Pr Rating System

[Keikan]

I know AMD uses the PR rating for it's processers but is it true to it's word? Like does the 2500+ actually perform at 2.5ghz? or 2800+ actually perform at 2.8ghz?

[clocker]

Edit:
My reply was nonsense, based on hearsay so I've removed it.

Thanks to Brainiac and VB I've learned something tonight.

[Virtualbody1234]

Pretty much.

Each perform faster with certain tasks. With some, AMD will be a bit faster and with others Intel will be faster. The differences cannot be noticed without benchmark testing.

[3rd gen noob]

the rating system is a comparison to the clock speed of an original athlon

so a 2500+ athlon xp is comparable to a 2.5GHz original athlon

[Mr. Elmo]

Originally posted by Virtualbody1234@1 January 2004 - 02:52
Pretty much.

Each perform faster with certain tasks. With some, AMD will be a bit faster and with others Intel will be faster. The differences cannot be noticed without benchmark testing.
wouldnt you notice the difference between the amd and p4 if you're doing taks such as video encoding?

[Virtualbody1234]

Video-Encoding benchmarks

http://www20.tomshardware.com/cpu/20030217/images/image026.png

I don't think you would notice any difference.

[Keikan]

no bartons in that picture

[Virtualbody1234]

Originally posted by Keikan@31 December 2003 - 22:16
no bartons in that picture
Yes there are. See the third one?

In any case the Barton would be a bit faster than the Thoroughbred and the chart shows the AMD to be faster than Intel for that particular encoding test.

I only posted this to show that there wouldn't be a noticeable difference between Intel and AMD.

[3RA1N1AC]

AMD's official explanation was that the PR rating is an equivalent to the performance of a Thunderbird Athlon running at that number of mhz. so an XP 2800+ is supposed to be the equivalent of a 2800mhz T-bird, except a 2800mhz T-bird doesn't exist, so it's all theoretical anyhow.

it's prolly a wise move to base the rating on a theoretically sped-up version of an earlier AMD chip, 'cause basing their rating system on a comparison to Intel's products would eventually trip them up when it turned out that the comparison isn't 100% accurate. it would just be embarrassing.

as some people on the forum (myself and Lamsey included) have mentioned at least a few times, a Mhz for one chip design isn't the same as a Mhz on another. the amount of work accomplished per cycle can vary wildly between Intel CPUs, AMD CPUs, Apple CPUs, etc, and even between different revisions of the same series. Mhz ratings and PR ratings will give you a very rough estimate of what you're actually getting, but they won't tell you nearly as much as a good set of benchmarks for precisely the programs you intend to use (i.e. if you play Quake 3, you'd want to look at Quake 3 tests; if you use Photoshop, you'd be interested in Photoshop tests).

i think the Mhz/PR thing is kinda similar to how horsepower ratings can tell you something about a car, but it's far from being the whole story.