PDA

View Full Version : Why doesn't Intel increase clock speeds (like OCing but factory set)



george12
25-10-2004, 10:48 PM
Here's my thinking:

A Celeron D 2.4Ghz (well, mine) humms along at 3.06GHZ with a stock HSF at 39 degrees C.

A Pentium 4 3.0Ghz (well, mine) humms along at 3.0Ghz with a stock HSF at 41 degrees C.

That's interesting!!! So, if you overclock a Celeron D to a 3GHZ P4, it runs as hot as a 3GHZ P4....

So why don't Intel just change the Cel D 2.4 to a 3.06 (like, change the default multiplier or FSB), and it won't even run hot?

Surely they could do this to all CPU's they make? I have had Intel CPU's (mostly Celerons) that run at 27 degrees. This is surely a waste of potential speed.

There is also the fact that (from experience) a P4 FSB 800 2.4GHZ runs at 33 degrees, whereas a 3.0GHZ (same in every other way) runs at 41 degrees.

This is a one to one ratio according to Windows Calculator, so does this mean that the chips are effectively identical?

George

george12
25-10-2004, 10:57 PM
> So, if you overclock a Celeron D to a 3GHZ P4

Just a little mistake there, I meant to say:

"if you overclock a Celeron D to 3GHZ"

A Celly D will never be a P4 no matter what speed it is.

Big John
25-10-2004, 11:22 PM
Also some chips just wont overclock. They are rated at that speed for a reason. Most likely they will overclock but there is the good chance it was actually a higher speced CPU that could not make the grade but could handle the lower speed so they set it at that. If you go and overclock it then there is a good chance it will be unstable even at a normal temperature and you wont know what is causing the problem.

the highlander
25-10-2004, 11:47 PM
>A Celeron D 2.4Ghz (well, mine) humms along at 3.06GHZ with a stock HSF at 39 degrees C.<

what board, memory,vid card on that setup george ? Not sceptical just interested since I am having problems getting past 2.8 on my 2.4 cel d
mind you it could be due the the crap memory and mobo its running on.

metla
26-10-2004, 12:18 AM
Because they would only need a small number out of each batch to fail and it would tarnish their name.

If they could run the p4 chip at those numbers across the board then you can garentee they would.

There is also many instances of OC'd chips running cool and still failing,i thought the term was called electromanitisim but google didnt give me any results so i must be mistaken about the term used.

PaulD
26-10-2004, 07:03 AM
The term is electromigration, a problem when voltages are lifted even if extreme cooling is used.

The big PC suppliers also want a range of speeds so they can offer different priced computers.

metla
26-10-2004, 07:27 AM
Bugger,and i spent an hour trying to find an article that linked magnetic forces caused by electric current with failure of P4 chips......Lmao.

JJJJJ
26-10-2004, 09:37 AM
Don't be silly.
Then we'll get next years model this year.
Jack

Pete O\'Neil
26-10-2004, 10:30 AM
They also need to have a gap between the P4 and Celeron in terms of speed. We all know that cellery's suck, but your average noob will see a P4 @ 3Ghz and a Cellery @ 3.0GHz and buy the cellery because its cheaper. If they want to sell P4's its in there best interest to keep the cellery slightly slower than the P4 thus making the P4 worth more. Dont expect to see faster Cellery's any time soon, Intel are struggling to scale the P4 higher, expect faster FSB's and more cache first. We wont be seeing a 4.0GHz P4 and I doubt we'll see a 3.8GHz this year.

AMD are suffering from the same problem, because their chips dont benefit from extra cache and fast memory they're struggling to differentiate between the Sempron, Athlon64 and AthlonFX. For example the only difference between the Sempron 3100+ and the Athlon64 2800+ is that the Sempron doesnt have 64bit capabilities. Another example is the new Athlon64 4000+ and the Athlon FX55, the only difference is that the Athlon FX55 has another 512 of lvl2 cache. The 4000+ is literally a rebadged Athlon FX53.

Ive managed to 3.6GHz out of a 2.4GHz Cellery D, what was quite freaky was that it was using a super cheap Albatron mobo, could only get 3.4Ghz out of two DFI lanpartys and one gigabyte board i tryed. One thing i did notice while trying to overclock my cellery was the lack of third party BIOS for intel boards.

PS if anybody knows how to do a vcore volt mod on a Albatron PX865PE Pro let me know

Greg S
26-10-2004, 01:00 PM
A nice reply by dansdata to a question under CPU Giblets (http://www.dansdata.com/io038.htm) partly answers your question

DangerousDave
26-10-2004, 03:26 PM
Believe me i wish it was that easy.

The difference between Celeron and Pentium is the cache size. Cache is expensive and therefore Celeron has little cache to keep costs down. A slow P4 chip can therefore run faster than a faster celeron as cache is extremely important for performance.

I don't think Intel 'overclocks' them for faster speeds in a FSB sense, how it is done is beyond me. Read http://webster.cs.ucr.edu/AoA/Windows/HTML/AoATOC.html
if you really want to know how a computer works ;)

- David

-=JM=-
26-10-2004, 03:57 PM
I'm picking that you would also be willing to spend the extra money they would charge for the 3.06GHz instead of the 2.4GHz?

Graham L
26-10-2004, 04:05 PM
Intel are not in the business of "overclocking". They sell CPUs with a specified speed. They can guarantee that if run at the specified speed they won't break. There's a fair chance the company won't go broke doing that.

Intel have in the past had problems with "entrepreneurs" buying Intel chips, grinding off the Intel markings, and selling them as faster ones. People complained to Intel when they failed.

If you want to overclock, that's up to you. But you lose the guarantee.

george12
26-10-2004, 04:21 PM
>what board, memory,vid card on that setup george ? Not sceptical just interested since I am having problems getting past 2.8 on my 2.4 cel d
mind you it could be due the the crap memory and mobo its running on

Well, my motherboard is a Foxconn 661FX4MR-ES (http://www.foxconnchannel.com/redesign/products_motherboard.cfm?pName=661FX4MR-ES), and my memory is cheap ADATA DDR400 (1 stick of 512).

Now, I had problems too, when I tried to put it above 2.88GHZ, it would not boot, or shutdown if it was already in Windows. And by shutdown I mean like pull out plug kind.

That was with FSB 162 or something. So, I suddenly realised that I was trying to run my cheap RAM at DDR 487 speeds.

So, in the BIOS I found the FSB:RAM ratio and changed it to 1:1. I found that this, even though only 170mhz RAM speed, offered the best performance, outperforming 5:4 (212mhz).

So I set the FSB now at 170mhz and all goes sweet as. PCI is a bit high (42 mhz), but no problems have occurred.

BTW, if anybody knows how to change the FSB:PCI ratio for this board please tell me.

Cheers