Welcome to Emulationworld

Forum Index | FAQ | New User | Login | Search

Make a New PostPrevious ThreadView All ThreadsNext ThreadShow in Flat Mode*


SubjectNot entirely Reply to this message
Posted bywildcat
Posted on08/27/07 11:03 PM



> A core 2 duo at 2.4ghz is twice the computer a P4 at 3.2ghz is.
>
> The Core 2 duo name means it has 2 processor cores instead of just one. So you
> don't have a 2.4ghz cpu, you have TWO, so if the program is smart enough to use
> both cores, you actually have a 4.8ghz.

I'd say it's closer to 4.0 to 4.5 equivalent. Yes, you have the dual core working for you, but there's also more overhead with memory, cache, cross-core communication, etc. You generally don't get a doubling (or quadrupling, or whatever) of performance with multi-core processors. Sometimes you do, but not generally.

> But even that isn't true, because the way the cpu works is also more refined, so
> what you have to look for is the performance name. For example, a core 2 duo at
> 2.4 is likely an e6600.

Yep. Conroe 65 nm.

> 6600 is the indicated performance. So, in that case, even two P4s at 3.2, aren't
> as good as a single 6600.

Intel doesn't work that way. AMD does (or at least they used to). The Athlon XP, Athlon 64 and Sempron model numbers are supposed to be relative to the performance of a Thunderbird Athlon (or maybe a P4 (probably a Willamette), although AMD never confirmed it) at that clock speed (so that an Palomino at 1.4 GHz gets the same performance as a TBird at 1.6 GHz, and an A64 Venice at 2.4 GHz is roughly equivalent to a TBird running at 3.8 GHz). They may not be using this scheme anymore, though (and they don't at all for Opteron and FX processors).

Intel, OTOH, assigns somewhat arbitrary model numbers. In general, yes, as the number gets higher, the processor gets faster, but it's not relative to any given processor, so you can't simply say that a E6600 runs like 2.1 P4 3.2s or a P4 at 6.6 GHz, or whatever (even though it may be true). All you can say is that the E6600 performs better than an E6550, but not as well as an E6700.



-
Entire Thread
Subject  Posted byPosted On
*Core 2 Duo  sdplayer08/27/07 05:57 PM
.*Performance hasn't been measured purely by clock speed in years...  Death Knight08/27/07 09:33 PM
...Not entirely  wildcat08/27/07 11:03 PM
...*I knew somebody would call me out on that...  Death Knight08/28/07 00:04 AM