Slashdot Mirror
C`t Throws Athlons And P4s In The Gladiator Pit
An unnamed correspondent writes: "In the most recent C`T "Computer technik" there is a great benchmark with a pentium 4 (1,5 and 1,4 ghz)vs a athlon thunderbird (1,2 ghz and 1,2 ghz ddr memory with the 760 chipset).
If you think that that isn`t a fair race ... then read it now here and here in English.
You should get a copy of the German paper version anyway -- great magazine, even beter benchmark.
Now does anyone know where to get a 760 mainboard ;-)" Unnamed's cousin Noname also contributes a link to GamePC, which
reviews in grand 13-page SE-style the 1.4 and 1.5 GHz P4 chips.
Not always true. Having just interviewed with a major company that deals with the different processes and fabs, a smaller line width isn't necessarily the holy grail most people think it is. The smaller the traces get, the more complex the die must become and the routing of the silicon becomes much much harder. Any two wires running in parallel constitute an antennae, so often the traces have to be seperated and/or made longer, so making the linewidth smaller can slow down the chip in some cases.
> the animosity towards Intel has been building for some time.
> And people seem really happy with AMD
I was an all-intel guy since the 80286 through PPro200.
What turned me off have been the CPUID thing. As soon as they released their chips with the CPUID, I decided not to buy any Intel chip any more
Had a lot of problems with my AMDs. But at least, there is now competition in PC arena. I am amazed at the perf of recent Athlons.
I tend to buy 1 personal machine per year, 3 or 4 for the work, and a couple of ones for friends. Now that intel is trying to fuck everyone with the RAMBUS thing, it is very unlikely that I buy any Intel CPU again.
Cheers,
--fred
Intel sacrifices die space for good design; I think this processor actually takes *both* branches of a jump, insofar as that is actually possible. I'm sure it does all sorts of complicated prediction on top of that, too.
Give me a Transmeta-style design any day...
Well, according to an article on www.theregister.co.uk, a 800 MHz or 900 MHz P3 outperforms a 1.4GHz P4... so it wasn't much of a challenge for AMD.... ;)
if you honestly believe the k6 series is as good as the p2 or p3 you are seriously misinformed. At first i thought we were a troll, but you probably aren't. most of what you say is correct, but you don't come off all that well when you start out by praising the k6 series chips.
Man, look at the POV-Ray scores.. Athlon 1.2GHz beats P4 1.6GHz by 960 to 834. Intel claims that "The Pentium 4 processors that we're announcing Monday have the highest performing floating point of any PC processor that's out there", and yet the Athlon is more than 50% faster at 3D rendering, clock for clock? What an embarrassment.
I'd like to see how an entirely native system does. Of course, this will be a *lot* more expensive than your average system, especially with the weird RAMBUS requirements, but a lot of people will be happy that it's not x86, and that it's faster than 500Mhz...
:)
However, my 800Mhz, non-RAMBUS Athlon should last me quite a while, thankyouverymuch.
---
pb Reply or e-mail; don't vaguely moderate.
pb Reply or e-mail; don't vaguely moderate.
Most consumers wouldn't be able to tell the difference in performance between the two. Those benchmarks are showing, 5-10% difference when running quake at 150 FPS. If it wasn't for the frame counter/speedometer you wouldn't be able to tell the difference.
The PIV is clearly designed to go to ultra high clocks, that has been the business. Intel has also done poorly, historically, with first revs of a new architecture. I'm guessing by summer there will be a refined PIV and by this time next year they will have tuned the caching and probably changed the branch prediction hardware enough to eek out that enough performance to sit on top, this is intel.
The one real thing that article does show is that AMD has put together a very competitive processor. It has been pretty clear for some time but it was nice to see a new generation Intel part come out that didn't destroy everything in it's path. Moore is catching up.
yep, alpha can use standard PC parts. ('cept soundcards which have a devious tendency to not implement a few of the higher order address lines out of cheapness).
Cheap though they are not. In fact they're damm expensive new - you'll never get a 21264 for anything less than 3k.
However, the previous generation of Alpha can still be had secondhand, even from some vendors, for a much more reasonably price. Eg you could get a 600MHz 21164A for about IEP£1k to £1.5k, or motherboard+chip for about £600.
I use Friend/Foe + mod-point modifiers as a karma/reputation system.
This is funny... If I looked hard -- even not SO hard -- I could prolly find 8 places in south KCMO that sell an athlon... maybe even at a competitive price to online retailers! :o)
Of course, Intel shot themselves in both feet (as well as their head) when they went RAMBUS, so no money from me!
--
--
Me spell chucker work grate. Need grandma chicken.
Correction. The initial generation of P4 doesn't have SMP support enabled. It's not that they're not designed for it (any more than P3 was).
The issue is that they're focusing on single-processor systems for roll-out. This is for ease of troubleshooting. They'll work out most of the bugs in the single-processor soloution before compounding any problems trying to prematurely release the SMP boards to market.
But if they did that, you'd want to gripe about how "buggy" their SMP boards were right?
At least try to appear impartial here. Brand-zealotry makes it too easy for inaccurate statements like yours to be made with impunity.
Chas - The one, the only.
THANK GOD!!!
Chas - The one, the only.
THANK GOD!!!
He never made the distinction about commodity components vs full systems from integrators. Nor was it important to the point I was making.
Chas - The one, the only.
THANK GOD!!!
Chas - The one, the only.
THANK GOD!!!
The processor was just released. So arguments about availability are kind of meaningless.
How available were Thunderbird processors when they were released?
How available was the Athlon?
How available was the P3?
Howabout P2?
Of course components that have had more time to penetrate the market will be more available.
Chas - The one, the only.
THANK GOD!!!
Chas - The one, the only.
THANK GOD!!!
At work, I get 4x rate on the mp3 creation, and at home I get 12x.
4x and 12x compared to what?
"onward!" cried the copper man, little knowing brass corrupts...
Well, I don't know what you drive, but my car has a knock sensor and will advance the timing until it detects a knock, then back off until the knocking stops. I always use 93 octane, and I do notice a difference when I get to the track and can fill up with 105 octane.
Just junk food for thought...
this is a very valid point. I have a system that runs all the applications that I need more than adequately. Even though I am college student taking some CS classes, paper writing, etc. I spend the majority of my CPU cycles checking email, IRC, AIM, etc. If I didn't run distributed stuff would I ever use more than 2% of my CPU? Probably not.
Problem is this: the man who dies w/the most toys wins..
A lot of people have a lot of money and are willing to buy systems w/it. They do it to beat the Jones.
Like someone said before, this was a marketing descision (to run up the clock speed but not the perfomance). People are going to rush towards the GHz machines just b/c of that.
Just my worthless rambling.
no the Alpha is slower b/c it is from 1993 and the fact that it is a known fact that it has an inflated clock speed... I wasn't comparing the two systems I was just listing their keys/s. Plus I wasn't really serious about using it :)
I think that RC5 is a good way to determine how fast the damn processor is...
:)
;-) I could care less about Quake3 benchmarks.. Quake3 sucks anyway :)
Alpha 166 UDB -- 184k keys/s
Intel 400mhz Celeron -- 1.1m keys/s (2.2 for dual
etc..
Personally, I think the average Joe has no need for all this high-power computing. What it really comes down to is that the average computer illiterate only uses the computer to check email, surf the web, play games, and maybe use apps like Word
How's that old 486SX holding up?
0 1 - just my two bits
Intel has always been way ahead of AMD though.
Until recently.
For older (read: non-pipelined) machines, you might use the inverse (clocks per instruction or CPI, a 68HC11 might take 3 clock cycles to execute an ADD instruction). It's generally an average over a plausible mix of instructions that would be a significant part of a normal software program.
Well, how do you make CPI into IPC? First, you pipeline the processor. Theoretically, you could get your CPI or IPC to near 1.0. However, certain things such as loads from memory might require you to wait a clock or two, pausing the entire pipeline. That'll keep your IPC from being big (big generally = better for IPC).
Okay, great. You did your best trying to get it near 1.0 and you couldn't quite do it. How about starting multiple instructions in each clock? Of course there are complications, you can't do an ADD that requires an answer before the multiply that supplies that answer has completed. So there are limitations. But for many instructions which are not inter-dependent, you can issue them all at the same time. Surprise! Now you can get an IPC > 1.0.
Yes, this is a grossly simplified answer computer architecture answer condensed in a can, but maybe it clears things up a little.
/ \
\ / ASCII ribbon campaign for peace
x
/ \
I can't do anything BUT disagree. The more "friendly" computers are expected to become, the more technology that is brought into the hands of the end consumer, the more CPU horsepower it will require. There is no shortage of ways to use the extra bandwidth. Will grandma need an Athlon 1.2 Ghz system to store her recipies? Probably not. But when she whips out that new word processor that is completely voice-driven - with real-time recognition - it's a whole new ball game.
A new motherboard and case shouldnt' be looked at as being this huge expense. After all, you had to buy a new mobo when you upgraded from Pentium or Celeron to an Athlon system... And the heat sink may be intimidating, but it won't be the sole factor that a potential buyer changes their mind...
But as for the price of the actual CPU's... Intel needs to get off their high horse and realize that they need to win back their market share, even if it means not living on 40%-60% margins on their high end chips...
But intel is much more successful at convincing software developers to adopt their new technologies than AMD's been, so far.
There's no more important set of benchmarks than the holy 3D Game Benchmarks!
However, that may be due to DirectX 8.0's P4-awareness.
Then again, I don't think the nVidia video drivers actually USE much DirectX. The only two other things that use would DirectX as far as I can tell in Quake3 would be DirectInput and DirectSound. Would they make THAT much of a difference?
Maybe we should benchmark a P4 using an A3D soundcard, thus bypassing DirectSound also.
:)
- Ed.
According to the other poster it's allready been done where it counts in the graphic libs. Chill out some you obvious intel basher, caus its going to happen.
I believe what he is saying is that the benchmarks should be optimized for whatever processor they are running on. P4 for P4 and Athlon for Athlon rather than P3(or whatever) for P4 and Athlon for Athlon.
-- themenace
The Pentium 4 shouldn't be confused with Itanium, the P4 is 32-bit just like all the other Pentiums
You won't like it when the guy dreaming us wakes up...
It is hardly fair for them to be releasing Pentium 4 benchmarks as Intel has delayed the P4 chip for "re-structuring." I know many reputable review sites have delayed/will re-do their reviews when the actual production Pentium 4 chips are out.
"There is no surer way to ruin a good discussion than to contaminate it with the facts." -- Cecil Adams
I used to read C'T when I was living in Germany, it's a great magazine full of technical information.
Can any one recommend an equivalent magazine available in the States either on the "high street" or by subscription? PC Magazine and PC World just don't cut it. Byte magazine was OK but is now web only.
Suggestions anyone?
Thanks
in europe they use , instead of . and . instead of ,
since c't is german i would say thats ok.
john
-- john
What do you need 64 bits for? To double the stack size on a context switch? 32 bits means each register can hold one of 4 billion different numbers. Thats more than enough for most applications and with those apps where its not enough, 64 bits isn't going to help much at all.
A big part of the new P4's strength wont be shown until people start compiling with the new V5 system and using the new instructions.
The branch prediction is supposed to be about 94% accurate and can be made near 100% with the use of some intelligent compiling technologies. Dont forget the performance on the new SSE-2 instructions as well.
I love AMD, but they refuse to initiate rather than imitate. Even their plan for a 64bit processor refuses to step out of the path. (64bit with x86 instructions) A vast majority of this site runs linux and are therefor mostly instruction independant. I would think this population would embrace a new architecture if it provided a performance increase.
FunOne
FunOne
SPECfp, iirc, is designed by the cpu manufacturer. Intel may have done funny things to their compiler, to make it look better. AMD, as you said, may not have cared.
Most all manufacturers do this, I wouldn't be suprised if AMD had done this in the past.
SPECfp, for the most part though, is useless. Sorta like the old Mac benchmarks that showed it to be better (ByteMark, was it?). They were written by Apple, and pulled all sorts of shortcuts.
Is it possible? Sure is!
Will it be expensive to implement? Sure will be!
The NB already has a ton of pins on it. I dont think they will want to add 80 or so more to the package. Not enough return in performance)
--
Don't lead me into temptation... I can find it myself.
For what it is worth :-) I had a heavy IO application running on a Celeron 500, and moved it to a Thunderbird 850. The configurations were similar otherwise. Based on my application the Athlon runs about 2.5 times faster than the Celeron. Even counting the bus speed difference (66 to 100), this is an impressive improvement.
Would you see the same improvement? I doubt it. But I think it backs up your point that the only "benchmark" with meaning is the one you run yourself...
A dingo ate my sig...
Just out of curiosity.. what is IPC? Can you explain this a little more. I would look it up myself, but I haven't the faintest clue about what you are talking about. Thx.
JOhn
Campaign for Liberty
Then why not trot on down to Microway and get yourself a smoking, 64 bit quad 667/750MHz Alpha 21264 system.
Sure, it's not a cheap way to go, but the Itanium isn't going to be a budget chip either, and the Alphas are already available, have been 64 bit for years, run Linux/Tru64 UNIX and a pokey version of NT4, and use standard PCI expansion cards.
Itanium will have to be a monster to compete with the 21264's FP performance, and the 21364 should kick sand in Itaniums face all the way down the beach.
You can play with some beasty Alpha systems on Compaq's TestDrive program. I used BMRT (Blue Moon Rendering Tools) to do an informal radiosiy render test and a quad 667-MHz Alpha with 2G of RAM certainly pumps the pixels.
I gots ta ding a ding dang my dang a long ling long
the RC5 algorithm uses 32-bit rotate operations, which is part of the hardware instruction set for Intel and AMD chips (and PowerPC chips too).
BLOCK STRUCTURE breathing apparatus required for special maneuvers!!
RC5 might be reasonable for comparing x86 cpu to x86 cpu, but it's unreasonable for comparing x86 cpus to non x86 cpus. RC5 uses instructions that many processors don't have (but the x86 (and PPC for that matter) does have them) (few applications need these instructions, so it's rarely a problem) and so they have to use more complicated routines on those cpus. That's part of the reason that your Alpha cpu is so much slower at RC5 than your x86 ones.
The 1.2 athlon itself doesn't outperform the P4, its the DDR memory. and what monopoly? that was gone when the original athlons came out.
It's not so bad, unless it starts to get stiff. Then, it's a problem....
Linux rocks!!! www.dedserius.com
www.dedserius.com
VB != VisualBasic
Oops, I forgot to add that I found 8 vendors off Pricewatch. I would imagine that the total number of vendors might even be as high as 9 but I'd want a hand recount ;-)
Yes.. I was unfortunate enough to see one of TV here in Scotland last night. It muttered something about the power of three, at which point I turned it over. Any other Wiccans out there that find that really offensive? Would they use things from Christianity to sell computer chips like that? Somehow I doubt it.
The advert didn't say 'put the holy trinity' in your PC'.. :op
;o>
;o>
The very term 'the power of three' is linked with Wicca/Pagan mythology, which is why it's known widely enough that they'd drop it in an ad in the first place.
Of course.. I could be reading this wrong. Perhaps Intel hired a Pagan ad agency.
Now, when we see someone on a cross being used to sell stuff from PC World, then I'll happily stop whining.
Sure some people look for clock speed over true performance, but I don't think that's the reason X86 chips have been outselling PowerPC. It has more to do with the fact that X86 chips are cheaper, more varied, and aftermarket parts are much easier to come by. Also that's where the software is.
just don't forget that those benchmarks aren't optimized for the t-bird either.
"It's such a fine line between stupid and clever" -- David St. Hubbins, Spinal Tap
*sigh* you're damn right. remember 3dnow? since k6-2 it was on the market. at least 6 months before intel came with isse. but only a few developers adopted 3dnow. then intel released p3 with isse (maybe a bit technically superior but not so well documented) and ho! everybody uses isse. thats the only reason adobe photoshop is faster on intel cpus. now almost noone remembers 3dnow even if athlon &co use it. even amd decided to drop 3dnow for their sledgehammer and use isse2. that's sick! but that's intel. oh well then amd will also have a cpu that makes internet even faster *bitter laugh*
microsoft uses same tactics... see websites. most of them now are optimized for ie. not netscape, not w3c html4 conform, ie.
always the same with (ex) monopolists.
"It's such a fine line between stupid and clever" -- David St. Hubbins, Spinal Tap
Your observation is pedestrian, and also shows your ignorance of P4 architecture.
The P4 was designed to cope with branch mispredict penalties and has the most advanced branch prediction of any processor in the market. Moreover, it has the execution trace cache which makes loops much, much cheaper than on other processors.
The main problem with P4 speed is not the long pipeline, but the following:
1. Slow shifts and multiplies. Hence, it does not do well on RC5, and other such benchmarks (Alpha, I might add, also does not)
2. Unoptimized code. Code, particularly FP code, needs to be compiled with SSE2 for top performance.
The SPEC2000 benchmarks (with optimized code) show P4 as the fastest processor in the world on integer (beating out Athlon and Alpha) and the second fastest on FP (second only to Alpha, and MUCH faster than Athlon and most other RISC processors).
It should also do well on memory benchmarks with the 400 MHz bus (vs. Athlon's 266).
The answer is that the Pentium 4s were designed to not be SMP capable
This is incorrect.
However, the Intel parts have an unfair advantage there, being clocked at 1.4 and 1.5 GHz rather than 1.2.
You're kidding, right?
Bad - The Pentium 4 is simply UNAVAILABLE. Intel can't produce enough, plus the parts for the P4 are 3 times as expensive.
Wow. That's a LOT of FUD for one sentence. The P4 hasn't even been LAUNCHED yet, but already is available from several vendors. Wait until tomorrow when it is lanuched, and when every major PC vendor will have a system with the part. What's your proof that "Intel can't produce enough"? Intel executives have said that hundreds of thousands will be available by the end of the year.
4x and 12x the speed you'd need to READ it in order to play the music without skipping.
The Web is like Usenet, but
the elephants are untrained.
Remember that the P4 1.5GHz costs more than twice as much as the TBird 1.2GHz.- -----
The thing would have to be a _lot_ faster to make up for this huge disadvantage...
--------------------------------
But intel isn't competing on the performance market ... they're leveraging their brand name and the biggest clock speed to sell chips ... they know they're not as fast as the AMD's ... they're not stupid :)
IMHO cpus should be sold by their Average CPI*Clock Speed ... whole computers should be sold by their CPI with cache missess * clock rate :)
just my oponion
Free Techno/Jazz/DNB/MI Music by guys obsessed with monkeys!
Good - People don't look at anything abut MHz rating when they buy computers. Bigger number means faster. So Intel makes it look like they have a faster computer and people without a clue (everybody) buys it.
Bad - The Pentium 4 is simply UNAVAILABLE. Intel can't produce enough, plus the parts for the P4 are 3 times as expensive. You gotta buy a special new case, power supply, motherboard and 1 POUND HEATSINK for this thing. You could replace the entire heating system of an Antarctic research station with one P4 machine. I would sa that this is borderline Kryotech excessive.
I think I'll stick with Athlon for a while.
The thing is, the Pentium 4 is slated to hit 2.0GHz by January 2001. The Athlons can't scale up as fast, and when they do scale up, they lose IPC like mad. Athlons and Pentium 4s are both good. However, the Pentium 4s will go faster in a shorter period of time (that was its design goal).
That could very well give the chip the edge especially if the optimizations are screwing with the branch prediction and out of order execution in the AMD chip by going against the chips better judgement. It would be nice to get some info on the state of the pipeline while the test executes. I am betting you would see lots of wasted time as various tweaks that kick butt on intels hardware confuse the AMD's optimizations and require it to flush the pipe again and again.
"You can now flame me, I am full of love,"
BTW the Specfp 2000 and SpecInt 2000 scores are done using the intels new compiler and are very heavy SSE2 optimized which will give you an idea of how good it gets. But all todays software can't use it and at legecy floating point code (especially without Nvdia drivers), like Unreal tournament, Quake II, 3d studio max or SETI at home, the Pentium 4 really sucks and can be killed by a Duron.
Actually the 266 Athlons were launch on the 3rd of november, some of the newer micron PCs have them. But the chips probably won't be available off the self for upgraders till january.
I had an old Athlon 550 one of the first from october last year. I just picked up an old Slot A Athlon 750 overclocked to 1GHz, it was a special offer from Overclockers.co.uk. I fitted it and its runs rock solid even for unreal tournament and overnight downloads.
It looks to me that in neutral apps (ones that have no optimization for AMD nor Intel) Athlon beats Intel CPUs in FP hands down. In the end it depends on compiler...
_________________________
_________________________
Spelling and grammar mistakes left as an exercise for the reader.
It's still unknown why the Quake 3 benchmark shows better results for P4 than for Athlon. Some people say it's only because of the good SSE support of Quake 3. For example look at those PovRay numbers which are not biased against either CPU. In fact the Quake 3 rating is the only real-world-like benchmark where P4 is faster. In addition SPECfp results are heavily biased towards P4 - at least according to this Anonymous Coward. I'll stick with my Duron for now.
_________________________
_________________________
Spelling and grammar mistakes left as an exercise for the reader.
The father, the son and the holy ghost. there you go. Christian doctrine being used to sell products.
The funny thing about the Blue Man Group in the advertisements is that they use PPC for their work not Pentiums.
http://www.apple.com/creative/stories/blueman/
Not quite correct. The Athlons FP pipeline
is 15 stages long.
also notice that the intel system had double the RAM...
Public perception is going to be the one that makes or breaks the P4. Public doesn't seem to be thrilled with RDRAM but Intel is finally realizing it and is working on DDR SDRAM chipset(s) for the P4. This is going to take a little bit of time to do properly. I think the biggest problem for Intel with this whole thing is the state of the current chipset. It doesn't support SMP and the market that it is going for is huge on SMP. The next revision of the chipset that supports SMP is not supposed to be out until later 2nd half 2001. This chip is aimed at workstation and server market. Dell said that wouldn't make a single CPU workstation model but supposedly have one on the way for the P4. Dell definitely has been loyal to Intel. The chances are very good that AMD will have a SMP chipset out before Intel and could let some of the steam out of the Intel SMP launch.
About the cost issue. I've got a 600mhz athlon system. Are you saying that I could stick a 1ghz athlon in it w/o modification?
Isn't 454g precisely 1 pound? Sounds like they used imperial for specification, and converted to metric. Although it seems strange for them to use pounds at all anywhere. I thought the whole silicon industry had gone metric by now.
Now calm down.... Just because he hasn't heard of Samsung (Korean), Siemens (German), Hitachi (Japanese) Doesn't mean that there is a silicon industry outside the US.
It makes me wonder what the company I work for in England really does though.
It seems that Athlon always blows away P3, P4 at all 3d games except quake. Does q3 use SSE or something? Why doesn't it have any 3dnow. (I think most hardcore games would rather buy an athlon for the excellent price performace ratio). I just hope Doom3 is not heavily optimized for P4 only, eventhough id might have the bucks to shell out for high end Intel stuff, we frugal gamers prefer the Athlon line.
They misunderestimated me. -- George W. Bush
well at least the pent 4s play quake well, that IS all that matters, isn't it?
Isn't that transmetta's thing? Narrow bandwidth? Yeah, that did them a lot of good...
If voting were effective, it would be illegal by now.
Do any of the upcoming DDR chipsets for either Intel or AMD support a dual channel DDR solution? Is that even possible?
If voting were effective, it would be illegal by now.
I wonder if the vast FCPGA north bridge is really a PIII clocked at 400mhz? It would certainly require a flexible chip to perform I/O functions for a P4 if their previous north bridge chip had trouble. In any case even if the north bridge is not a P!!! it is still manufactured on the same process I assume cutting into P!!! production. Combined with the vast P4 die size, I am guessing if this is anything more than a paper release it will break Intel.
If voting were effective, it would be illegal by now.
This is only going to be a problem for Intel in the short term. They've designed this CPU to scale to ~3GHZ, and at that speed the misprediction will be less of a problem. Intel has a long history of making tradeoffs between current performance, and long term scalability.
Think outside the... Hey, where'd the friggin' box go?
Gaming is the most CPU intensive work that ~90% of CPUs ever do. Would you have them test with weather simluations, aerodynamic analysis, and PERL scripts? Most people never do anything more than E-mail, word processing, and yes - games.
Think outside the... Hey, where'd the friggin' box go?
in europe they use , instead of . and . instead of ,
Except in the UK, where we use the same as most of the world; '.' for a decimal separator and ',' for a thousand separator... So, despite the UK being part of Europe (the continent) we're much closer to the USA than we are say France.
Five years ago I bought a Cyrix 486 clone, because it was the best I could afford ... it wasn't great, but it was cheap, and I haven't bought an Intel since. My AMD K6-2 was as good as a Pentium II, and my AMD K6-3 is better than a Pentium III, plus I saved about $200 on it. What a long way AMD has come ... when I bought the K6-2, it was being touted as a bare-bones processor for low end machines. Now AMD is clearly pulling ahead of Intel in the race for the best possible processor, and they are still cheaper!
As for Power PCs, there are two major problems with them: (1) Apple is the only manufacturer to adopt them heavily (2) Mac hardware is ridiculously overpriced. Now that Power PCs and x86s can both run Linux, I'd be happy to have a PPC system, except that I would pay way too much for peripherals. My roommate paid $30 for a mouse for his Mac and $100 for a CD-ROM drive, that is ridiculous!
Finally, I'd like to get an Alpha system if I can. I've heard they're super fast and pretty cheap and they can use most standard PC peripherals as far as I know ... am I right about that?
"If we couldn't laugh at things that didn't make sense,
My bicyles
The whole point of RAMBUS is that it has a much lower track count to impliment in comparison to SDRAM. This means that you _can_ have multiple channels on a motherboard at a much lower cost. Idealy RAMBUS based systems should be designed with 2 or more channels to take advantage of this.
RAMBUS memory solves a lot of problems when it comes manufacturing and when used correctly (ie. the Playstation 2 with 4 channels) it can be devestatingly effective.
Unfortunately it is not easy to work with, good chipset support is very difficult. Couple this to cost and the fact that the stuff is so patent overloaded the only real reference comes from RAMBUS its self. You do not get many companies wanting to create better platforms.
Rambus type technology will have its place in future computers but it is only going to take off long after Rambus the company is dead.
"The P4 is using dual RDRAM. I've seen a lot of benchmarks showing dual channel RDRAM to be slower than PC133."
That is the point. Intel is touting RAMBUS as the greatest thing ever, and its really a POS.
"And remember the P4 has much better frequency potential than the K7. Possibly being released at 2GHZ in q1 2001."
Megahertz doesn't mean jack. Oh wow, they have 2GHZ, but it doesnt mean crap when an Athlon 1.5Ghz is beating its ass.
You are the exact person that Intel is aiming at - someone who doesn't know jack.
The problem with the z80 analogy, and the entire idea is that the P4 basicly requires a compiler to do anything. Whereas most (if not all) z80 programs are written in assembly. You neglect the fact that the older processor will have more efficient code (unless there is in the future a debian-2.3-intelp4, debian-2.3-amdathlon, debian-2.3-intelp3, debian-2.3-amdk6-3, debian-2.3-intelp2 and any other processor).
Nice idea in theory, but wrong in real life. However the part about increase of transistors should be right.
Now we have 16-instructions per clock cycle chips (Alphas 21264), which had 4 instructions per cycle from at least 1995 (original 21164s).
The G4 can do 4 or 8 per cycle.
Anyone know any others?
The reasons are future compatibility, and 999mhz1000Mhz.
>If money is no object and all you care about is FP speed, buy the Intel. Otherwise buy the Athlon. No, buy an Alpha 21264, which will beat them both in integer, and will cream both of them in FP. That is if money is no object. Otherwise get the athlon.
AMD has already started to build a brand, it's just not as pervasive as intels is yet.
Metamuscle.com - News in the Iro
And, like IBM, you are showing your true colors as a short-sighted moron. Remember, high-end is what drives the market. You're complaining about progress like it is a bad thing. Please remember that if it weren't for the constant improvement of PC hardware, things like the internet as we know it today and smart word-processing apps would still be a dream. You are wrong about performance not making a machine more useable. Having a higher performance processor allows the PC to better bridge the gap between what people want and what the machine can do. It allows us to design more sophisticated interfaces, eventually, even language based interfaces will become more common once the processing power is there. I also think that the idea of companies focusing on manufacturing rather than making faster processors displays a complete misunderstanding of the market. Most companies make the majority of their money from the high end. The low-end processors make very little. So, to tell a microprocessor company like intel to drop their cash cow and go for the low end is the stupidest idea I've heard in a long time. I think you are naive to say that people can't find uses for this processing power. I think games are an easy example, they easily can use up the CPU cycles on a processor. Now, I'm sure you're going to say that we don't need them. Who gives a shit? Quit trying to place some warped sense of puritanic philosophy on technology. I don't really care if I don't need it. I'm glad that I can do it. Getting away from consumer applications, the software that I write for a living models geographic data collected via sattelite to view in a 3D virtual world. WE could use that power and so can most of our customers. You are also making the mistake of viewing the PC as one appliance. It's NOT an appliance, it's a computer. It's flexable, that's what makes it so powerful, it can do many things very well, as long as the processing power is there. There will always be a need for more power, because, as PC's become more powerful, their function changes. They are able to morph into what we want them to be. If we have the power, we can use our PC as a vcr, a stereo, a telephone(over the net), these are all things that require power that didn't exist 5 years ago, and I am sure that as they get more powerful, their function will change yet again, to allow them to solve even more problems. If we limited this power, we would limit the amount of utility that we could get from these devices, which, in the long run, would hurt us.
Yeah well p4 isn't out yet so of course it's price will be alot higher
"This is where god would go if he wanted to get off blow!"
Actually Intel is planning on releasing the 1.4 at $625 base price...which is something like $20 more than the current price for 1.2 Tbird IIRC..
This isn't gonna be about price, its gonna be about 2 things. Availabilty (volume) and public perception (which Intel has a strong userbase backing because of its more public image and persistence in the market).
but its good to see competition in the market, We shouldn't be rooting for AMD or Intel, but rather both at the same time...technology advances with great competition
"This is where god would go if he wanted to get off blow!"
have AMD machines now and when the dual-Athlon DDR SMP motherboards come out (purr purr) I will be getting another one.
Have you heard of who is going to be offering one of these units? I am looking to do the same when its available and all this 'on again' - 'off again' confusion has me flustered - Does anyone know the "REAL" skinny on the Dual Athlon MoBos?
Screw 32 bits, that is SOOO 5 years ago.
:-)). The Amiga had 32 bit sw since the beginning, that must be about 15 years ago now.
In Homecomputers? Much longer. 10 years ago I wrote my masters thesis in TeX on my private Amiga with 32 bit sw and hw (and printed it on mainframes; There is even TeX for Crays
All right, so we have a bunch of benchmarks that test some games that we're all playing right now to prove that the Athalon is better than Intel. That may be true, but if I want to use my computer one year from now, I'd place my money on the Pentium 4.
I think we all know that a benchmark that simply focuses on one strength of a processor and doesn't reflect how fast your system will run in actual applications is useless. And yet, everyone seems to ignore where PCs are heading, and what sort of applications we will be using in the future.
What use are all of the Pentium IV's fancy multimedia extensions? Better multimedia. Many of you have probably downloaded a Simpsons episode off of Scour and watched it on your PC. I bet you're hoping that that 1 1/2" window will get bigger and bigger as the years go by. I bet we'll be streaming DVD's and watching HDTV off of the net very soon now. And with these high memory bandwidth requirements, the PIV will really shine.
For years, the slow memory system of a machine didn't matter. So long as you had a lightening fast cache you were fine. Most programs in use today obey the laws of locality of reference. But that's going to change in the future. We won't have tiled video game worlds, we'll have a single huge non-repeating world. We'll be streaming videos where the contents of the cache are used in a single frame, and then discarded in the next frame as the movie advances. I believe Intel saw this future, and so decided upon RDRAM with it's high latency but superior bandwidth.
Sure, if you want to crack RC-5, then a quad Athalon system is your best choice. But I think that the bulk of users will find a single CPU'd high bandwidth memory PIV to be their choice for the future.
I was reading reviews and the Thunderbird seems to use twice as much power as a PIII. I don't know if this holds true for the P4 also. I was looking into doing a small Beowulf cluster and was thinking about waiting for the SMP Athelon boards. I have my doubts now.
#define OH_YES_INDEED 1
int microsoftsuxbadly = OH_YES_INDEED;
Oh dear, poor little AC - feeling all insecure, are we?
Did the nasty man make you think there was something outside your US? There there...
You are right - the "whole silicon industry" is based in the US (well, apart from the large portion of it which is based outside the US, but don't you worry your pretty little head about that part - if you keep your head under the covers, maybe it'll go away?).
However, the WHOLE silicon industry (including the fraction which IS US-based) is completely metric and has been for goodness-knows how long.
So what was the point of your little tantrum, oh insecure one?
--
People should not be afraid of their governments - Governments should be afraid of their people.
with dual procescor celery300@450mhz beos itl'l top dual proc P4@1.5ghz win2k thx to BeOS threading EVERYTHING.
Hey, glitch! The Thunderbird 850 has a 200MHz bus! Sometime early 2001, AMD will be releasing a modified version of the Athlon that supports a 266MHz bus.
-atrowe: Card-carrying Mensa member. I have no toleranse for stupidity.
Actually, the majority of people DON'T respect Microsoft. It's just that they think they have no valid alternatives. And for a lot of them, it's true - it's gonna be a long while before my parents are using Linux, even though they despise windows. They don't know much, but they do get pissed when an hour of writing just dissapears because windows decided to take a dive right before they saved.
now matter how correct you are, the only reason people really need the speed is for gaming, and since a lot of smart computer people use AMD, i find it hard to believe games are gonna be designed specifically to work with p4. Therefore, I think in a way this is a very usefull benchmark tool...
I just like the fact that an AMD beat a p4 on a MMX test.
Um... maybe there's money to be made selling P4 heatsinks to car owners. ;)
--
A feeling of having made the same mistake before: Deja Foobar
Sounds like you should like your CDROM drive instead. I think that difference is far too big to be explained by processor differences alone.
Oops...
Last time I looked at developer.intel.com there was no sign of a P4, so I just assumed IA64 was it.
IIRC, the PPro did significantly worse than the Pentium in benchmarks for executing 16bit code. We'll start seeing representative benchmarks when Win64 shows up.
Yes a shorter trace helps increase clock. I was refering to this:
The problem is that it's a bit of a false gain. Most of the performance gained in clock speed is lost again to the serious hit you take at each branch misprediction. If you could keep your ultra-long pipe full, you'd be cruising, but you can't. Occasionally you will mispredict, and have to flush that pipe. One your pipe becomes as deep at the P4, that performance hit starts eating your lunch. Suddenly, most of your processor is sitting empty most of the time.
So, clock-for-clock P4's get slaughtered by Athlons or PIII's. But Intel doesn't care. They know that the majority of consumers buy based solely on that magical MHz/GHz number. Most consumers are not sophisticated enough to realize that there is more to performance to clock rate
Although the amiga was definitely 16 bit Acorn were turning out 32 bit Arm based machines at the same time. They only really caught on in the uk education world, which is a shame cos I think they were probably pretty good.
Yeah...I'm sure it'll only take a few weeks, right? And, BTW, I think that those optimizations are more for the SSE instructions than anything else.
I would personally rather see a review based on prices, not parts. Since most people have a budget, wouldn't it make more since to compare a $1500 Athlon system to a $1500 P3 to a $1500 P4?
I don't normally like to confuse people with facts they clearly are not capable of understanding but here goes.
Those P4s have a 400MHz FSB connected to twin RDRAM channels giving it a memory bandwidth of 3.2 Gb/s, compared to the athlon with a 200MHz FSB and its DDR RAM with a 2.1 Gb/s bandwidth.
The DDR RAM only slaughters the single channel RDRAM systems and the P3s with low FSBs.
Against the P4 it's more even (taking into account the latency problems of RDRAM).
grab your ankles bitch
He was talking about the memory speed not that of the FSB.
grab your ankles bitch
Actually only the French and other degenerate has-beens do that.
grab your ankles bitch
Direct X and virtually all graphics card drivers have heavy 3D Now optimisations.
grab your ankles bitch
"Just for example, changing our video drivers from DirectX 7.0/Non P4-optimized to DirectX 8.0/P4-optimized gained us roughly 40% in gaming performance in some situations. We imagine that within a few weeks when developers have had a chance to update their software, overall Pentium 4 performance will be leaps and bounds ahead of where it is now. "
So for God's sake, before going crazy over the fact that the P4 was outperformed by Athlons in a few benchmarks, use some sense. Let's see how programs re-compiled to take advantage of P4 optimizations fare in a head-to-head processor comparison.
---
To engineer a site for Nutscrape is as bad (if not worse) than designing for IE. All sites should be designed for W3C compliance.
It's much more that no killer app has emerged yet for this sort of power (excepting games). The Internet pushed people to (for want of a better term) 686 machines (PII, K6-II and -III, their PowerPC equivalents). Since then, nothing has emerged. The 'net isn't chewing up CPU cycles like it used to (and won't until Cable/DSL connections become significantly more common). When broadband internet becomes more prevalent, we'll see another boom in high-end sales (which have been in a slump for the last few years).
What makes you think that this is the last core that AMD is going to release? They have a 64 bit core in the works already. You say that "Intel is a very big company." So is IBM, and you see what a powerhouse IBM is in the PC industry now. It was people like you who said that Compaq was going to fade away because IBM was such a "big company."
Just look at the absurd blunders by Intel. They had the Pentium floating point bug and then tried to avoid replacing the chips. They put the serial number in the P3s, upsetting many consumers and privacy advocates. They released their 1.13ghz P3 and then had to recall it. They have been unable to deliver their high-clock-rate CPUs on time or in quantity. Their i820 chipset had a bug that required the replacement of thousands of motherboards from many different manufacturers. The 440BX chipset was the chipset of choice for the last couple of years because their i8xx chips sucked. Then they got in bed with the Rambus extortium, concentrating their efforts on a flawed, proprietary, overpriced technology.
So, don't be so quick to assume that AMD will not come out on top.
-- Fred Maxwell --
I'm amazed that the Athlon 1.2gHz so consistently outperforms the P4 1.5gHz... is this the sound of a monopoloy crumbling?
"The good thing about Alzheimer's is that you can hide your own Easter eggs."
"People should be allowed to keep midgets as pets."
- Gov. Jesse Ventura
C'mon what are Joe public gonna be swayed by.
The average PeeCee user buys Books like:
"Computers for dummies too stupid to understand a florida ballot paper"
They watch the Blue Man adverts and go "need intel peecee with windows, must have intel peecee"
and they'll wobble off to PC World and buy the fastest pentium IV they can find.
Are you sure the FPU runs at double clock? I would assume it's just the integer units... since the FUs are already pipelined anyway you wouldn't really gain too much...
I must burn in hell, suffer and pay for my sins
But Gods the one who's losing, Satan always wins!
there's a huge difference between optimizing for P6 vs Athlon and the P4... P4 has significantly different pipeline design and branch prediction than current market processors...
I must burn in hell, suffer and pay for my sins
But Gods the one who's losing, Satan always wins!
As the other person responded instructions per clock(cycle)... back in the bad old days, you would measure things in cycles per instruction... this is pre-superscalar, meaning you only could execute a maximum of one instruciton per cycle, and since you never could really achieve this, it took more than one cycle to execute an instruction... starting with the pentium(and I suppose the K6 but I don't really know) Intel had a superscalar design... the P5 could execute up to two instructions per cycle... I don't know what its actual IPC was, but it didn't even have a real branch predictor, so it was rather far from 2 I'm sure...
I must burn in hell, suffer and pay for my sins
But Gods the one who's losing, Satan always wins!
You're just forgetting 1 small thing..
The avrage joe - when he see's a commercial of a major brand selling PC, or an add of Dell - he see's quite clear the animated logo - "Intel Inside".
This little animated logo makes the differences for the avrage joe. He'll see in the commercial a flashy PC and he will goto the store and buy it - with this little tiny logo that he see's on TV.
Thats how you sell Intel PC's. Maybe AMD would do this some-day...
Hetz (Heunique)
Well. From what I learned back in grade school, "unavailable" means you can't get it. Period.
Having it priced beyond your means doesn't make it unavailable. A better term would have been "less readily available".
I DO agree with the basic sentiment of it though. The focusing on the best price/performance ratio rather than best price or highest rated speed. As a P4 system with the mandatory RAMBUS *GACK!*, easily outpaces an Athlon.....in terms of price.
Chas - The one, the only.
THANK GOD!!!
Chas - The one, the only.
THANK GOD!!!
Sure, Athlon kicks Intel around the block.
Doesn't matter. The media is frantically pumping Intel's "next big thing," to the exclusion of AMD's *already existing* big thing.
It's been most interesting to watch the mainstream (C|Net, ZDNews) suck it up to Intel. Two months ago, they were mentioning AMD about as much as Intel. Today, they talk exclusively about Intel's "to be" chips, and ignore AMDs existing superiority.
The winner in the chip wars will be the company that best manipulates the popular media.
Which means that Intel, no matter how badly it shoots itself in the foot with poor designs, poor performance and poor planning, will succeed -- because Intel is a Master of Marketing.
--
--
Don't like it? Respond with words, not karma.
There's no problem with defining time_t to be 64 bits on a 32-bit architecture. There can be a small performance penalty if it isn't done carefully, but that's it.
I thought my K6-2/350 was quite fast enough - then I installed the latest bloatware, i.e. Mozilla, ...
VMware is also quite CPU hungry - it's fine for most things, running Windows on Linux on this hardware with 256 MB RAM, but it could still do with a speed boost when booting (which doesn't run at full speed).
This isn't a trend I particularly like, but it seems to be happening in the Linux world as the GUI applications get better...
I figured I'd get my say in about why the p4 currently performs the way it does. All of the arguments I'll make here are based on data from Intel's various spec sheets and documentation on the Pentium4.
:-) ... I wouldn't be terribly surprised if a future p4 variant started making use of the huge number of transistors available on a 0.13 micron process to do some fairly massive optimization or redordering of code in the trace cache, similar to what Crusoe's code morpher does in software. Intel seems to have finally found a way to remove the "CISC penalty" of IA-32 code. I say good luck to them and AMD both, as long as they compete, we the consumer wins.
1. Why doesn't the P4 blow away the P3 on normal integer code, despite having double clocked ALUs?
Two reasons. First, is the issue that everyone keeps bringing up: mispredicted branches causing part of the pipeline to be flushed. At 20+ stages, there's a fairly large penalty for this. I don't think this is as big of a problem as most people have been led to believe. The branch predictor is 33% better than the one on the P3 according to intel, which helps a lot with this. With the use of a P4 aware compiler, branches are also laid out in such a way as to help out the static predictor (in the case the branch is not in the branch history table, or is unpredictable), and the P4 can use branch hints emitted by the compiler or assembly writer. So at least with newly compiled programs, branching shouldn't be a huge issue. With older programs, the improved branch predictor should help a lot. The real problem with the P4's integer performance (and why it performs dismally on RC5 is that shifts, rotates, and multiplies all have increased latencies (although the throughput remains either the same or very close per clock) compared to the p3. So code that expects to get the result of a shift or rotate back very soon is not gonna be happy when the p4 takes 4 cycles to do so. Small shifts can be replaced with a series of adds which have both higher throughput and 8x lower latency than shifts on the P4, but the compiler has to know about this in order to optimize it.
2. Why does the FPU perform so poorly in some things, and great in Quake3
My answer here would be that the FPU in the P4 is almost identical to the one in the P3 in terms of throughput, but the operations have a longer latency. So again, code that expects p3 or athlon latency instructions will get a rude awakening. Quake3 (and this is total speculation) probably uses a large unrolled loop of FPU ops that would hide the latency issue, but benefit from the same per-clock throughput as a p3...at 1.5x the clock speed. If you check out JC News (www.jc-news.com/pc) there's is someone claiming that Quake3 has no SSE/SSE2 optimizations whatsoever, so the FPU routines are probably just tuned to get as much throuput out of a pipelined FPU as possible. Use of the scalar SSE2 FPU ops on a P4 (by the compiler, or the assembly writer) can decrease latency while using the SIMD ops can increase FPU throughput. Obviously, Intel has decided to make some tradeoffs here. Shifts, rotates, multiplies are all slower, while add/sub/not/or/xor/and are all faster. The trace cache is perhaps the most interesting thing about the P4, as the x86 instruction set becomes a one-time cost for the core most of the time. It is in essence, a very simple code-morpher (x86 Ops -> cached uOps).
Speculation time
Digital video is going to be incredibly popular, as it gives consumers the ability to actually turn their home movies into something actually *watchable* (in essence, returning to to what 8mm film offered 50 years ago). Video will chew up CPU, memory, and disk for the forseeable future, thank you very much.
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)
Some software assumes time_t is int. If that were not the case, we could simply define time_t as unsigned int on 32-bit systems (provided systems don't use dates from before 1970) and continue along for an extra seven decades.
If we just make ints 64 bits we won't need to clean up such brokenness. Of course there is still 64-bit uncleanness (stuff that assumes sizeof(int) == 4), but that will have to be fixed regardless.
Reviews are all over the net now...here is some of them:
Anandtech
HardOCP - on HardOCP's frontpage you can find more links to reviews.
Toms Hardware hasn't got his review up yet, but I bet it will be soon...
Greetings Joergen
Well, well, well.
The Athlons rule in everything but raw FP speed, where the Intel parts actually win. That translates over to things like Quake 3 and Unreal Tournament runing faster on Intel.
However, the Intel parts have an unfair advantage there, being clocked at 1.4 and 1.5 GHz rather than 1.2.
THat really means Intel and AMD have equally capable FP units, maybe AMD has a slightly better one.
If money is no object and all you care about is FP speed, buy the Intel. Otherwise buy the Athlon.
Glückwünsche, haben Sie Slashdot ermordet, indem Sie zum korporativen Druck beugten und Subskriptionen einlei
> I'm a big believer that, public perception be damned, if your dealing in tech and you lose the respect of the technical community then, over time, you will lose the respect of the rest of the market.
:-).
That would explain why Microsoft is so close to bankruptcy
I have a P-II 500 at work, and a Athlon-600 at home. I use Grip and GoGo to rip mp3s. GoGo is essentially LAME with MMX optimization by some nice Japanese folks. At work, I get 4x rate on the mp3 creation, and at home I get 12x. All the time. That's a 300% difference for 20% clock speed difference. Foo! I like my Athlon.
________________________________________
Napster-to-go says "Fill and refill your compatible MP3 player", which is a lie. It's not MP3. It's WMA with DRM.
Ah! That makes sense. Well, 3DNow! beats the pants off MMX, then. Still like my Athlon! :)
________________________________________
Napster-to-go says "Fill and refill your compatible MP3 player", which is a lie. It's not MP3. It's WMA with DRM.
Don't forget that the P4 systems use RAMBUS as well, making the system even more expensive than the Athlon PC133 systems, and Athlon DDR is a big question mark as PC2600 seem even thinner on the ground than RAMBUS right now, hopefully the manufacturers will start to ramp up production soon, then we might see PC2600 on Pricewatch etc.
Any sufficiently advanced man is indistinguishable from God
I seem to recall from some class that, when you have a deep pipeline, there are benefits from being able to make instructions "conditional". That is, given a conditional branch over only a couple of instructions, it is more effective if you can make the test, set some bit in a register, and have the next two instructions execute or turn into a NOP based on that bit. The two NOPs cost less in performance than the hit you take when the branch prediction fails and the pipeline gets you in trouble. This is also an alternative (don't remember if it's all that good) to spending lots of transisters on branch prediction.
Anything like this make it into the P4? You'd need P4-optimized code, but hey, that's one of the things I like about running Open Source stuff...
MB's can be very expensive.. Especially when they're in low availability. When the Athalon came out, the MB's were almost as expensive as the chips.. Throw in the fact that they were buggy (since it was all brand new.. including the EV6 connectors). The P4 requires an all new powering structure, so it's possible that we're going to find MB's that blow up the CPU's.. Thus you might want to wait a few rounds before considering these guys (this is usually a good rule of thumb anyway).
Given that it's brand new, you're also going to be at the mercy of the MB manufacturers in terms of features.. If you like ultra-SCSI-RAID on your MB, it might be a while. There is a similar argument to be made for your case.
Expect to pay premium dollars for this combo.
The problem with the head-sink is logistics and ergonomics.. Where do you put everything? Unless you have a monster tower case, you're not going to be able to fit too many full lenght boards.. You have a massively over-heated PAIR of RDRAM RIMMS, a MASSIVE CPU, a larger than normal power-supply.. And then you get to have your first AGP card. A Mid-Tower is probably too small for these guys.. That might not be a problem for some.. But personally, I like stacking my computers, so it's a problem for people like me.
The heat-issue also means real-estate for lots of cooling.. And an increased risk of heart-failure.. So to speak.
My guess is that you're looking at $300 - $400 for the MB + case alone. That's more than I usually spend on an entire bare-bones system. Throw in roughly $600 for each of CPU and memory, and you've got a rather large handi-cap for an almost miniscule performance gain.
As for the pricing of their CPUs... To be fair, they _have_ to depretiate their costs.. It's not cheap to design and entire CPU from scratch.. Remember, that this is their FIRST CPU redisign since the pentium Pro some 5 years ago!! Everything else (with the exception of the vapor-ware Italium) has been an add-on to their old archetecture.
I don't know what their margins are per CPU - I wouldn't be surprised if it was 100% - But when you take in the cost of multiple billions of dollars, it's not all money in the bank. AT&T use to depretiate their hardware-costs over 40 years.. Intel doesn't have such a luxury.
-Michael
-Michael
Most of the RISC/UNIX camp are already in 64 bit land, having just emerged from several years of teething pains.
Check out:
In many cases the chip hardware has reached 64 bit before the OS has. Most of this development has been below the radar in the mainstream press because the solutions are not Wintel.
Extrapolating the story here suggests the Transmeta approach of using VLIW to emulate lower bit width processors can help to compensate for the slow rate of change in OS.
Also, it suggests there might be practical merit in the AMD K8 approach of abusing 64 bits for double 32 bit processing.
"Provided by the management for your protection."
Personally, I think the average Joe has no need for all this high-power computing. What it really comes down to is that the average computer illiterate only uses the computer to check email, surf the web, play games, and maybe use apps like Word (I know, but they usually don't get Linux). And all they see is the clock speed, true, but then again they can never really use much of the processing power.
;P )
But the whole performance rating really doesn't matter much once you get past the point of human usability. I mean, unless you're running more programs than you can possibly use at once, who can truely use all that computing power personally?
Both companies should just forget pushing processors farther and faster, and go for manufacturing. Once they can mass produce those processors, the whole world can have a computer. (Hence, the government has more to worry about
Your numerical program in all likelihood take advantage of neither.
Also don't forget that Intel has far greater resources to make sure all the compilers, etc are fully optimized.
Actually i just found an article on ZDnet really panning the P4. Here/A& gt;. Although at the end he said he'd by one just for unreal tournament (800 pounds of chip just for one game) and he's wrong, Athlon 1.2 is faster at UT, but the P4 wins at Quake III (but only with new nvdia drivers).
Intel's horn has been honked by the marketing & management types now for a while. This is only the latest in the series of disasters that they have come up with. First Rambust, then the P4, then the pricing on the celeron series chips. The pricing here was dictated by the requirement to keep the average ASP up rather than sell chips. They have all but totally ceeded the low and mid market to AMD, and now it looks like then high end is going to go bust too. Quite frankly they deserve it.
Windows already has a 32 bit OS they're making for compaq in preperation for intels release of their 64 bit chip ...
Ignore the "p2p is theft" trolls, they're just uninformed
Hmm... lets see. Comparison of Z80 and P3 at 1 MHz.
P3 has two 32-bit wide ALUs (arithmetic-logical units) and one floating point unit. Each one of them can do one operation in one clock cycle (excluding multiply, divide, legacy instructions and branches), therefore in theory being able to execute 2 instructions per clock (in perfect conditions, perfect pipelining (no mem/reg dependencies) and no cache misses, page faults or interrupts).
Z80 has just one 8-bit wide ALU and no floating point OR multiply/divide instruction (and yes, it has two 16 bit registers too, but still it's an 8-bit system internally). Z80 is not pipelined, so it has to spend one (two?) complete cycles just for memory access and then two cycles for execution (minimum cycle time being 4 cycles).
Even in basic situations, at 1MHz, P3 would win Z80 by 800%. But in general each P3 instruction does a *lot* more useful work than a Z80 instruction. If you code multiply routine in Z80, it probably needs at least C+N*8 to C+N*16 cycles (C = some initialization N=number of bits), probably more. Original Pentium needs 9 or 11 cycles for multiply (AMD K5, K6 and K7 have troughput of one cycle per multiply IIRC). N being usually at least 16 and C maybe 32, Z80 would probably be able to do one 16-bit multiply in 200-400 cycles, all modern processors would be a *lot* ahead in this case. Even in case of general code, 1MHz P3 would probably come ahead some 2000%, at least 1000% faster, than 1 MHz Z80.
But I wouldn't be surprised if Z80 performance per transistor per MHz would be faster than P3's counterpart, though...
K6s and Athlons have AMD's counterpart of MMX called 3DNow!, in addition to MMX. Apparently 3DNow! easily outperforms MMX but only a few applications support it. However, one of these happens to be Gogo (which also supports MMX). So, you're not comparing the MMX unit of the two processors. I've noticed the huge advantage of K6 over Pentium when using Gogo, and it looks like it is because of 3DNow! vs. MMX.
--
Escher was the first MC and Giger invented the HR department.
No vendor is going to go to this trouble unless you are buying at least a hundred machines.
Cheapest P4 1.4 Ghz =$920US
Cheapest Thunderbird 1.2 Ghz = $488US.
Sounds like a signifigant difference to me.
-atrowe: Card-carrying Mensa member. I have no toleranse for stupidity.
Still Wrong.
-atrowe: Card-carrying Mensa member. I have no toleranse for stupidity.
It looks to me like the floating point performance was the main thing hampering the P4's. As a non-gamer who doesn't do much graphics, but does regularly use the full integer capacity of my computer, I think the P4 was significantly faster. I think they did a lot of work to reduce branch misses, and it paid off. I'd also like to see the further improvement when (if?) we get a mature chipset using RDRAM, as it is a much faster breed of memory. There are issues with it, but because of the immature hardware to work with it, we've only seen the bad side so far, not the good.
If you are modding me down because you disagree with me, use the "Flamebait" category, not the "Troll" one.
64 may seem like too high end and too far down the road, from the present, but the fact is 64 bit systems will be common place, probably in as soon as 3 years. They already exist in great numbers in engineering workstations and servers, from a variety of vendors.
--
A feeling of having made the same mistake before: Deja Foobar
Intel Set to Unveil Next-Generation, Speedy Pentium 4
By Duncan Martell
SAN FRANCISCO (Reuters) - About every five years, Intel Corp. (NasdaqNM:INTC - news), the world's No. 1 chip maker, undertakes a mammoth transition.
After viewing the benchmarks, the world mammoth does come to mind. Old, huge, slow (well, slower than a sabretoothed tiger)
Granted these benchmarks are not ground to the finest tuning on either platform, they benchmark _is_ fair.
Why? Because when I acquire software for an x86 platform I'm not getting something tuned specifically for that processor, that memory, those controllers, all I get is an approximation that's "good enough" Therefor, unless there was some specific nefarious activity to downgrade the P4 results Intel is producing a throwback. Should the press be so unkind as to publish Athlon DDR benches and translate it into laymans terms, Intel could be seeing the next big hit on their credibility.
Chances are, Intel will escape unscathed, as the press are either ignorant savages or too afraid to stake their names to a story to rain on Intel's mediablitz parade. For the New York times to proclaim "so what?" at the bottom of Page 1, would be removing the crumbling keystone from Intel and sending them into the "get serious" restructuring they are so badly in need of. Anyone with doubts need only look at the Merced project to see where a Bay of Pigs mentality took root and manifested itself at Intel.
--
A feeling of having made the same mistake before: Deja Foobar
Begining:After viewing the benchmarks, the world (typo included) are my own words. Sorry about that.
--
A feeling of having made the same mistake before: Deja Foobar
This depends on the compiler. New versions from Microsoft or Borland should have this option.
Keep in mind that C|Net is paritally owned by Intel, and ZDNet was recently acquired by C|Net. No wonder they "suck up" to Intel.
I'm not a very demanding fellow. I just want 64-bit systems to be everywhere before 32-bit time_t overflows in 2039.
Personally, I think it's gonna be tight. We've been hearing about 64-bit for a long time now and yet most of us are still stuck with 32-bit.
And I really hope MS moves their OS from 32->64 in less time than it took them to go 16->32. Wasn't the 80386 released some time around 1987? Past experience suggests that the "fully 64-bit" Windows 2015 will still run some 32-bit code under the hood.
There are a small handful of folks there who do benchmarks I trust
(eg. C't and Ars Technica). It isn't always so easy to do your own
benchmarks since one needs access to all the sets of hardware you want
to compare.
I don't know if I agree with this.. These benchmarks show that Intel's 1.5 is roughly equivalent to AMD's 1.2, and we were told to expect this by the previewers.
.13u before Intel), then they can start pulling away.
.13u is going to be hard to beat. There are no benchmarks for AMD's sledgehammer, so there's no point in speculating about it.
I believe AMD uses a fully pipelined FPU (multiple ones at that). I'm sure that the P4 also fully pipelined their FPU, BUT, they added several additional stages to the basic instruction as well.
The addition of stages does two things: first it increases your max clock rate all-else-held-equal. And second it increases the latency for missed branch predictions. Again, all-things-being-equal, the missed branchs will tend to hurt more than the higher clock helps, except in a few special cases.
Intel, therefore included with those extra stages a highly advanced branch predictor that lives in the MIDDLE of the pipe. More-so, successfully predicted branches can skip the first several stages thanks to the branch cache. Thus, well-behaved code will get the benifit of heavier pipelining with fewer of the pit-falls. To make things even more tantelizing, they're using a 2x clocked Integer Unit. Thus they have a 3GHZ integer unit on these benchmarks. That means that they can further extend their pipelines with almost no visible penalty (even in branch-misses).
Unfortunately, they still seem to be plagued with branch misses (the only logical explanation for why AMD can still keep up or even surpase them). Obviously the memory played an important role in these benchmarks.. The KT133 v.s. AMD560 really only differ in memory speed, and that was enough to sway several percentage points. A more fair comparison would be between VIA's up and comming DDR-SDRAM P4 chipset.
But, as was pointed out; if Intel can get the P4 up to 2GHZ before AMD can (last rumor I heard was that AMD was going to hit
Unfortunately, as several sites pointed out, buyers don't look at benchmarks, they look at CPU speed, so Intel should be able to wrongfully win people over on this synthetic basis. Thankfully, the only people that are going to be willing to buy P4's are people needing servers (or maybe even Q3). We'd have to see NT ASP/Sql Server and or Linux+Apache+PHP+Oracle, etc to determine who's king (including memory types). Unfortunately I rarely see benchmarks on these grounds.
Sooner or later AMD is going to come out with their 64bit proc. With Mustang gone, this is their only next-great-hope. An all new design - hopefully without a tremendous cost - that has started from scratch (as the P4 did). I'm sure it too will have a heavy pipeline, but several of it's new features (such as the flat-memory archetecture) should enhance the playing field.
By then, however, the P4 will have found a new chipset that handles DDR-SDRAM and will have enough volume MB's and cases that it'll be cheap enough for the hard core gamer and possibly even casual gamer to purchase. A 2 - 3GHZ processor running at
I totally agree with another poster that said we should be rooting for both Intel AND AMD since competition is good.
-Michael
-Michael
Intel has a good public perception, but Rambus has a hideous one. I would think that until the SDRAM boards come out for the P4, the fact that it requires dual channel Rambus is going to hurt it somewhat.
(thats another thing that I find interesting, that they need to use a *dual* channel 800mhz rambus setup to be able to compete witha single channel sdram setup. Thats pathetic.)
-- "So they told me that using the download page to download something was not something they anticipated." - Bill Gates
The more benchmarks I read, the more I steadfastly believe that you must test yourself. Benchmarks don't mean squat if the machine sucks running YOUR app. Get your vendor to supply 2 test machines, and then pick the better one for you. Oh, and this also solves that "other" problem. The vendors can only get you machines THAT THEY HAVE. Machines that aren't "really" avaialble to the rest of the public for another 6 months don't mean CRAP to the guy that needs his system now.
"Science is about ego as much as it is about discovery and truth " - I said it, so sue me.
Here's a question I've got.
I have AMD machines now and when the dual-Athlon DDR SMP motherboards come out (purr purr) I will be getting another one.
A lot of the benchmarks, etc. claim that certain things are bettered by optimisations, saying that recompiling or rebuilding with P4 or Athlon optimising in place will radically change the numbers.
So for the Linux/FreeBSD crowd in the know: given that we rebuild kernels, what are going to be the chances that gcc and/or buildscripts are going to support/offer optimisations for either the P4 or the Athlon? I think the PentiumGCC people are working on K6/Pentium optimisation, any chance of it going further?
I'd hate to think that but for the want of code optimisation options for my silicon, I'd be unable to take full advantage...
--- Jump!! Fire!! Bullet time!! - Lego version of the Matrix
So what you're saying is that if I could buy the P4, It would cost twice as much as an Athlon that can beat the hell out of it in most performance benchmarks.
Thanks for clearing that up for me.
-atrowe: Card-carrying Mensa member. I have no toleranse for stupidity.
(And I was getting worried...)
For example, if you compared a Z80 running at 1MHz with a P3 running at 1MHz, you would find that the Z*) does much more work.
KTB:Lover, Poet, Artiste, Aesthete, Programmer.
KTB:Lover, Poet, Artiste, Aesthete, Programmer.
There is no
Thanks for any information!
Ah well. Right now I'm really waiting for Itanium anyway. Once that comes out, I'm hoping someone'll do a price/performance comparason of the assorted 64 bit chips on the market that will run Linux. I'm also hoping it'll push 64 bit processor prices down a bit. I'll happily go for whoever offers me the most bang for my buck.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
As the GamePC article points out, a comparison between the Athon's and the P4's cannot be "apples to apples" at present. The P4's have significant hardware changes (such as a larger pipeline) that present software doesn't take advantage of. Since the P4 is "serialized" it really shouldn't come as a shock that the Athlon's performed better in this test. Let's not jump the gun here. Do another test in a few months and see what happens. Besides that, I don't really think game playing is the best benchmark that could be done. Most of the benchmarks involve 3D graphics. Quite frankly, with a processor like these, there are much better ways to use those cycles that would be more indicative of their power.
"I believe that a scientist looking at nonscientific problems is just as dumb as the next guy." -Richard Feynman
And people seem really happy with AMD, sure there was a minor flap when they were accused of covering up a bug with their chipset not doing full AGP speeds with Nvidia boards, but over all I see people who know what they are doing with hardware drooling over AMD and raising thier eyebrows at Intel.
Until now, it's been kind of hard to tell someone who doesn't understand technology why they should like AMD (except for the price). Not every one is willing to listen to a lecture about the evils of unfair patent law and the whole Rambus affair. Not everybody can even understand or care, how schizophrenic Intel as a company has become, shipping chips with no decent chipset support (i820 anyone?), announcing releases of high-speed chips that they can't supply in any reasonable quantity, and ignoring the needs of not only large accounts (I worked at a school board last year and had to fight tooth an nail to gaurantee a supply of celerons for student workstations) but also niche accounts that hold the strength of their image in their hands.
I'm a big believer that, public perception be damned, if your deailing in tech and you lose the respect of the technical community then, over time, you will lose the respect of the rest of the market.
I don't think the public has had much bad intel publicity that they can fully understand. However, I think that Intel's move to increase clock speed at the expense of performance will ultimately have a negative effect across all segments of the market.
What is the message going to be from you people when you are asked about which computer people should buy? It will be, "yeah you could get the intel system, but it's actually way slower than the AMD that costs a lot less too." And the psuedo-experts who read the free computer monthly will pick the argument up and spread it even further.
And then Joe Lunchpail goes to work and tells his buddies, "yeah, I never heard of this AMD, but I guess they are making faster computers than intel, even though intel says they're faster, so that's what I bought."
And this kernel of information, meme if you must, will start to weaken the Intel brand and the public's perception of Mhz. It isn't hard to understand. Faster clock speeds are just for marketing. Even John Dvorak could bold that entire line in his zdnet column.
And for all the PowerPC zeolots, give it up, you can't actually buy those chips yet either.
I'm not saying that this issue will kill Intel, but it will damage them. It is a short-sighted and ignorant move by their marketing department, how, like most marketing departments, overestimates word-of-mouth when it's in their favour, and underestimates it when it's potentially negative.
Metamuscle.com - News in the Iro
What I want to know is this: how will these two processors perform against each other in dual-processor configurations?
The answer is that the Pentium 4s were designed to not be SMP capable, while the Athlons will be using the same SMP architecture that is used currently on DEC Alpha systems, which means that each processor has two dedicated connections to the North Bridge of the motherboard, as opposed to Intel's Xeon SMP configurations, which require all the processors to share bandwidth to the North Bridge.
Friends don't let friends use multiple inheritance.
It seems as though the athlon processor kicked Intel's butt according to the numbers, but there are some other things that put athlon way ahead of the p4....for instance... PRICE! athlon chips are a lot cheaper than these P4s.....also...the p4s require you to buy an entire new system...new mobo, new powersupply, new case, and a new 454 gram heatsink (454grams is about a 1 pound). I think that if these things were added in, there is no way that anyone in their right mind would take a P4 over the AMD chips.
The anti-salmon
I hope you realize that this is a fundimental CPU design. If AMD wants the Athlon to go beyond 2GHz they will have to make a deeoper pipeline. The Athlon has already incresed the pipe from ten stages to twelve. If any CPU maker decides to make a fast CPU it must have something to feed it data. Pipelining is the way bot AMD and Intel solve this problem.
A lot of the discrepancy between GHz and performance seen in P4 chips is explainable by Intel's choice of pipeline design. Intel chose to make extraordinarily deep pipelines on the P4 chips, which allows them to crank the clock speed up up up. Sounds great, huh?
The problem is that it's a bit of a false gain. Most of the performance gained in clock speed is lost again to the serious hit you take at each branch misprediction. If you could keep your ultra-long pipe full, you'd be cruising, but you can't. Occasionally you will mispredict, and have to flush that pipe. One your pipe becomes as deep at the P4, that performance hit starts eating your lunch. Suddenly, most of your processor is sitting empty most of the time.
So, clock-for-clock P4's get slaughtered by Athlons or PIII's. But Intel doesn't care. They know that the majority of consumers buy based solely on that magical MHz/GHz number. Most consumers are not sophisticated enough to realize that there is more to performance to clock rate.
This move on Intel's part was motivated by marketing rather than management. They are playing on the uneducated masses. It is all but directly deceptive, and I hope they get their clock cleaned by the press for it.
Buyer beware.
--Lenny
I hope you realize that this is a fundimental CPU design. If AMD wants the Athlon to go beyond 2GHz they will have to make a deeoper pipeline.
:).
...Or move to a finer linewidth.
Pipelining lets you increase the clock rate at a given linewidth. It isn't a requirement for faster clock rates in general.
Sometimes it's a good idea to use a larger pipeline, and sometimes not. For a given linewidth, increasing the pipeline depth will increase the clock speed at the expense of mispredict penalty and hardware complexity and timing sensitivity. Sometimes the increase in clock speed is enough to offset the disadvantages, but beyond a certain point, a deeper pipeline makes performance _worse_. What pipeline depth makes sense depends on your branch predictor, your cache, and a few other things.
However, shrinking linewidth will always let you increase clock speed, regardless of pipeline depth. It gives a straight factor-of-x speedup of all logic, no matter how the pipeline of the chip is set up.
I've been studying this for five years, so I have a good idea of what the tradeoffs are
Its all great to look at benchmarks but a chip that is unavailable scores a 0 each time you test it. There's two reasons a chip is usually unavailable. It is priced well beyond reach and reason or it is being produced in such low quantities that they might as well not make it
Since the release of the Athlon, AMD's chips are more readily available at higher clock speeds. Right now, there are 4 full pages of vendors selling the AMD 1.1 Ghz Thunderbird. The chip sells for as low as $341. That is an available chip.
Intel's 1.4 and 1.5 Ghz chips are available from 8 vendors and will cost you between $950 and $1100. In my book that chip is not available.
Pentium 4 1.5 GHz: $1099 (Pricewatch)
AMD T-Bird 1.2 GHz: $488 (Pricewatch)
Marketing to convince consumers that Pentium 4 is faster: $4 million
Look on Intel managers' face after seeing sales statistics: Priceless
------------------
A picture is worth 500 DWORDS.