is this really true?

[mark2k]

that xp chips are just good for moneybut p4s are better? i always thought that a xo 200 was better than p4 2.0 i always that at same speed ath was better.=(

[mark2k]

helllo>?

[eh?]

they are about even in raw performance, however the p4 does oc higher. and amd is just put its .13 micron technolgy.

[bigfoot]

that xp chips are just good for moneybut p4s are better? i always thought that a xo 200 was better than p4 2.0 i always that at same speed ath was better.=(

NO. think of it this way, AMD is now starting to make there processors on a .13 micron process. There the T-bred Athlons and they have recently been showen to reach almost 2.3 Ghz, now Intel's processors top out at around 3Ghz. BUT AMD processors process 9 things per Cycle, Intel's only can do 6 per cycle, so...

AMD: 2300x9=20,700
Intel: 3000x6=18,000

This gose to show that SPEED still means nothing. AMD has Intel beaten.

[Yodums]

Originally posted by bigfoot


NO. think of it this way, AMD is now starting to make there processors on a .13 micron process. There the T-bred Athlons and they have recently been showen to reach almost 2.3 Ghz, now Intel's processors top out at around 3Ghz. BUT AMD processors process 9 things per Cycle, Intel's only can do 6 per cycle, so...

AMD: 2300x9=20,700
Intel: 3000x6=18,000

This goes to show that SPEED still means nothing. AND has Intel beaten.

Cycles aren't everything to a chip. Your not going to tell me 700mhz does nothing in performance. Yes AMD may do more IPC but why does the Northwood score so high in benchmarks, and why are all scores at MadOnion are powered by Northwoods?

It's really hard to compare the both chips in architecture since they're made up of two completely different ones.

The only thing I'd recommend AMD at the moment is for Folding at the moment.

[bigfoot]

Maybe intel has a thing or two with branch prediction and it helping in 3D gameing but like you just said, It really sucks when you actually do REAL processing like folding!

EDIT: I almost forgot the most important reason, Memory, just wait till the nForce 2 chipset comes out... Intel crushing memory scores:D

[bigfoot]

Oh yea.. have you seen any of those Mpeg compressing bechmarks where an athlon would kick a P4s' ass without its SSE2 instructions enabled, even with it though, the Athlon still beats it a much much lower clock speed.

[Mr. $T$]

I love competition Because it brings the best out of a company like Intel or AMD. If AMD was not there what would push Intel to 3Ghz and the oppsite way around, I just like getting fast chips at a good price. ex: pII $700 486DX in 1993 $280 now a day AMD pally 1600+ is $70-100 and a 2.0 Intel is $100-120, Gotta love It!!!!

[takiwa]

Cycles aren't everything to a chip. Your not going to tell me 700mhz does nothing in performance. Yes AMD may do more IPC but why does the Northwood score so high in benchmarks, and why are all scores at MadOnion are powered by Northwoods? RAMBUS, young jedi....RAMBUS.

The sad thing about cpu's is that you can't test a chip without it being independant of the chipset and memory. Overall, I do think AMD's are a harder working processor...even on DDR they have shown that they can hold their own against a P4 on a RAMBUS board (which is a pathetic attempt from Intel, IMHO)

[Yodums]

Originally posted by bigfoot
Oh yea.. have you seen any of those Mpeg4 compressing bechmarks where an athlon would kick a P4s' ass without its SSE2 instructions enabled, even with it though, the Athlon still beats it a much much lower clock speed.

Source? If it's THG, I'll faint because he had all the P4 win everything but like 4.

[DaddyB]

the athlons definetly do more work per clock then a P4 and the best way to compare a P4 to an athlon is to compare the P4 clock speed against the XP PR rating. so yes the xp2000 is about the same level of performance as a P4 2.0, however those 2 processors will each win certain benchmarks. so an XP2000 will beat the P4 2.0 in 3dmark but the P4 will beat the XP office apps etc etc.... they are not identicle, they are similar in performance.

as for that whole 9 IPC VS 6 IPC thing its just not true, i mean yes the xp can do 9 IPC but it doesnt do 9 IPC for every cycle and the P4 doesnt do 6 for every cycle. It depends on how the code is written and processed.

If we compare an xp2000 to a P4 using the IPC then:

P4 2.0 is 6x2000=12,000
XP2000 is 9x1667=15,003

yet we all know that the 2 processors perform very similarly, the P4 beats the XP in some benchmarks. but according to the numbers above the xp2000 should do more work and win every benchmark. the IPC is more of a theoretical maximum, so neither processor does the maximum for every cycle.

[james.miller]

actually the xp2000 should stomp on a 2gig p4
this is my xp1800 on 256mb sdram @ 133

[james.miller]

and thats a 500mhz difference. now i know that this is only one benchmark so not totally accurate.but it is an indication.
now how would a 2gig xp compare to a 2gig p4?

[DaddyB]

exactly my point, it "should" but it doesn't. Sandra code is written to take full advantage of both intel's and AMD's procs, so it make use of all of the IPC for both procs. real world performance does not reflect the results that sandra gives b/c real applications do not make full use of all of the IPC for either proc. If you look at the benchmarks below you will see that the P4 2.0 beats the xp2200 in some benchmarks, clearly indicating that XP doesnt make full use of all 9 IPC for every cycle.

anandtechs benchmarks (http://www.anandtech.com/cpu/showdoc.html?i=1635&p=7)

Tom's review (http://www6.tomshardware.com/cpu/02q2/020610/thoroughbred-20.html)

as far as a P4 2.0 Vs a 2000mhz XP , yes the xp @ 2000mhz would win most of the benchmarks but thats not what I said. I was talking about the xp2000 and the P4 2.0, which have very similar performance.

[james.miller]

how on earth does the xp2200
run at 2gig standard when the 2100 is 1.73?

the xp2200's real STANDARD running speed is 1.80gig, not 2gig

http://www.athlonxp.com/modules.php?op=modload&name=Diner_Wrapper&file=index&req=ShowFile&file_wrap=html/reviews_2200.html

[DaddyB]

sorry, I just got up a while ago and im not even done first cup of coffee yet...

that doesnt affect the point I was making though. The P4 2.0 and xp2000 perform about the same regarless of mhz, IPC or Sandra scores, each will win certain benches but on the whole perform about the same.

[james.miller]

lol dont worry about it.that was kinda my point aswell. thats why im interested to find out just how well a 2gig xp will perform.its should be right up there with the 2.5p4's at least

[james.miller]

well, i think

[james.miller]

ive been up for about 12 hours now.im getting that way already!

[Emericana]

it is a tough question. my cpu that i usually run a little over 1.8ghz is at par with the 2.4ghz Pentium 4 in sandra results.

mhz for mhz amd destroys intel.
just look at a 1900+ vs a 1.6a or a 2200+ vs a 1.8a

however, i do not care what you guys say, if you have a 3ghz pentium 4 overclocked from around a 2.26 chip then the p4 will win in all benchies vs a 2ghz AMD. Once amds start being able to clock to around 2.4ghz with watercooling or worse, that will be a whole different story.

best bang for the buck is amd... just look at the 1600+.

in my opinon there is no reason not to get amd. nobody can telll the difference between a 2200+ and a 3ghz pentium 4. what do you use your computer for benchmarks?

cpus have become so fast that the videocards are what are holding them back not the actual cpu. with geforce 4's you can play almost any game with a smooth frame rate at 1600x1200 with aa on.

what more could you want?

[james.miller]

agreed.soon there will be no reason to buy pcs on the basis of speed.i think in time quality of the display and the components (sound cards,printers,scanners ect.) will be more of an issue than how many fps you can squeese out of quake3

[Yodums]

Originally posted by james.miller
actually the xp2000 should stomp on a 2gig p4
this is my xp1800 on 256mb sdram @ 133

Your comparing against a Williamette (First generation) Pentium 4, which was a horrible chip. Compare to the Northwoods. I wouldn't necessarly compare at stock because the Northwood are popular for their ability to overclock so much.

[james.miller]

but we are talking about stock speed here

[Yodums]

Originally posted by Emericana
it is a tough question. my cpu that i usually run a little over 1.8ghz is at par with the 2.4ghz Pentium 4 in sandra results.

mhz for mhz amd destroys intel.
just look at a 1900+ vs a 1.6a or a 2200+ vs a 1.8a

however, i do not care what you guys say, if you have a 3ghz pentium 4 overclocked from around a 2.26 chip then the p4 will win in all benchies vs a 2ghz AMD. Once amds start being able to clock to around 2.4ghz with watercooling or worse, that will be a whole different story.

best bang for the buck is amd... just look at the 1600+.

in my opinon there is no reason not to get amd. nobody can telll the difference between a 2200+ and a 3ghz pentium 4. what do you use your computer for benchmarks?

cpus have become so fast that the videocards are what are holding them back not the actual cpu. with geforce 4's you can play almost any game with a smooth frame rate at 1600x1200 with aa on.

what more could you want?

Again, your going to overclock your chip right? People wanted the 1.6a since they could get nearly 2.6gigs on good components + cooling of course. The way your looking at the best bang of the buck is for stock, which in this case it shouldn't be since your in an overclocking community.

Not necessarly do we do this for benchmarks, this is a hobby for most of us.

[Yodums]

Originally posted by james.miller
but we are talking about stock speed here


What's the point of that? Not many people run at stock here.

[james.miller]

why would anybody overclock if it wasnt for benchmarks.
theres no difference between 80 and 160fps besides the obvious, right?

well then.why overclock at all unless we benchmark it, because i for one cant tell the difference in the speed.can u?

[Yodums]

Originally posted by james.miller
why would anybody overclock if it wasnt for benchmarks.
theres no difference between 80 and 160fps besides the obvious, right?

well then.why overclock at all unless we benchmark it, because i for one cant tell the difference in the speed.can u?

Yes I can.

Stock running SETI: Average 3:45 hours for a 0.417 AR Work Unit

Overclocked @ 1.7gigs: Average 3:12 hours for a 0.417 AR Work Unit

It's a hobby, you do lots of things that have no beneficial results. Games are getting very detailed and in SOF2 w/ an ATI Radeon 8500 128mb DDR I can drop to 35-40fps on a very detailed place.

[james.miller]

erm.thats still a 'benchmark' isnt it? seeing how fast you can get a certain task done? am i wrong?

[Yodums]

Originally posted by james.miller
erm.thats still a 'benchmark' isnt it? seeing how fast you can get a certain task done? am i wrong?

Using anything to compare in between one another is a benchmark. . . I gave you games and SETI. You can also use Folding as an example. They use 100% CPU Power.

[CharlesHF]

Yes........join the Fold!!!!!! :D

[DogfooD]

Even before I unlocked my 1700+ Sisoft Sandra rated my machine the equivalant of a p4 2.0. Now it's cooler and OC'd.... hmm, wonder what it compares to now????

AMD = good ( I Picture AMD being a group of people sitting around a hookah thinking up ideas just to **** off the man )
INTEL = bad ( Intel is like the feds. Just people I don't want around my house period. )

[DaddyB]

I was just trying to answer the question that was originally asked. which was, as i understood it, is the xp2000 better then the P4 2.0 and i'd say they are almost the same.

If you're talking about overclocking then the P4 1.6A can OC to 2.6ghz or more (keep in mind this was intel's cheapest and slowest chip and was recently discontinued) and the xp2200 can OC to 2.2ghz or so, a pr rating of about 2800+... so intel's worst chip OC'd can perform almost as well as AMD's best chip OC'd. also note that the P4 can do it on air but the XP will need water cooling or better.

[james.miller]

Originally posted by Yodums


Not necessarly do we do this for benchmarks, this is a hobby for most of us.

so what does this mean then? does this not aply to you afterall.seeing as you just gave me some benchmarks

[james.miller]

anyway im not about to start an argument dont worry.

at stock.yes,i believe it does compare very well

[Yodums]

Originally posted by james.miller


so what does this mean then? does this not aply to you afterall.seeing as you just gave me some benchmarks

I'm saying that I care about the actual computing and stuff more than I care about results of a benchmark.

[james.miller]

Originally posted by Yodums


Yes I can.

Stock running SETI: Average 3:45 hours for a 0.417 AR Work Unit

Overclocked @ 1.7gigs: Average 3:12 hours for a 0.417 AR Work Unit



oh my word! computing stuff,games whatever.
it doesnt matter what you call it. what you stated here is quite clearly a benchmark.

it can do THIS much at THIS speed,but does even more at THIS speed

that is a benchmark.which you said you wasnt interested in.

[james.miller]

anyway like i said im not here to argue. im here to (hopefully) help other and get some help myself, and thats it

[Yodums]

Originally posted by james.miller


oh my word! computing stuff,games whatever.
it doesnt matter what you call it. what you stated here is quite clearly a benchmark.

it can do THIS much at THIS speed,but does even more at THIS speed

that is a benchmark.which you said you wasnt interested in.

No this is my point of view. I'm interested in SETI. And being interested, I would like to crunch WU faster, so I overclock it to be able to get it faster. I don't necessarly look at SETI as an overall benchmark, but I just gave you an example. I think of SETI is to find Extra Terrestrial Intelligence and explore the universe, again I just used it as an example.

[james.miller]

yeah i know.but you still gave me a 'benchmark' of sorts. its just the reason your overclocking is far more benificial than to simple play games faster.

honestly i totally understand why your doing it and your point off view.but please understand mine - we all do this to achieve things faster wether it be seti or quake or whatever.

no hard feelings

[james.miller]

the way i see it is if youve got the horsepower to propell game to 200fps+. limit it to 80 guys. youre processor does less work,stays cooler and lasts longer

im sure that is the way to go

[Emericana]

that is not the truth at all about overclocking.

if i ran my xp1600 and geforce 3 ti200 at stock speeds there is no way at all i would be able to run it like i do now: 1280x1024 w/ all effects to max and quincix AA with EAX enabled. it is just not a possibility. by overclocking you make it a possibility.

another thing , which is not a concern for soe people but it is for me, is how it performs in photoshop.

this is simply how fast the cpu is not vid card. by lowering the mult and killing the fsb you get a HUGE difference than you would if you run at stock. if you enjoy making 1600x1200 images that are 10mb+ you notice a huge difference at 400mhz DDR than you do at standard 266mhz.

[james.miller]

yeah i know.different tasks need more power than others.but your missing the point i made. i ment for game i think you should keep it limited to 80.why go any higher?
as for applications,well i think you should settle for a happy medium

[james.miller]

for example. when you have a massivly fast system running quake3 and you have the game limited to 80fps.the prosessor and g/card dont waste their time trying to render the extra 100frames per second or whatever. BUT the power is still there where the scences get very complex

THATS the only time i think the power of the pc should be used,not to render extra frame wich the eye cant pick up on anyway

apllications are a whole different ball game though.

i hope you can understand me more now

[DogfooD]

I don't play Q3 but I do play CounterStrike. CS isn't an extremely system taxing game, but limiting the fps can be bad. When you're running along, turn a corner and BAM! there's 8 Terrorists in front of you, all firing guns, throwing grenades, ect. Then your team shows up milliseconds behind you. All of a sudden there's 20 people in close quarters all doing their own shooting/nading/dieing/jumping/running, ect your system useage jumps drematically. For the average machine, say a 1Ghz with a meager vid card, your system is going to drop down in the single digit FPS easily. Having a fast machine and a decient connection keeps you alive (hopefully) I won't even get into the technical issues of your machine being able to process bullet trajectories, ect.... stuff handled client side. CS is limited to 99fps by their programming, by no means by me worrying about my $90 proccessor's life....

[james.miller]

oh yeah definatly.
But when running cs on, say a xp2200 with a g4 4600ti, it will hit that upper fps limit very easily and have no problems staying there.
when those 20 odd people start throwing grenades about and causing chaos,there is vastly more processing power available so those complex scenes wont slow down anywhere near as much.
and thats what im talking about, with the limit on - the power is there when needed,not wasted

[james.miller]

having that limit has no effect on complex scenes when they slow down.
that 99fps limit mean it will not render above 99 frames per second.if it drops below that for any reason,its because the processor and g/card just cant hack it

[james.miller]

try running any game and run the fps as fast as you can and let it keep running for 10-15 mins. then exit and check your temps.

let your system return to normal

now run the game again but this time limit the fps to 80.run for 10-15mins and check the temps.
you should see a clear drop in the temp of the cpu but no change in the quality or speed of the game. try it

[Yodums]

James Miller why don't you just post one post instead of 2-3 consecutive ones?

[Emericana]

cs isnt a good example hehe seeing how it is based on technology based in 1998. it ran fine @ 1280x1024 on my athy 850 with geforce ddr card back in the day :)

[old skool]

Where are you people getting this information?

The Athlon can only execute up to 3 instructions per clock cycle. No more. The Pentium III was the same (except the FPU wasn't fully pipelined). Whoever said it can execute 9 instructions per cycle is severely mis informed.

The Pentium IV, on the other hand, can only execute up to 2 instructions per cycle. It needs a 50% clock speed advantage to produce better results than either Athlon XP or the Pentium III.

However, the fast memory bandwidth of the Pentium IV RAMBUS systems help make up some of that difference. Looks as though the integrated memory controller on the upcoming hammer should remedy that situation.

[DaddyB]

the athlon can do 9 IPC, it has 3 floating point units, 3 interger execution units and the decoder can handle 3 instructions at a time... 3+3+3=9.

from your statement " (the P4) needs a 50% clock speed advantage to produce better results than either Athlon XP or the Pentium III" it is apparant just who is misinformed. An XP @ 1667mhz (xp2000) is roughly equal to a P4 at 2.0ghz, a 50% clock speed increase would have the P4 at 2.5GHz. the 2.53Ghz is untouchable by any present AMD CPU especially the lowly xp2000.

and saying the athlon has 3 and the P4 has 2 it seems obvious to me that you are refering to floating point units (FPU) alone. the CPU can handle a lot more then just FPU instructions. here (http://www.emulators.com/docs/pentium_1.htm#Processor%20Basics) is some reading that will help you better understand the way a CPU works if you are interested, mind you it is a long read.

[adamtekh]

did some one say memory scores , oh and im just warming up

[james.miller]

blimey.whats that benchmark running on??

oh and what is your problem yodum?

[bigfoot]

Thanks for clearing that up DaddyB.