stuff like that.
Finally it can be used for the ultimate user interface... specialized devices which are cheap, disposable, portable and secure. The high end of technology pulls the low end with it.
Since when does a sugarpill induce vomiting, nausea, headaches in 0.0001% of patients studied?
Having both a degree in Psychology, and working in an IT job, I'd bet that it's more like 1% of patients reporting a lot of this stuff. My grandmother swore that Sudafed made her drowsy, never mind that it's a stimulant. People always like to blame their meds for causing some problem with them. They'll report that the test drug caused diahhrea, and forget to mention that they ate four pounds of greasy Mexican food the night before.
Kind of like the users who swear that installing the latest version of Netscape caused their printer to break.
In short, you'd be surprised at the large large number of people who don't or can't grasp the concept that correlation doesn't equal causation.
---
When in danger or in doubt, run in circles, scream and shout. --Robert A. Heinlein
Every word in that sentence is merely a gross misspelling of the word tomato.
This comment is guaranteed*
*not guaranteed
I find it funny how intel marketting and pr driods can say the names for their processors and keep a straight face. Also the amazing creativity in naming their processors once released is astounding. At this point if Intel released a processor that wasn't named **pentium**** people would probably think it was made by AMD.
I have a feeling that the G4 at 1.2 Ghz would be putting out more than 21W.
The Tualatin, just like the Willamette, and the Columbia, and the Deschutes, and some more I can't think of, are in fact, rivers.
The clever practice of naming chips after rivers has been revealed by someone who's fallen into the Tualatin River and survived without mutation.
And it's pronounced 'too-wall-ah-tin'. There's a city. With a crawfish festival.
MHz != real life speed
The 750MHz PPC is much faster than a Pentium III at 750MHz. As far as I know the 750MHz PPC is about equal to ~ 1.1GHz PIII in most operations.
There is of course faster x86s than PIII-1.1GHz, but it is definitly not the slowest x86.
two-AH-lah-tin
It's a river in Oregon.
thursday
Apple's really pushing the whole 'megahertz myth' concept as a way to reassure Mac buyers that an 800Mhz G4 is just as good as these 1.xGhz monsters from AMD/Intel. They had the full dog-and-pony-show including animated slides at MacWorld a couple of weeks back.
What's the consensus around here about such things? I get the impression that the AMD/Intel competition keeps the x86 arch on the cheaper end that PPC for obvious reasons.
It's kind of annoying that it is hard to know what to trust and how to benchmark different chips. Overall, the CPU only makes up a portion of the package when you buy a PC (RAM, HDD, Optical disk, video, etc being others), but it has traditionally had mindshare of being important beyond such other things. And people tend to only look at clockspeed.
What's the best way to compare these things?
----------------------------
Sure, but with a mass-produced hardware encoder you could do the same with a 100Mhz part.
Consider how much slower a general-purpose CPU like the P3 is compared to custom hardware like the NVidia GeForce3 for doing realtime 3D. You probably would neet a 10GHz or better P3 to equal the performance of the GeForce3 at this task.
With the increasing popularity of video-processing on consumer desktops, it would be nice to see hardware manufacturers putting some hardware into their cards to support encoding functionality, as well as just augmenting software decoding.
Realtime MPEG-4 encoding is not out of the question, since realtime MPEG-2 encoding is now a consumer-level proposition - TiVO, cards from hauppage etc..
It would be interesting to see Matrox take back some market share by building a programmable video compression engine onto one of it's upcoming cards.
They have tried this with the 'Rainbow Runner' and it's ilk, but these products were never billed as a 'Complete PVR and DVD-ripping station', which i'm sure would be vastly more attractive to joe average than 'Record and edit your own home videos'
This is not to say that a 10GHz CPU would not be nifty, but rather you could get more done with a set of lower-clocked chips, each optimised for specific functions.
I gots ta ding a ding dang my dang a long ling long
Damn. I hope these ones won't be as buggy as the 810/815. I've tried a few ones, and all I got was random halts on my penguin OS....
BTW I would like to know if any of you ran memtest86 on i810/815 boards... I did on several ones and I got some pretty nice errors... so I fall back to my oldie PII. At least this one works 7/24.
Álvaro Lopes
Indeed this looks like an Intel something...
How is this flamebait? Seems like quite a good summery of the respective respectibilities of the two companies!
A deep unwavering belief is a sure sign you're missing something...
Actually, NVIDIA has been consistantly making quality hardware for years now. Just as you don't expect a reliable company like Intel to produce buggy chipsets, you don't expect NVIDIA to produce buggy products either. It sometimes does happen (Pentium FDIV and i820) but not often enough to get cynical about.
A deep unwavering belief is a sure sign you're missing something...
Really? Empirically, NVIDIA's graphics drivers are the best out there, quality-wise. They've never frozen on my machine (the GF2/3s might be different) and they have by far the best OpenGL support of any consumer cards.
A deep unwavering belief is a sure sign you're missing something...
Linux or Windows 2000? Because there are SMP problems on Linux, but I haven't heard of any on Windows.
A deep unwavering belief is a sure sign you're missing something...
Yep! And I'm still wondering why the school decided to upgrade the 486 machines in my lab to PIII 800's (the data streams in at one rate and the IBM Aptiva's handled it perfectly).. It all comes down to Marketing! If you're going to show off the lab to perspective students, you need new hardware.. Dumb!
Intel isn't the only guys to have motherboards out for this chip, currently Asus has the TUSL2 and I believe that Abit is putting out a motherboard shortly.
How do you pronounce this? Tualatin? Towel a Tin?
Intel is getting ridiculous. At least i can pronounce my processor. Power PC G4 "GEEE FOUR" Easy =)
--------========+++Dont Feed The Lab Techs+++========--------
The result is that the new CPU compares poorly to the Athlon CPU, which processes three FPU instructions per CPU cycle compared one instruction per CPU cycle on PIII's.
P4 also only does 2 FP ops/second but outperforms K7 by about 50% on SPECfp. FP performance typically has more to do with memory bandwidth than with the speed of the FP unit.
It's better than having a "chip" on your shoulder.
Friends don't help friends install M$ junk.
does anyone else read that article the first time and see 12ghz? its obviously a simple typo and easily forgivable but I damn near crapped my pants before reality caught up with me.
-
i compiled xfree4 yesterday - 10 hours (and it didn't work :( on my p133 laptop.
:)
:)
.oO0Oo.
a 10Ghz cpu would really help voice & handwriting recognition.
besides other intensive apps like compiling, editing video (more in demand than you might think) faster processors mean you can revise your approach to your problems.
Remember when OS's and a lot of apps (such as Windows 3) were written mainly in assembler?
As machines get faster we are able to increase the complexity of the systems we use. As usual it is a double edged sword. There does come a time when we should throw out all of our slow cpu based software and start again. The free unixes are fabulous and I use them every day but I also want them to die. I'm a plan9 user and fan (I had to mention it in somewhere
It does amuse me, however, that I use faster and faster computers financed mostly by editing plain text with vi
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
At some point the die size is gonna stop shrinking and we are going to have to start going multi processor. Any guesses on when this will start happening?
It would be kind of interesting to see some estimations on current mix of assembly instructions by category, for "typical set of applications" (defined in some reasonably sane sense). Wouldn't be too surprised if dudes at Intel had used such estimations on deciding to use an outdated FPU, actually.
--
I like paying taxes. With them I buy civilization -- Oliver Wendell Holmes
"Athlon T-bird become the value market chip once Athlon 4's hit the streets?"
I'm just making an educated guess here, but I would strongly suspect that any 'Duron 4' would be a reduced cache version of the Athlon 4, just as the Celeron and Duron were cache-inhibited versions of their siblings. As well as this, I'm sure AMD want to maximise the number of chips with their new instructions on as well; the enhanced performance of Athlon 4, just like the Pentium 4 relies to some extent on those new features being exploited.
"I Know You Are But What Am I?"
beats the Athlon4 the old athlon wasn't called the athlon3 (I know it was the third evolution of that chip, so what). Which is blatantly trying to have the same name as the P4.
I do.
We saw the same thing with the celery FSB clock speed.
Intel is always extremely conserned with it's lowend processors competing with its top of the line. Check out the Tom's Hardware review about 2 months ago when they got their hands on a high power Tulatin. Those marketing fools just figured they could kill 2 birds with 1 stone by lowering power (which hurts performance) and calling it "mobile" or "low power".
The only thing you can do, is try and look at a variety of benchmarks, that put the chips on equal playing ground.
Make sure benchmarks or apps have optimized code for the processor they are running on, and make sure the setup is the same for both processors.
Much of the code that runs on the P4 isn't optimized for it, the optimized stuff runs much faster. Apple's G4 can perform really well on certain benchmarks but they never show the ones that suffer.
Anyway it is a difficult process to say the least. Thats why Tom writes entire articles comparing chips for a living.
When someone created a "red-alert" type game in our assembly class. Real-time war, lots of units on the screen, 100 or so. The new p3 - 900mhz processors in the lab handled it no problem, however it might have crashed or run really slowly on an old 486. But anyway your right, and the argument still stands. Gaming is about the only thing that might require speed.
Don't feed the trolls!
About motherboards: Do you know who is making the new MP chipset? AMD themselves--and according to all reports, it rocks. And it's not like Intel hasn't had its share of problems with chipsets. Sure, Athlon+VIA will always probably suck, but for the same amount of money as a P4+MB you'll always be able to buy a faster Athlon with a better MB.
I personally think the P4 is Intel's curse, and bad news for them is far from over. I would be biting my fingernails if the earnings of my company depended entirely on something as shoddy as the P4 design.
Anyone who's checked pricewatch would have to be insane to buy a P4.
You do know that the Tualitin uses the same P6 core that debuted on the Pentium Pro, and is clock for clock, faster than the P4 which is the one that has the utterly stupid lengthened pipelines which allow it to ramp up to such high clock speeds, right?
The PPC chip is also as fast as the slowest x86. Wow.
I love idealists not because I am one, but because they make life bearable for pragmatists such as myself.
You "assume" that it will be quality chipset. What evidence do you have to back this up? Personally Im a HUGE fan of Nvidia but for you to simply ASSUME that their first venture (ahem.... well I guess you could count the xbox) into the field of integrated chipsets will be shining with quality is proposterous speculation to say the least.
As far as the NForce goes - I say wait and see what happens. Its bit pricy for an intergated chipset so it will be interesting to gauge the market acceptance for such a product.
If anything the Nforce has brought us Dual Channel DDR Ram - perhaps now DDR will be able to fully compete with Rambus memory (Say what you will but Rambus is clearly superior to its competitors - submissions proving otherwise will be considered.)
Oh yeah one more thing - the Nforce will force the fools at Creative to get off their collective asses and produce something New for change. Afterall the sound portion of the Nforce makes the SBLive look like crap from what I understand.
Gam
"Flame at WIll"
I love idealists not because I am one, but because they make life bearable for pragmatists such as myself.
Yeah the 6.49/6.5 drivers seem to provide the best reliability. 12.x may be faster but there are a few games which dont work with them.
Gam
I love idealists not because I am one, but because they make life bearable for pragmatists such as myself.
IBM was working on this for it's Power line (not PowerPC). Dammit, can't find any links to anything concrete.
Most people look at a chip primarily by the clockspeed--higher MHz sells better.
Where are we going and why am I in this handbasket?
The problem is that there are so many varients of the P3 and Athlon that the outside world still keeps these names to tell the difference.
Admittedly as clock speeds increase it's obvious what version it is (i.e. there is only one 1.3 GHz Athlon - for now). But oftentimes the overlap requires those who follow the industry to mention the development name so they know exactly which one they are talking about.
Right now you can call up most computer stores and say, "I want a 1.3 GHz Thunderbird." and they will know exactly what you want.
Until AMD and Intel come up with distinct names for each chip generation, the internal names will be known and used by people in the hardware business.
While Athalon's may be fast they are not always stable and can run exceptionally hot. This aside I would still buy an AMD chip over an Intel at the moment, (just based on price alone, never mind performance.) Yet some don't necessarily trust AMD, be it Intel's propaganda machine, or consumer distrust after such great processor manufacturers like Cyrix. To this day I know people who prefer Intel's exclusively (not including the P4's) the Intel line has always been reliable (okay ignoring the Pentium bug as well.) So it almost goes back to the old fable of the tortoise and the hare, while AMD might be a superior processor speed wise, Intel has maintained stability. While AMD completes for market share with low prices, Intel plods along with new technology a little slower, and a lot more expensive.
This won't be the same forever, as Intel brings its price down, and now that AMD has (in my mind) more then established itself as a solid competitor we should soon see a all out pitched battle for supremacy that should in the end have a net benefit for the consumer. (Well here's hoping at least)
Geoffrey Cameron Peart
McMaster Software Engineering
Geoffrey Cameron Peart
McMaster Software Engineering
Monkies? I like Monkies
Judging from reading the Anandtech review of the new 1,200 MHz Pentium III CPU, I think the problem is that the CPU--while it is very fast indeed--still sports the older-style FPU unit. The result is that the new CPU compares poorly to the Athlon CPU, which processes three FPU instructions per CPU cycle compared one instruction per CPU cycle on PIII's.
The poor FPU performance is why I don't think there will be much interest in the new CPU, especially since the 1,200 MHz Athlon CPU will substantially out-perform the new PIII CPU with any application that is FPU-intensive such as CAD and illustration programs.
I haven't had any experience with a TiVO, but a lot of consumer-level (and that's all I can afford) MPEG-2 encoding hardware for PC produces truly lousy quality video compared to what the software encoders can do.
Still, I'm sure you're right, and I have faith that the market for this sort of thing will improve. But I was really just trying to make a point about the old "we don't need more speed" argument...
You ALWAYS need more speed. If someone can't figure out what to do with more CPU power, they don't have much imagination.
STOP . AMERICA . NOW
Encoding DVDs.
With my 1GHz Athlon, it takes ~10 hours to encode 1 hour of MPEG-2 DVD video (720x480 MPEG-2 7M-9M/s).
A 10GHz CPU with linear scaling from my current 1GHz CPU would reduce the time from about 10 hours to about 1 hour. So, if I could just get a 100GHz CPU, I would finally be able to encode 2 hours of MJPEG DV or VHS capture in just about 12 minutes.
That would be nice.
STOP . AMERICA . NOW
Anyhow, anyone who buys the P3 will probably want it because it's cheaper, it works in their old P3 socket, it uses less power, etc. It could replace the Celeron line temporarily.
If it does have the old FPU (and we don't really know that it does) then so what? It's a shrink. And it's an EOL product. Intel has said that this will likely be the last desktop rev of the P3. I wouldn't expect them to spend the development costs to design and integrate a new FPU into a product that will not continue to evolve. That would be insane.
Intersting name, it has to have been named after the sound it made when throw across the room.... As there have been so many problems in their release they had to have thrown a bunch of them away.. (actual recording from Intel devlopment labs) "Ralph another one did not pass" " Well Joe, toss it" *Finggggg* *Tulana* *Tin* *chink* *Tink* *tink* *plink* "Now Joe, I did not say "name it", I said toss it on the pile"
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Some people are alive, only because it is against the law to Kill them!
Gee, I'd love to have one of those on my desktop... no idea what I'd use it for, but I'd still love to have one!
MadCow.
I used to have a sig, but I set it free and it never came back.
Just-dreaming-ly-yours,
Madcow.
I used to have a sig, but I set it free and it never came back.
Oh come on... this guy is lying... don't any of you know what VLIW is? AMD doesn't have any VLIW offerings... The Intel Itanium is basically a VLIW processor, and we all see how well that worked out...
feh
I must burn in hell, suffer and pay for my sins
But Gods the one who's losing, Satan always wins!
If you can point me to some documentation of this fact I'd be grateful then...
I must burn in hell, suffer and pay for my sins
But Gods the one who's losing, Satan always wins!
How about actually having a chance at winning the RSA contest posted last week ;-)
NVIDIA's drivers have been giving me random opengl lockup problems on two different dual proc boxes. After some usenet searching it seems to be a problem with their drivers and dual proc machines.
I have had problems in both.
Cheers
Dwain SnydersResearch and Development, AMD
2DUP * ;
At AMD, the latest generation of chips are currently being designed with flow-through core transistors, so it'll really be more like a "smart capacitor" than a integrated circuit, like most CPUs. I have been tasked with writing the VLIW compiler for the new chips, and I can tell you that they really do fly, and use less power than teh traditional Athlon/Duron series, while retaining the power and in fact doing a lot of optimisation thanks to the new VLIW instructions that are being ingrained into the core.
All I can say is, folks, look out for this one. It will be hot. (but not because of excessive power consumptionDwain Snyders
Research and Development, AMD2DUP * ;
Don't feel bad... my "chip" is only 4"
Oh God... you're all talking about microprocessors... I thought you were talking about... I'm so embarrassed...
- For the complete works of Shakespeare: cat
"This new Pentium III, code-named Tualatin during its development"
Did you see that? It says 'code named', meaning it's not neccesarily going to ship under that name. Read the article...
Remember the AMD Duron's code name was 'Spitfire'. You don't need me to tell you it didn't ship under that name.
Sham on
Build a beowulf cluster, what else!
badness 10000
The reason Intel won't attach 166 or 200 MHz system bus to P!!! is this time is that it will destroy the flagship PIV if Intel does that. P!!! can benefit from higher bus speeds cuz' it has 2 SIMD units that can process up to 8 simultaneous ops if 2 instructions were decoded, no matter its slow data moves, units' latency are smaller than Will. P!!! is now really Celery of family, efficient but cut down. Intel could add P!!! better FSB for data prefetch when PIV achieve 3 GHz and quad-pumped 133 MHz system bus. "Athlon + DDR is mutha-fucker, if you optimize your software for them"
SOI (Silicon on Insulator), .13 process, crafty enginners, scrappers from AMD doing all for Athlon to put Intel in shame.
Of course AMD will have no problem ramping up stable Athlons to higher GHz.
AMD 760 isn't a good chipset for all of you?
nvidia to my knowledge implemented tile rendering technology in GeFORCE chipsets, but doesn't works with hardware T&L, this is NOT a bug? Intel also intended to launch an OVERCLOCKED 1.13 GHz into market. Homewer, Winfast GeFORCE 2 MX cards ships overclocked and Hercules GeFORCE 2 MX400 has the memory OC'ED, but we doesn't consider this as a bug, right? Explain to me what happens to 733 MHz (yes, those have been shipped at default clockspeeds) Coppermines and its "strange being" with certain benchmarks? And finally, if Intel chipsets have bugs or haven't, this isn't a problem. Unlike VIA or ALi chipsets they have good default drivers from Microsoft and Intel is a good supporter (well, when Intel intends to make the hardware offers of competition appear to be buggy or crappy, when the problem is mainly software or firmware) to the Linux community.
As long as I have to wait for my machine to render, I can always use more power. When I can 3D model in realtime with raytracing and shadows, etc - then we can start having this discussion. Of course - I can't wait for the next time I step into BestBuy and hear the salesman tellin the 75 year old guy who just wants to be able to read the email his granddaughter wants to send him that "oh yeah - you definitely want the power of that 1.7 gHz machine as opposed to that pokey little 1.2 gHz box. it'll make your net connection a lot faster." two sides to every arguement.
anything i tell you will cloud your opinion.
Has anyone profiled the efficiency of these cache-inhibited chips against their "real" bretheren? I learned in computer pre-school (university) that the processor's efficiency increases dramatically as cache is added (256KB=fair; 512KB=good; 1MB=quite good; 4MB=very very good) even for low MHz processors. Just how much of the "value" processor's time is completely wasted due to pipeline flushes, etc.? Are "value" customers taken to be ignorant drool-buckets, or are they really getting better price/performance?
Healthcare article at Kuro5hin
It's great to have speed, but it's even better if you actually have stuff you do that uses all of it; what's your favorite CPU-intensive activity? Decryption? Creating 3D movies? AI? Computing the entire Othello game tree? Playing hand-coded-assembly Pong at a blinding pace? Or has clock speed finally surpassed practical use?
Alsea, Butte Creek, Calapooia, Clackamas, Crooked, Metolius, Molalla, Owyhee, Siletz, Umpqua
http://www.awa.org/awa/river_project/Oregon/