Cpu speed

[Lord of Shadows]

Kinda of a newb question but I dont know the answer and would like to know.

Whats the difference between a slower cpu and a faster cpu of the same core other than the multipler? Im upgrading my cpu and was set on getting a 3ghz, but if Im going to watercool the thing shortly after, why not just get a slower and cheaper cpu of the same core?

I want to have a nice balance between cash, maximum cpu speed, and fsb speed.

 

[tristam]

well, if you get a 2.4ghz proc as opposed to a 3.0ghz proc (assuming cache, core, etc are all the same), the reason you want a 3ghz is because it will have a much higher OC ceiling than a 2.4 since it is of higher quality and has a much higher FSB multiplier. That said, i have no idea if Intel even makes procs with that much ghz difference with the same core/cache/vcore/etc.


[/foot in mouth]

P.S. oh thats right, Intel multipliers are locked, so you want to find a proc with a high multi...cuz if you get the FSB to like 285 (for example), a 13x multi vs a 14x multi is 285mhz!!!

[/other foot]

 

[KoZLop]

It will run cooler if you dont overclock a faster processor.

 

[Lord of Shadows]

When your overclocking you obviously up the fsb at the given multipler, but is the performance gained here greater than what would be attained if the multipler is incremented and the fsb decreased slightly.

I guess my answer deals with what multipler is used when overclocking an unlocked cpu to attain maximum performance. Is it always the lowest? the highest? or something between.

Edit: But that assumes there is nothing different other than the multipler between cpus

 

[nst6563]

Higher FSB = Greater overal system performance.

Intel Multipliers are locked. no chance of unlocking them unless you have an engineering sample.

I have a 3.0c and a 2.8c. The 3.0 will overclock to 3.6 stable. My 2.8 will overclock to 3.7 stable. However, if I only overclock the 2.8 to what the 3.0 maxes out at (3.6) then I still get a faster system than if I overclocked the 3.0 to 3.6 (240fsb for the 3.0c vs 257fsb for the 2.8c).

Some higher clocked cpu's like the 3.0c, 3.2, 3.4 won't oc as well as a cpu with a lower mulitplier such as the 2.4 or 2.8 (although, I don't think I've seen many cases of a 2.4 at 3.7)

Basically....it's all luck. Get what you want. I got a 3.0c first, then stumbled on this 2.8c and got it to play with. I was pleasently surprised at the results, but every cpu is different.

 

[Lord of Shadows]

Thanks for the info; guess im going to have to read alot of reviews. =)

 

[Lord of Shadows]

Thought I might as well post to anyone reading this what the difference is other than the multipler. I knew this at some point, came outa my head somehow. (darn Q-tips)

Anyway, I've never been to a manufactorer's site so im being very general. When you mass produce anything you want to reduce variances in the parts used and the methods used to create them to keep the maximum amount of working products.

When it comes to cpu's these small variances can cause a chip to fail at certain workloads, so what do you do when you have a faster chip that wont run at its rated speed, but will run slightly slower? You change its rating and sell it as a slower chip of course.

So manufactorer's test their chips at certain speeds and retest fail chips at slower speeds (normally in 200mhz decrements) rather than just throwing them out. The problem with this is that you rely on failing chips for the lower speeds, and when you need more slower chips and dont have anymore flunkies, you must ship out some slower rated chips that pass at higher speeds to meet the shipping quota. And the overclockers rejoice at the thought of getting a higher rated chip at the cost of a flunky.

(Hope I didnt get too technical there)