Will AMD regain OC crown off Pentium D 805, is AMD slipping?

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.

J-Micro

Distinguished
May 23, 2006
13
0
18,510
To my knowledge, AMD hasn't held an overclocking crown since the release of their 386DX-40MHz CPU. They've exchanged the performance crown with Intel several times since then but Intel procs generally overclock better than AMD, and for the most part, always have.

This isn't to say that they necessarily scale as well, especially with the current AMD CPU technology, where a 200MHz clock speed increase is a serious and noticable increase in performance. But when you look at highest clock as a percentage over stock, Intel has always held that crown.

Regardless of the percentage of the overclock, taking in consideration
that Intel chips seem to run hotter than AMD chips, and AMD chips execute more instructions per cycle; hence more performance per 100 MHZ of overclocking than the Pentium/Celerons, how is it that Intel has always held the crown?
 

edwuave

Distinguished
Mar 5, 2006
154
0
18,680
let stop talking about those high-end processor, i got a sempron palermo 2500++ which i can overclock from 1.4GHz to all the way to 2.6GHz without raising any core voltage! which means i got an overclock of 85% from the current speed.
whats more, its the cheapest processor around, even way cheaper than 805D.
 

J-Micro

Distinguished
May 23, 2006
13
0
18,510
let stop talking about those high-end processor, i got a sempron palermo 2500++ which i can overclock from 1.4GHz to all the way to 2.6GHz without raising any core voltage! which means i got an overclock of 85% from the current speed.
whats more, its the cheapest processor around, even way cheaper than 805D.

Umm... 1.2GHZ with no bump in vcore? Is it Prime 95/Super PI stable...
and if so for how long? I have a hard time believing that w/o a screen print. (sorry!)
 

edwuave

Distinguished
Mar 5, 2006
154
0
18,680
well, if u do a research over any overclocking website, u can see that sempron palermo processor can be overclock to 2.5GHz withouth raising core voltage, provided that u pair with a good motherboard-DFI UT250GB. the board can tweak its FSB to all a maximum of 400+ MHz.
 

J-Micro

Distinguished
May 23, 2006
13
0
18,510
well, if u do a research over any overclocking website, u can see that sempron palermo processor can be overclock to 2.5GHz withouth raising core voltage, provided that u pair with a good motherboard-DFI UT250GB. the board can tweak its FSB to all a maximum of 400+ MHz.

Ok, I kind of get where you're going with this, but it's not completely
the case for each processor. This is why...

I have had two Sempron processors on the 90NM Palermo core, and one on the 130NM core (Paris). The Sempron on the Palermo core ran at 1.6GHZ; the 2600 with
64 bit extensions. I overclocked it to 2.0GHZ with no increase in voltage. That is a 25% increase, at stock voltage. This processor started to have Prime 95
errors at 2.1 GHZ (275HTT), with a voltage increase to 1.5, the processor was stable 24 hours+....etc..etc.

I overclocked the 3100+ (1800MHZ) from 1800 to 2.2 with no voltage increase in voltage, but at 2.3 and above; once again Prim 95 errors, upped the Vcore to 1.5,
then to 1.68..to push the processor to 2.6GHZ (306 FSB)

Now I have the NF8 and it goes to 450 MHZ HTT, and it also has the NF250 GB chipset. What I'm saying is, you cannot conclude, that with a Palermo core; all Semprons can be pushed to 2.5 GHZ, as well as being pushed to that speed at stock voltage. That is not an overall true statement as each chip varies in overclockability. The thing about the Sempron is that most of the cpus have had headroom to overclock. This is due to lower power requirements, lower voltage, cooler
running, and 90NM technology.

Now, the closer your processor is to 2.5 STOCK speed, the better your chance of overclocking higher, with lower voltage. If your processor is 2 GHZ stock, (Palermo), most likely it can be overclocked to 2.4, with no voltage change. If I'm not mistaken, that would be the 3200+.

Conclusion: Such generalizations cannot be made when overclocking.