On Mon, 21 Feb 2005 03:43:04 -0500, George Macdonald
<fammacd=!SPAM^(E-Mail Removed)> wrote:
>On Sun, 20 Feb 2005 22:20:38 -0500, Tony Hill <(E-Mail Removed)>
>wrote:
>
>>This article seems like it should be decent except for one REALLY
>>stupid move that kind of invalidates the whole damn thing... They
>>tested the 5xxJ series chips on a different motherboard than the 6xx
>>series chips! Given that the whole idea of the article is to compare
>>these chips and they WILL run on identical boards, I really have NO
>>IDEA why the Tech Report people were so brain-dead as to do such a
>>thing. I can only guess it's because they were in a rush to get the
>>first review out the door or some such thing.
>
>Umm, recycled "results" I think. Note that the Ahlon64 config uses a VIA
>AGP chipset mbrd, though he has also done a A64 PCI-Express review...
>apparently not enough tests for the "comparison" though.
I was thinking about the same thing. There are a few other
comparisons out as well:
http://www.xbitlabs.com/articles/cpu...tium4-6xx.html
http://www.anandtech.com/cpuchipsets...oc.aspx?i=2353
One thing that Anand noted is that the L2 cache latency went up from
23 clock cycles to 27 clock cycles when Intel doubled the cache. This
certainly helps to explain the rather lackluster improvements in
performance.
Fortunately these articles do test the P4 chips on the same
motherboards, and they do seem to indicate that, on average, the 6xx
series is probably worth the extra cost vs. a 5xxJ series P4.
Generally speaking it would seem that the 540J (3.2GHz) and the 630
(3.0GHz) should perform about the same and cost about the same, with
the latter having the advantage of 64-bit support. Of course, a
Socket 939 Athlon64 3200+ is cheaper still, and runs cooler, is just
as fast and has 64-bit support, but I guess some things don't really
change too much.
>>The end result is that their data tends to point to show the 6xx
>>series chips in a rather less encouraging light than reality, due
>>solely to the fact that the Abit board used for the 5xxJ chips seems
>>to push performance somewhat higher than the Intel board used for the
>>6xx chips (not at all surprising, Abit always strives for maximum
>>performance first and reliability second). Just how much this changes
>>things is really tough to say.
>>
>>In any case, it looks like this new chip finally brings Intel up to
>>par with AMD's Athlon64.... Only a year and a half later!
>
>Yes, they even run at lower power draw, for reasons "unknown", though still
>a good bit more than A64s.
The power consumption numbers seem a bit here and there, looking at
the various tests. All seem to indicate that the 6xx chips are better
at idle, but they're rather varied when the system is loaded down.
I'm beginning to think that a lot of this just amounts to the
statistical variability of one chip to the next.
> Gaming still favors AMD and video processing
>still favors Intel so the "balance" is still there. Hopefully we'll get
>the promised 64-bit tests real soon now.:-)
Hopefully we'll get the promised 64-bit Windows real soon now! :>
-------------
Tony Hill
hilla <underscore> 20 <at> yahoo <dot> ca
Similar Threads
