Slashdot Mirror
Pentium III hits 1Ghz
Frac writes "somebody has finally overclocked a Pentium III to over 1 gigahertz. Here's the translated version of the original French article. "
← Back to Stories (view on slashdot.org)
← Back to Stories (view on slashdot.org)
The big deal about 1GHz+ processors is now I can use the microwave radiation from my cpu to make popcorn, cook Hotpockets, etc...
Are you kidding? Gigahertz speeds are essential to seeing Netscape crash faster than ever.
----------
In a real emergency, we would have all fled in terror, and you would not have been notified.
What I'd like to see is AMD's Athlon technology (micro-ops, along with all its weird Funky massive parallelism and the like) applied to a decent architecture such as PowerPC. Sure, the clock speed would only be a 'measly' 700MHz, but it would be able to do so much *more* at that frequency...
---
"'Is not a quine' is not a quine" is a quine.
"'Is not a quine' is not a quine" is a quine.
Quine "quine?
Did you read the article?
They used the ASETEK Vapochill system to cool down the coppermine.
Now, since the multipliers are locked on Coppermine processors, there's no way that these people could get a 7.5 multipler. The highest is a "7" with a front-side bus of 100 mhz
Intel engineering samples are not multiplier-locked. I guess we can throw your "simple facts" table from Tom's right out the window.
it'd be wise to check the facts first.
I agree. Do you plan on following your own advice though?
People are questioning the validity of this, and some funny quotes. 1. Don't forget that this is translated... 2. Intel Engineering samples are not multiplier locked. They could've put that thing at 1.066 GHz, too. Wouldn't matter... 3. I still think that those copper Athlons (not coppermine, it's still made of aluminum) are going to rock. A 1.2 GHz Athlon? Oh, yeah...
Someday Intel chips will get so hot and so powerful that only the five richest kings of Europe will be able to cool them.
Gates' Law: Every 18 months, the speed of software halves.
What exactly is the big deal about speeds in the GHz? May be a mile stone, but it's only logical progression.
aÍÍ©ÍÌÍ£Ì'̽ͩÌÍzÍYÌÍÌY
yep.. exactly. I always wonder why laptops never went scsi.. cheap scsi. The ide-to-scsi convertion chips people seem intent on claiming all scsi is, and is thus slower.. except for the extremely pricey drives. Not much slower, and with the scsi bus, no extra weight or really lack of power.. laptops should be scsi!
My old P5-200 SCSI desktop usually outperformed friends P2s on anything not cpu driven, as windows is disk driven.... they used to claim firewire would bring a better standard than scsi at a cost for home and corperate users.. killing the pesky ide. Would hae been nice.. since all of the IDE plusses (ease of installation) has always been nicer on scsi (just stick on a unique ID.. plug in.. make sure your card is detected and usable..)
"Open Source?" - Press any key to continue
Here.
This was done a few months ago with 2 P3 500s by Hardwarecentral. They said it wasn't terribly stable at 1055 Mhz (1.055 GHz), but it DID go over 1 GHz. In fact, they did it with DUAL CPUs.
Results.
(Yes, their CPUs were unlocked, they modified the CPU for the cooling, but it does count)
I definitely hope so. For many tasks parallelism is the easier and more logical way to gain speed, for instance, neurocomputing, computer graphics, games, signal processing etc. Many of the problems could be reduced to separate problems that can be solved concurrently.
Hopefully we will also see transition from programming languages that don't support parallelism (like C/C++) to better languages that are easier to parallelize even implicitly by the compiler (like the functional and logic programming languages).
The programming tools are not really up to massive parallelism yet and it may be one factor affecting development; the change requires a paradigm shift. I don't think many people would want to program a 64 threaded program in C.
I also hope the parallelism will become fine grained where individual statements in procedures can be executed in parallel, which of course puts the burden on the compiler.
1) If I'm not mistaken, the L1 cache in the PIII is at least 64k (32+32). According to the screenshot (which can easily be faked) it's 16 and 16.
2) The L2 cache of the 733 PIII is 256 kb ONBOARD! Why is this important? Well first of all, there is nothing listed under the L2 setting. Did they have to DISABLE THE L2 CACHE to get it to work? If this is not the case, then why isnt' the L2 amount shown? If it was indeed disabled, then it's extremely doubtful that those benchmarks are reliable.
Performance of todays processors is almost as much dependent on cache performance as archetecture and design. By increasing cache performance and memory throughput through the chip, you are decreasing the amount of time that the processors instruction pipelines are left empty. With no L2 cache, this would prove an incredible crippling of the chips ability to execute instructions.
The only other possibility here is that the L2 is disabled, but the test scores are real. This being based on the idea that perhaps, all of these benchmarks are running within the L1 cache, which if true, would mean that the tests are not indicative or real world performance.
But then again, I could be wrong,
Big Din K.R.
I really don't think you understand how this works. An R12k SGI Mips processor running at 250MHz on a typical O2 will absolutely blow away a 1 GHz PC. Even the older R10k at 180MHz is much faster than a 500MHz PC. One reason is that they don't use a bus. All memory is shared instead, which gives you over 2.1Gb per second data transfer!!!! And then we go up to 1000+ processor machines with over 1 Terabyte of RAM (That's 1000 Gigabytes - roughly)
We like the Cray T3E
There is no contest. Believe me! PC's are currently baby food.
On the super PC note, you must have seen Slashdot's article about SGI's new Linux Supercluster. Should be the fastest machine in the world.
No, they're not.
So? Generating 32-bit constants is not what interesting programs spend most of their time doing. Optimize the common case.
To sum it up, I'd have to disagree on several points:
- Complex instruction decoders are a bottleneck to issuing large numbers of parallel instructions. It's more than just a power issue.
- The 32-bit constant generation example you have is not a typical bottleneck in most code. Control flow is more likely to be.
- x86 is less efficient than alot of other machines due to being a register-starved architecture. It has to rely on tricks such as store-to-load forwarding and micro-caches to turn a small cache of memory addresses into a pseudo-register file.
- While Intel has generally been the "Floating Point King" in x86 world (with Athlon recently taking the title away), x86 floating point as always sucked compared to architectures that have traditionally taken FP seriously. RISC wins squarely here, period.
--Joe--
Program Intellivision!
Not exactly, but close. Signal propogation delay is getting worse and worse with respect to clock cycle. In the early days, you could consider wire to have zero delay, since transistors were so slow compared to the propogation rate on the wire. As wires have gotten smaller and smaller, their resistance has gone up. Meanwhile we've packed them closer and closer together, so we have tons of capacitance between wires. And finally, transistors have gotten orders of magnitudes faster.
End result: It takes bloody ages for a signal to get anywhere on the chip with respect to how long it took to generate the signal.
Most pieces of silicon nowadays operate with various "domains", each of which has its own local clocking. Depending on how fast you're running, it can take several clocks just for a signal to travel from one end of the chip to the other, so designers tend to subdivide problems into domains that aren't more than a clock-cycle wide. Pipelines and replicated hardware help a little, but physics really starts to bite you in the arse. Copper helps a little here (since it lowers resistance), but it's not a cure-all either.
Our friendly constant 'c' is a couple orders of magnitude above the propogation rate on the chip, so it's not the main limiter. To put it into perspective, light travels 30cm every nanosecond, and chip dimensions are usually closer to 1cm on a side. But we're getting uncomfortably close. :-)
--Joe--
Program Intellivision!
Awhile back, the X3J11 group responsible for the ANSI C standard was looking at some rather nifty data-parallel C extensions that retained the otherwise "serial programming" nature of the control code. I suspect that an evolutionary approach such as this is likely to gain more ground than forcing people to think about programming in a completely different paradigm.
Postscript and text files containing the Data Parallel C Extensions draft is available here: ftp://ftp.dmk.com/DMK/sc22wg14/data-parallel/
--Joe--
Program Intellivision!
So 733MHz goes to 1GHz.
Next year we get an Athalon 1.1GHz, which we overclock to 2GHz. And so forth.
But why is this such a big deal?
Probably the same reason as going from 700 mph to breaking the sound barrier in the air was, or having the first car exceed 100 miles per hour. Its a logical progrression, but milestones are important in our culture. Why do we celebrate our 50th birthdays or 50th anniversaries? Well, because they are milestones.
I'm still using a p/233MHz and it still offers me instant gratification for anything I want it to do. What's the rush?
OFTC: By the community, for the community
Now the big question is -- where the hell did they get a PIII CPU with a multiplier of 7.5? Coppermine 733 has bus speed 133 and multiplier 5.5. Since the multipliers are locked, there is no way to change them. The PIII would have to work at 750MHz (100MHz bus) to have a multiplier of 7.5, but such CPU does not exist. The only CPU that has a multiplier of 7.5 is Celeron 500.
So does that mean it's a hoax?
___
If you think big enough, you'll never have to do it.
Then, after the solid whooping, I'd like the other one to come back and do the same. A nice see-saw of continuous ingenuity that allows me to maximize my utility, while one company or the other maximizes it's profits.
Let the best man win!
-----------
"You can't shake the Devil's hand and say you're only kidding."
According to an article I read in New Scientist, isn't 1GHz around the frequency one uses in domestic microwaves? So as well as yer average Pentium running hot enough to fry eggs, will it be that the inside of a computer is going to resemble a combination oven now...?
Matthew @ Bytemark Hosting
It's a fairly run-of-the-mill "Hey I overclocked my processor to Ludicrous Speed!" article. However, I nearly fell out of my chair laughing at some of the truly odd translated phrases...
"If no processor given rhythm at this frequency..." - these processors have rhythm baby! This line reminds me of the dancing bunny-suit Intel engineers.
"While waiting, greediest in MHz can fold back itself on the systems KryoTech Cool Athlon 800 and 900..." - greediest in MHz? Fold back itself while waiting? Sounds like a contortionist that wants all the MHz for itself...
"While waiting, history to dream a little and to see what will have in the belly the processors of demains..." - no comment. Too good for words.
"...on a chipset BX the AGP is to 88 MHz when one uses a drunk face side of 133 MHz..." - my question is, if it's drunk then is it still greedy? Does it still have rhythm? (I know that I lose any rhythm I have when I'm trashed)...
"...the profit of 39% is a little less significant but remains impressive, even if it is also with the read-write memory given rhythm to 133 MHz." - ok it's getting dumb now (the joke, not the article) so I'll end it here... but I like to think that my read-write memory has rhythm too, since my Celeron 300a doesn't shake its stuff too well yet...
The simple facts (from Tom's Hardware)
Coppermine processors:
Rated Speed Bus Speed Multiplier
733 133 5.5
700 100 7
667 133 5
650 100 6.5
600 EB 133 4.5
600 E 100 6
533 EB 133 4
550 E (PPGA) 100 5.5
500 E (PPGA) 100 5
Now, since the multipliers are locked on Coppermine processors, there's no way that these people could get a 7.5 multipler. The highest is a "7" with a front-side bus of 100 mhz.
Note that they also make absolutely no mention of any method of cooling this thing. The only thing presented are the benchmarks, all of which are directly proportionate to the increase in mHz rating. Coincidence? I doubt it. There's no mention _at all_ of the hardware this was tested on, except for one screenshot that mentions an ABIT BX6 (Rev. 2) motherboard.
Before you go expounding on the wonders of 1ghz Coppermines, it'd be wise to check the facts first.
Well, the good news for AMD is that this is just an Overclocked 733 (although, I can hardware junkies trying out the 153mhz bus speed on some 'overclocker' boards like the bx6-2), so right now AMD still has the fastest chip. Of course, if the yields are high enough over at Intel, they could hit 1Ghz quicker then AMD.
:)
I personally hope that AMD can keep ahead of Intel; their competition has dramatically lowered the cost of PCs in the past few years. When I purchased my p75, it cost me $1500, and that was about as far down the line as you could get in terms of CPUs (The chips out were the 75, 90, 100, and 120 (little did I know at the time that my little 75 could be overclocked to 133). This Saturday I saw an Athilon 600 system for the same price (made by IBM no less). Cheap PCs are good for everyone
Of course there are probably other reasons as well, such as the emergence of the PC as an 'internet appliance' But if it weren't for AMD, they'd all be using Pentium 200s, instead of celeron 466's (witch by the way, was selling for about $500 in a Compaq that day.)
--
"Subtle mind control? Why do all these HTML buttons say 'Submit' ?"
ReadThe ReflectionEngine, a cyberpunk style n
The way I see it, we'd have had retail 1GHz P!!! processors even before now if only Intel had its mind on the right technologies instead of pie-in-the-sky technologies that no one really wants. How many of us really care about Rambus, considering that we haven't nearly reached the limits of SDRAM? Think about what graphics cards manufacturers are using in their next-generation (or for the GeForce, current-generation) cards--Double Data Rate SDRAM, much as the Athlon FSB is really 100MHz but transfers data twice in the time it takes a normal FSB to transfer once, DDR SDRAM would effectively double memory performance whereas Rambus will start out more expensive than SDRAM and perform worse than 133MHz SDRAM.
We could have DDR SDRAM 200Mhz memory right now if Intel had supported it instead of Rambus, and DDR SDRAM would quickly reach effective speeds of 300MHz--150MHz x 2 times the transfer rate.
But what does this have to do with the 1GHz mark? Intel pushed Rambus for its own agenda, not caring about customers' needs; likewise, development has been almost completely shifted to Merced (oops, make that stupidly-named-Itanium) instead of pushing x86 to its limits first. Were it not for AMD--and this is supported by Intel's own "development roadmaps"--we wouldn't even have seen 700MHz Coppermine this year (and we still won't get it in quantity this year). We have AMD to thank for 700MHz P!!!, which is reason enough in my book to buy Athlon--Intel simply does not care about the consumer, they care about pushing the unnecessary and too-expensive technology of their Rambus partner, and they care about finishing their high-end server processor Itanium; they do *not* care about making their soon-to-be-low-end-compared-to-Merced P!!! run as fast as it can for their customers. Why do you think Intel is suing VIA, whose 133MHz SDRAM chipset beats Rambus performance to Hell and back? This is not off-topic--this is why we don't have 1GHz retail processors by now, which is as on-topic as it gets. Intel, we want x86 at well over 1GHz before we even want to think about Merced. We want you to care about what we care about. But since you don't, a lot of us are going to start thinking AMD Inside instead of caring about Intel, and by the time your prized Itanium rolls around we just might drop it in favour of AMD's 64-bit x86-on-PCP offering. Think about it.
"The more corrupt the state, the more numerous the laws."--Tacitus, *The Annals*