Slashdot Mirror
Pentium 4 And Brookdale Update
ravedaddy writes: "With the Pentium 4 in mail order stores now (before Intel's release date), [Sharky Extreme] felt it was time to give an update on the status of Intel's next generation chip as well as a look at some more information on Intel's upcoming SDR and DDR chipsets (Brookdale) for the Pentium 4." Key words: "Don't be foolish and buy now. You can't actually buy a Pentium 4 motherboard yet, so you won't be able to use a Pentium 4 right away, anyway."
I've been told performance isn't related to size.
This is not targetted specifically at home users... This is supposed to be the flagship product of Intel...
I will guarantee, though, that the benchmarks that will be run will not take this into mind. You will see Quake3, Incoming, Winbench, and the like used to see what the processor will do. Great. Fine. Here we have software that is either bound by the bottle neck of the GPU, memory bus, or some other technology that is holding things back. Even those programs you mention will be limited in some way regarding these technologies. Bah, I say.
Bryan R.
Bryan R.
The price of freedom is eternal vigilance, or $12.50 as seen on eBay.....
> Obviously, the pentium 4's performance lags (or will) behind Thunderbirds of similar clock speeds. But this thing is also a .18 die process.
.18 micron process. This means that an individual gate measures 0.18 micron wide (or is it long? Damn...where's my VLSI text). Now, the smaller you make each individual gate, the more gates you can pack into the same square area of silicon. Or, if you keep the same design with the same number of transistors and "shrink" it to a smaller process, the same part can be manufactured in a smaller die size (square chunk of silicon).
You are being a bit loose with terms. The Pentium 4 is fabricated with a
Cost for semiconductor manufacturing is primarily a factor of silicon area, so reducing the die size is a big cost-savings per part. Also, smaller die sizes increase yields. Defects are inevitable, but with careful fab management, you reduce defects to a certain number per wafer. Each one of those will (likely) ruin a die. But if the dies are smaller, less of the overall wafer's area is lost to that particular defect. Thus, the smaller your die size, the less any given defect will cost you.
But there's another very important advantage you get from a smaller process. A smaller process generally translates to a higher clock speed. Why? Gate capacitance. Each gate is essentially a capacitor that has to be charged or discharged to switch the gate on->off or off->on. A larger device has a greater capacitance and hence takes a longer time to charge/discharge. So, larger devices take longer to change state, and hence can not be clocked as fast.
So, in summary, a process-shrink should improve yield for Intel, which means lower cost per part. It should also help their engineers up clock speeds on the component. However, at any *particular* clock speed, performance will not be affected. Heat dissipation / power consumption should be reduced, but otherwise, clock for clock the consumer will not notice a difference between processes.
...anyway, that's some of the engineering behind this. In terms of forecasts, I think Intel has been caught with it's pants down. They have an inferior product, and, if the world is sane, AMD will clean their clock in the coming year.
--Lenny
I'm not worried about that. My Celery 300A system will be in service until 2002. The motherboard only supports CPU's up to 550 MHZ. When I bought the thing the fastest Celery processor was 433 MHZ.
I have no high hopes that I will be able to put a 2 GHZ processor in that board, so I guess that it was an evolutionary dead end.
I have never upgraded the processor on any of my computers without switching the motherboard as well. I just run them too long for that. I typically run a computer for at least 4 years, and my current machine will go for 5 years.
My next machine will be 1000 times faster than the 386SX that I used from 1990 to 1994, so I expect that it will serve me even longer.
If tits were wings it'd be flying around.
Hey 1.4Ghz means that the buttons on Java GUIs will go up and down almost in *real-time*. Wow. those wierdos in the Java development camp will be dancing in the air!
First of all, I am a software guy and I took only as many hardware classes as were required, so consider most of the following to be phrased in the form of a question (as in "Does this make any sense at all?"). That said:
It seems the speed of light must be becoming a significant factor here. At 1 GHz, light in a vacuum will only propagate about 30 cm during a clock cycle (3*10^8 m/s / 10^9 cycles/s = 3*10^-1 m/cycle), and at 5 GHz that drops to 6 cm. I don't know the speed of electrical impulses in silicon, but even if it's not much less, that means a signal can't be doing too many laps back and forth across the chip (up and down the data path, etc.), i.e., you're working in a time scale where you can see the clock pulses rippling across the chip. Then no matter how fast the gates themselves are switching, the number of them that the signal can go through is limited by the time it takes for it to travel the total distance.
When the speed-of-light propagation time from one end of the chip to the other is a measurably large percentage of the clock period, that would seem to make for some incredibly funky new design problems, due to clock events having occurred in one place but not another, signals that take different paths not reaching the same place at the same time, etc. Or if you just wait for everything to catch up between steps, that would cut down the number of steps per cycle even more. This would be another reason why smaller feature size allows higher clock rates for the same design: it reduces the distance the signals have to travel.
David Gould
David Gould
main(i){putchar(340056100>>(i-1)*5&31|!!(i<6)<< 6)&&main(++i);}
I heard that the reason Palm went from the Palm III to the Palm V was because the number 4 is extremely unlucky in Oriental cultures.
Maybe Intel know this and are releasing this as a sacrificial processor before they release the Pentium 5
SteveB.
You arent really supposed to buy one now.
.13micron soon.
Why?
Because of several reasons
1)The chipset is quite large and it will be migrated down to the
2)It potentially will reach speeds of 4Ghz, just wait for them to get it up to 1.8-2.0 and watch AMD fumble the ball. The P4 at 1.4 is only about as fast as a P3 at 1.1 anyway.
3)The Price obviously
4)Like the article said it doesnt support most memory types now, so when it does it will drastically improve performance. Why, because memory speed is the primary limiter for processor performance. Why do you think Intel tried to do that whole Rambus thing??? If you have fast memory it helps your processer out a lot.
Personally Im going to wait for someone to make a mobo that supports the 64bit Itanium, or two of them! then run BeOs -drool-.
Could someone explain something to me?
.18 die process.
.13 micron process, I know their power consumption will decrease, but will their performance increase at all? I mean, if they can't get the thing to 5+ gigahertz speeds, will it be very effective at all in the consumer marketplace?
Obviously, the pentium 4's performance lags (or will) behind Thunderbirds of similar clock speeds. But this thing is also a
When Intel moves to the
---
No motherboards are available? Oh no, I'll have to postpone my multimedia strategy. I wanted to empower the internet users by utilizing an ubiquitous 3D paradigm. Looks like you'll have to settle with the "old" internet until I have a P4.
Who's grepping my arse?
First of all, I was wondering if maybe Intel could some up with some more creative chip names then just using stupid numbers...at least with AMD you have Duron and thunderbird and stuff....intel has Xeon and Pentium 1,2,3,4.
.18 die set...it would make more sense to me to center new products around .13 or something a little smaller...
Second of all, who would want a Pentium 4? Besides the stupid kid in my computer class who saw it in a magazine....it will require a 454 gram heat sink...that is a full pound of steenkin aluminum (or whatever metal they use) A brand new mobo design is needed, and a new case with supports for the processor and the heatsink! Unless they sell a conversion kit with it....
Third of all, they are using
also..I heard some talk (rumors..not necessarily true) that a new powersupply is needed as well. This new Pentium 4 chip is full of heat, sucks up the wattage, and requires redesigned cases and mobos. Unless they are faster than 5 Ghz...i think I will stick with my AMD processor, thank you very much. It has a much smaller price tag, and better performance if you ask me.....
anyways...that is my opinion....
The anti-salmon
We are proud to announce that the first one million Pentium 4's sold will come with a drip tray and an endorsement by George Foreman free of charge.
-atrowe: Card-carrying Mensa member. I have no toleranse for stupidity.
Only one reason... SMP. Of course, once AMD releases the 760MP, they will most likely steal Intel's thunder here, too. The Xeons won't have much of a competitor for a while, unless AMD allows for more than dual CPU configurations with the 760MP. The P4 doesn't even support SMP with the i850 chipset, so it is obvious that Intel has made a mistake here by shooting to overtake AMD in the "I got a higher megahertz than you" pissing contest. The P4 may well be better than AMD's offerings for a *little* while, but AMD will soon be wiping the floor with them.
This comes as a great loss to all of us whom have come to know and love Intel over the past many years. They have just made too many mistakes recently, which will take Chipzilla a long time to recover from completely.
wolf31o2 Developer, Gentoo Linux Games Team
I'm sick and tired of trying to keep up with the advances in this technology, only to find out that there's always some glitch or gotcha that gets 'figured out' in the next generation of components.
Frankly, I don't give a crap about mHZ ratings or benchmarks any more. From now on I'm going to base my component-buying decisions on whether or not the company is *honest* about the problems with its product lines, and whether or not they fix them, standardize on ways to do things that won't isolate existing customers when newer revisions become available, etc.
Right now I have an AMD 750 CPU in my main development system which runs just great. It's not the fastest computer I've ever used, but it is *plenty* fast, and I probably won't need to upgrade it for at least another year (hopefully).
I also have a Mac G4, which, for as much as I've despised Apple for Mac OS9.0, is a really, really terrific architecture. Sure, it's not a dual-proc machine, but it is *fast*, and it's *EASY* to upgrade.
Being a long-time computer geek, I've come to appreciate this simplicity of Apple gear more and more - to the point where the x86 way of life is really just too frustrating. Give me Mac OX X on a fast and well-designed G4, sitting in an *available* (i.e. non vaporware mobo) architecture, with sufficient RAM bandwidth and i/o options (Firewire rocks serious ass), and I'm happy.
From now on, Intel are the last CPU mfr's on my list to pay attention to... I'm so tired of being fed a turd while being told it's chocolate.
; -- the corruption of government starts with its secrets. a truly free people keep no secrets. --
what tomshardware has to say about this turkey.
Yep, you can't get a motherboard for it yet, since it is incompatable with any existing mobo, and worse, will be incompatible with any other FUTURE mobo. It's a dead end. An evolutionary abortion.
Rather than a revolutionary new package to compete with AMD it's something pushed out the door long before its gestation period is up, rudely stamped, deformed, unfinished, sent into this breathing world scarce half made up, and that so lamely and unfashionable that dogs bark at it as it halts by them.
Oh, sorry.
Look, Intel screwed the pooch with the whole Rambus fiasco and not figuring that AMD would EVER be real competition. Now they are behind and scrambling. The P4 is a stopgap measure to get SOMETHING out the door that they can call new and great.It also complies with the already repudiated Rambus contract that they are trying like mad to get out of. They plan to dump the whole thing as soon as they can and cease all support for it.
I don't blame them either.
Wait for the Pentium Squared.
KFG
Why would anyone want one of these chips? Looking at Pricewatch, the chips are going for nearly $1k! Meanwhile, you have a great Thunderbird going for $480'ish. Currently, the low end systems for most games are still running PII 266 with a good Voodoo2! With a combo of Athlon/PIII and a good GPU from Nvidia, you will have a system that runs the latest software at the greatest of speeds. I am sure that Intel will release a great benchmark tool that will demonstrate why we need it, but my question is where is the software to really demonstrate the power of this chip?
Bryan R.
Bryan R.
The price of freedom is eternal vigilance, or $12.50 as seen on eBay.....