|
|||||||
![]() |
|
|
LinkBack | Thread Tools | Search this Thread | Rate Thread | Display Modes |
|
|
#1 |
|
Member (10 bit)
Join Date: Dec 2002
Location: Ireland
Posts: 739
|
Difference in performace between same speed Pentiums and Celerons?
What is the difference in performance between same spec Intel Pentium(tm) and Celeron(tm) processors?
Take the 1ghz. PIII. Why is it priced at about $100 more than it's counterpart (boasting the same clock speed)? Yes I know Pentiums have a "name". But surely consumers wouldn't pay THAT much extra for a name now would they? Or is there some other difference in power where the Pentiums excell? If anyone could spread some light on this it would be much appreciated!! JTH |
|
|
|
|
|
#2 |
|
Banned
Join Date: Feb 2002
Location: in harms way
Posts: 2,768
|
There may be some noticeable diff, created by the increased cache in the p3.
|
|
|
|
|
|
#3 |
|
Red-eyed Moderator
Staff
Premium Member
Join Date: Dec 1999
Location: Regina, Saskatchewan, Canada
Posts: 17,575
|
There are a few differences as you go down the line.
The first Celerons released at 266 and 300Mhz had no cache on them at all. They were killer overclockers quite often capable of double clocking with little or no effort. At factory speeds, the 266Mhz Celeron could barely keep up and often lost to a first generation Pentium 233MMX. Then came the "A" revisions of the Celerons at 300 and 333Mhz. Again, dirt cheap, but this time, they had 128k cache on die running at full processor speed instead of 256K cache running at half speed like the PII. Because of this nifty little feature... they were only about 5-7% slower than their PII counterparts. As the PII grew into the 100Mhz FSB and changed into the PIII, the performance gaps grew more considerably as the Celeron maintained a 66Mhz FSB all the way up to 766Mhz. This made the performance loss as great as 20%... cheap, but at a price. Then came the 133Mhz FSB PIII's... these literally crushed the 66Mhz Celerons... no competition, hands down the PIII won. Finally, at 800Mhz, the Celeron grew to a 100Mhz FSB. While once again, performance was close to that of the 100Mhz FSB PIII's, those PIII's were almost gone from the market and replaced by the 133Mhz FSB version... so while performance differences came closer together again, there was still a considerable gap. Along comes the Tualatin core for both the Celeron and the PIII. The Tualatin Celerons doubled up the cache to 256k and the Tualatin PIII's were available in 256k and 512k. Tualatin Celerons are pretty good performers keeping up quite nicely with the 256k cache coppermine PIII's... essentially, they're the same chip. The Tualatin PIII's never really took off too well as their price was sometimes higher than a lower end P4 chip. The P4 makes its debut and goes thru the growning pains of RDRAM and finally to DDR, Socket 423 to Socket 478, but doesn't really come alive until the Northwood is introduced with a 512k cache. Around this time, the P4 Celeron is introduced... the same chip, but this time with only 1/4 of the cache at 128k. While these chips clock for clock perform similarly to P4's doing everyday mundane tasks, they really begin to lag when programs are run that really thrash the cache around. In those situations, the Celeron really loses it and the P4 races ahead.
__________________
-At Ford, quality is job #1, job #2 is making them explode. ~Norm MacDonald, SNL News -Switching to Glide..Balancing in my head..inside of me... taking the glide path instead. |
|
|
|
|
|
#4 |
|
Forum Administrator
Staff
Premium Member
Join Date: May 2000
Location: Joplin MO
Posts: 37,766
|
Same clock speed, but look at the comparative FSB speeds and amount of cache. P3-1000's were available in 133 FSB and (rare) 100 FSB, and have 256k cache. Celerons at 1000 have 100 FSB and either 128k or 256k cache.
|
|
|
|
|
|
#5 |
|
Member (10 bit)
Join Date: Apr 2003
Location: Jacksonville Beach, FL
Posts: 879
|
i believe that celeron's also lack some support for certain instruction sets as well. Not sure if that is still the case though.
|
|
|
|
|
|
#6 |
|
Red-eyed Moderator
Staff
Premium Member
Join Date: Dec 1999
Location: Regina, Saskatchewan, Canada
Posts: 17,575
|
No... they carry the same instruction sets... but ...damn... trying to remember... PIII's were 8 way associative and Coppermine Celerons were only 4 way.... Does that sound right G? I forgot already.
|
|
|
|
|
|
#7 |
|
Member (10 bit)
Join Date: Dec 2002
Location: Ireland
Posts: 739
|
Thanks... and while I'm at it what exactly defines the cache? And how does it affect the CPU (might as well advance my learning seeing as I'm at!)
|
|
|
|
|
|
#8 |
|
Shiro Usagi
Premium Member
Join Date: Sep 1999
Location: Kaneohe, Hawaii
Posts: 34,002
|
|
|
|
|
|
|
#9 |
|
Member (9 bit)
|
p4 architecture needs it's cache. Celerons stink for gaming etc. The name sells them. Buying one over an athlon xp is ludicrous.
|
|
|
|
|
|
#10 |
|
Forum Administrator
Staff
Premium Member
Join Date: May 2000
Location: Joplin MO
Posts: 37,766
|
We aren't talking about the P4 architecture here, these are P3's.
|
|
|
|
|
|
#11 |
|
Member (9 bit)
|
I saw his example but thought we were being more general. Yes, I am talking about current Celeron's based on p4''s.
|
|
|
|
|
|
#12 |
|
Forum Administrator
Staff
Premium Member
Join Date: May 2000
Location: Joplin MO
Posts: 37,766
|
The P4-based Celerons are fine for general office machines, we sell a lot of them and they are very stable, cool-running, and reliable.
|
|
|
|
![]() |
| Bookmarks |
| Thread Tools | Search this Thread |
| Display Modes | Rate This Thread |
|
|