p3 more expensive??

[CalsonicGTR]

why is the p3 processor much more than an equally powerful celeron?

[Yodums]

Well if you were to compare a 1ghz Celeron and a 1ghz P3, the P3 would be faster because the Celeron has way lower bus speeds than like the P3 133mhz FSB.

While the Celerons have higher multipliers and lower FSB's ..

Much more harder performance comes out of the P3 .

[Godfodda]

Originally posted by CalsonicGTR
why is the p3 processor much more than an equally powerful celeron?

FSB speeds, twice the cache. There was some discussion around here a few weeks ago about whether the Celeron was also missing some instructions included in the P3. I don't remember what was ever decided.

Yodums is right about the performance difference, too. Higher FSB, extra cache, and lower multiplier make the P3 a better performer than a Celly at equal speed. And since performance = $$$, the P3 costs more.

[Bmxpunk86pl]

the fsb is higher the cache is higher and i think it supports SSE2 and MMX+

[Snoopy]

My PIII 850MHz (cC0, 100MHz x8.5, 256KB 850MHz L2 cache) does not support SSE2 or MMX+ (at least according with WCPUID).

[Yodums]

Well basically all you need is that it has twice the cache and the FSB is twice more ..

And the performance is very noticeable.

Hope this answers your question..

[Dtown]

Neither does my PIII 700@933! :D

[tainice]

Originally posted by Yodums
Well if you were to compare a 1ghz Celeron and a 1ghz P3, the P3 would be faster because the Celeron has way lower bus speeds than like the P3 133mhz FSB.

While the Celerons have higher multipliers and lower FSB's ..

Much more harder performance comes out of the P3 .


what does multiplier have to do with performace? is low the better, or higher the better?

[jazztrumpet216]

Originally posted by Bmxpunk86pl
the fsb is higher the cache is higher and i think it supports SSE2 and MMX+

I think the P4 supports SSE2 and MMX+, not the P3.

[Pinky]

Celeron slower if the bus speed is at default for the same MHZ P3 chip... and the fact the P3 has twice the L2 cache doesn't help either.

Now, from a PC guru I work with (no lie, the guy's forgotten more than I'll ever know) he thought the Celeron was missing some advanced multimedia instructions, but said they were for the integration of sound and video applications, and since we now use superior integrated components to when these instructions were first developed and included, we wouldn't notice any difference on that end.

I can't verify a damn bit of this. Intel doesn't exactly give this secret away easily, possibly because it's in part with the multiplier locking scheme of things ;)... hate to get that conspiracy going again :p !!

[funnyperson1]

well lets see, a 1.1 celly will perform like an 850 pIII, unless of course you overclock the snot oout of that poor unsuspecting celly...:D

[Ebola]

there is a huge difference in performance between the p3 and celly. I had a celly running at 952 mhz which was 112fsb and my p3 1000E at default kicked the crap out of it in everything and thats running 100 fsb.

now for folding i have to admit, a celly is equal to a p3 mhz for mhz.

[Ebola]

a p3 also can be run as a dually unlike the cellys. and if you look at the bottom of the chip you can see a lot more things on the p3.

[Godfodda]

Originally posted by tainice
what does multiplier have to do with performace? is low the better, or higher the better?

My understanding is that lower is better. This has something to do with how long the rest of the system has to wait for information from the CPU. I'm not sure if that's exactly it, but maybe someone else can clear it up. It's been a while since I read about it.

[Philip]

Higher latency, what does that mean?

[6502kid]

I think that is the number of CPU cycles the processor waits
before grabbing the next chunk out of the L2 cache.

I could be wrong......

I got mine set at 0. All full speed 512k of it.:D

[Philip]

How do you guys come up with the PR rating for your processors? do you get that with a benchmark program or just by looking up some technical documents?

My apologies if this is a silly question as I have not follow the Intel CPU stuffs for quite some time as you can tell from my signature.

[Yodums]

Originally posted by tainice



what does multiplier have to do with performace? is low the better, or higher the better?

... The FSB is the main performance the mutiplier you dont worry about because it can't be changed..

But FSB is really important since it overclocks everything..

Your ram, PCI, yada yada..

While multiplier could have been useful where you hit the max FSB and you turn the multiplier up..

So a 1gig 133 FSB p3(7.5 multiplier) would run better than a 100 fsb p3 10 multiplier

[Monaco]

I did a bit of research on this topic a whiler back- the cores all come off the same wafer, so P3 and Celeron CPUs all have the same instruction set. It turns out the big difference is in the on-chip cache.

A Pentium3 has 256KB of 8-way associative cache.

A Coppermine Celeron has 128KB of 2-way associative cache.

It's not just how much cache, but how the chip hits up the cahce and uses it. If you've ever seen a Linpack benchmark graphed, you know that a large on-chip cache can frequently go to waste- or at least not be the big performance boost it's supposed to be.

In my experience, all things being equal, a Celeron is equal to a P3 clocked 150-200mhz slower.

[tainice]

but how come according to cpu-z that my celeron has a 4-way associative......instead of 2-way.. Anyway, if what you said is true, then why a Celeron 300a after overclocked to 450(with fsb 100) is roughly equaled to a PII 400~450..performance wise?

Anyway, i even seen this article once, which basically saying that a celey300a@450 could head on with a piii450 no problem..

:D Well, since piii is SSE ready, i don't think a celey1 can compete with it in term of floating point......anyway, i am just bsing, so leave any suggestion or thought

[Monaco]

Well I could be wrong, could be 4-way not 2-way. Trust the program over me, I would. I make mistakes!:)

Pentium 2 generation Celerons are very different from Pentium 3 genreation Celerons. That is why I specifically stated those numbers were for Coppermine-type chips.

P2 type Celerons were basically just a P2 stuck at a 66mhz FSB and low clockspeed.The success of the overclcocked 300A spurred Intel to reduce the newer Coppermine Celeron's set associativity in order to make the P3s seem worth triple the price of a Celeron.