• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

The difference between a Pentium and a Celeron

M

messerschmitt_

n00b
Joined
Apr 7, 2004
Messages
46
Why is A 2.6 GHz Pentium 4 so much more expensive than a 2.6GHz Celeron. Both of them theoritically run at the same 2.6 GHz speed. What part of the picture am i missing? Just curious.

Thnks,
 
the larger L2 Cache size and instruction set on the pentium = faster more efficient cpu
 
the celeron is very castrated, in the scheme of things, even an "ancient" 1.6 ghz 400mhz fsb norwood is better than almost all celerons. I believe the celeron only has 1/4th the cache, which cripples it for games, they are still descent for office programs though.
 
LOL, I saw that "ancient" remark that you made. Weird how I was using that CPU till just last week. Those CPUs are pretty cool tho, had it OCed at 133FSB.

Alex
 
Yeah, I believe that the best way to explain it is that the Celerons run at about maybe 100MHz LESS than what they are listed as.
 
I remember reading a review in which the Duron 1.6 smacked the 2.4ghz celeron around.
 
Originally posted by Punkrulz
Yeah, I believe that the best way to explain it is that the Celerons run at about maybe 100MHz LESS than what they are listed as.
Click to expand...
Heh, I saw a 3.2 ghz celeron(overclocked) getting r0xed by a 2 ghz(I think) p4 norwood. I think its quite a bit more than 100 mhz less.
 
Actually i've read that say a 2.6 celeron is exactly the same as a 2.6 northwood.. the only thing is that they close off the L2 on the chip.. so if you could somehow open it.. you got yourself a northwood 2.6.. but i heard that trying to open it = breaking the CPU.
 
Originally posted by RancidWAnnaRIot
Actually i've read that say a 2.6 celeron is exactly the same as a 2.6 northwood.. the only thing is that they close off the L2 on the chip.. so if you could somehow open it.. you got yourself a northwood 2.6.. but i heard that trying to open it = breaking the CPU.
Click to expand...

Its impossible to enable the full lvl2 cache on the celerons.
 
The high-speed memory in Level 2 cache is very expensive... the Celeron is the exact same processor, but with 1/4th the Level 2 cache. That means smaller chip, fewer transistors (ie, cheaper), less power consumption, etc... And it also means it sucks, basically.

which cripples it for games, they are still descent for office programs though
Click to expand...

Technically, it sucks for office apps too, it's just that you don't really notice it very much, since any even semi-new computer can (or at least SHOULD be able to) run any office app at the speed of the user.
 
Originally posted by Black Morty Rackham
The high-speed memory in Level 2 cache is very expensive... the Celeron is the exact same processor, but with 1/4th the Level 2 cache. That means smaller chip, fewer transistors (ie, cheaper), less power consumption, etc... And it also means it sucks, basically.
Click to expand...
L2 cache is expensive, yes, but the Celeron's are not a different CPU line. Take a NW, chop 384K of L2 and disable HT, remark as Celeron.

My understanding is the center of the wafer becomes Mobile P4's, outside of that is the P4, outside of that is the Celeron. The center of the wafer is higher quality, the edges of the wafer turns out the weakest chips, which become Celeron's rather than throw away an imperfect P4.

It is not on a different core, no fewer transistors, not cheaper, just fewer transistors in working order.
 
Originally posted by 0ldman
It is not on a different core, no fewer transistors, not cheaper, just fewer transistors in working order.
Click to expand...


Err, well, I suppose it's as many transistors. It's just that fewer are turned on. But it IS cheaper (they cost less, you know). :p

Other than that, yeah, that's what I said. It's the same basic chip, just "neutered".
 
For the [H]ardtype, Celerons aren't very useful; however, they run everyday processes just fine, and also have a great deal of headroom for overclocking (Your mileage may vary). I got 2.8 out 2.2 chip and it was completely stable at stock voltage.
 
I can attest to the following. An oc'ed Celery 2.5@3.0GHz can't outrun a Northwood 2.0A@2.4GHz. This much I tested at home.

Also that same 3.0GHz Celery can't even remotely compete with a stock 3.0C P4 Northwood. I have tested all these chips together at home in the last couple weeks.
 
Originally posted by Black Morty Rackham
Err, well, I suppose it's as many transistors. It's just that fewer are turned on. But it IS cheaper (they cost less, you know). :p

Other than that, yeah, that's what I said. It's the same basic chip, just "neutered".
Click to expand...
Didn't understand you right. Celeron isn't cheaper to produce, but sold cheaper as a cheaper line of chips.
 
Originally posted by 0ldman
Didn't understand you right. Celeron isn't cheaper to produce, but sold cheaper as a cheaper line of chips.
Click to expand...

Technically, it's cheaper to produce, too. It's like with the Athlon64 that has only half the cache... since the quality is worse, some of the transistors aren't fully functional. If they don't all work, the entire processor won't work at all. But instead of just throwing away a broken P4, they disable the nonfunctioning cache, remark it and sell it as a Celeron. So in a certain sense, it IS cheaper to produce, since it's either throw away or sell cheaper.



I wouldn't recomend anyone buying a Celeron, ever. AMD own the low-end/low-cost market. You can get an AthlonXP for less than a Celeron, and it will perform way, WAY, better. Even the Durons can outrun the Celerons, if you OC them a little.
 
it would be cheaper to produce assuming Intel would trash the faulty chips, but with the overclocks people are getting with the Celeron's, I have my doubts that the current crop of Celeron's are remarked crappy chips, just chopped chips from the weaker part of the wafer.
 
I don't remember if I read it somewhere or if it was in that Bob Colwell vid...

Center of the wafer is highest quality silicon, the farther from the center the lower the quality. Being Intel and having done this for a little while, even the low quality stuff is pretty damn good...

Word was in the past the Celeron's were remarked PIII's with nonfunctional portions of the L2, much like the Thorton is to the Barton Athlon XP's.

Even tho the Celeron is just a chopped Pentium III or Pentium 4, depending on the version, it is still the cheaper route to take over the Athlon/Duron having different cores and different manufacturing lines.

One manufacturing line producing identical chips, just disabling or not connecting parts of the L2 and other features while packaging. Cheaper costs involved even though the product is essentially identical.
 
Originally posted by Black Morty Rackham
Technically, it's cheaper to produce, too. It's like with the Athlon64 that has only half the cache... since the quality is worse, some of the transistors aren't fully functional. If they don't all work, the entire processor won't work at all. But instead of just throwing away a broken P4, they disable the nonfunctioning cache, remark it and sell it as a Celeron. So in a certain sense, it IS cheaper to produce, since it's either throw away or sell cheaper.
Click to expand...

Dude, you need to learn something about economics. They are not cheaper to produce. As my man, Oldman, already said, they are the same die, made from the same wafer. All wafers cost the same to make. You are right on the fact that they are rejected P4s, but saying the quality is worse??? I dont get that.

There are more differences to the Celeron then just the cache. They also do not have HyperThreading and run a a 400 MHz bus speed.
 
well they are cheaper to produce in the sense that it is profits where you would have otherwise seen loss, (no money from dead silicon), Intel doesn't have to spend a dime extra and they get all these cheap chips to dump into the market.
 
A P4 with half the L2 cache not working is not a functioning P4. Rather than throwing it away, it disables that portion of the cache, and the hyper threading for some damn reason, and remarks it a Celeron. More processors for the same cost for Intel, meaning it's, in a sense, cheaper to produce.
 
I would never use a celeron...

budget CPU....

its like saying you go and get a nice car... Ferari or something and put a Lebaron K car motor in it....

:p
 
(Disclaimer...)

I really have no idea as to how Intel produces the Celerons... this is just me, err, guessing, pretty much. :p

If I'm wrong, please prove me so, so I can know how they really do it. :D
 
Here's the difference:

Pentium 4: goodness, generally, like the northwoods

Celeron: badness, for most recent generations
 
It bites that the Celeron's are such bad chips. They weren't when the 300A came out. If they just let them have a 533MHz FSB and let them keep Hyperthreading they could still cut the cache memory in half and we'd have a decent processor.

The problem is anything you do to improve them such as increasing the FSB or let them keep their Hyperthreading would make them too close to Pentium 4's in performance.

Back when the Celeron A's came out Intel lost alot of sales that would normally have gone to higher end Pentium II's to the Celeron's thus reducing the amount of money they made.

Intel would rather sell of the castrated junk to OEM's and system integrators and that way they profit all around. The Celeron is liked well enough in coporate PC's and they are not BAD chips, they just aren't for the enthusiest.
 
Originally posted by Black Morty Rackham
(Disclaimer...)

I really have no idea as to how Intel produces the Celerons... this is just me, err, guessing, pretty much. :p

If I'm wrong, please prove me so, so I can know how they really do it. :D
Click to expand...
Noone can prove you wrong as Intel will never officially release any information relating directly to internal business policy or manufacturing methods.

Any info I've gathered cannot be linked or proven, just general information gathered over a *few* years in the business.

All info gathered is completely unofficial, as is the current line of info that says the Celerons are nothing more than nuked P4's, not damaged, not recouping losses or preventing losses, just neutered P4s.

As I've said numerous times, the Celeron comes from the outside of the wafer, which has the potential for weaker chips, not necessarily defective chips. As the Celerons overclock like beasts, I really have my doubts that any have defective L2 cache.
 
I had a celeron, before my laptop. It was a 433 or 466, forget which. Really wasn't half bad at the time.
 
the athlon xp-m 2400 in my laptop ( which benchmarks at around an athlon xp 2200 level ) can beat a celeron clocked at 3.3ghz.

celerons werent always bad, back in the p2 days intel had the bright idea of using a new core with the celerons before they used them in the p2's so althought they had 1/4 the cache, the cache was running at the same speed of the cpu versus 1/2 the speed of the cpu in the normal p2's. so a celeron 300 overclocked to 450, would outpreform p2's at the same clock speed.

coppermine celerons werent all that bad either, i was lucky and got some good overclocking chips. i had a celeron 600@1035, and a celeron 1100@1250. the celeron 600@1035 preformed about as good as a p3 933, and the celeron 1100@1250 could beat a p3 1ghz no problem.

tualatin celeron were the last good celeron chips imo. they had 256kb cache and ran the newer, small, better, tualatin core. i had 2 1.3's that ran at 1.6ghz. the only bad part about tualatins was that they did not have many ddr boards. but the cpu itself could beat some p4's.

willy celerons......junk, never NEVER get one. if you do, im sorry, dont expect to get anything over 2ghz out of it.

northwood celerons, if you like to see high clock frequency's from low speed chips, thats fine, but dont expect good performance. as said in this thread, a celeron clocked at 3ghz can be beaten by cpu's with 600mhz or more less. the 128kb of cache and lack of HT impacts performance ALOT.

prescott celeron i think will be really good cpu's. they have 256kb cache, run on 133fsb, and use the newer prescott core. im very interested to see if prescott celerons can hit 200fsb.
 
You must log in or register to reply here.
Back
Top