[vox-tech] processor shootout: new intel, amd offerings

[Bill Broadley]

On Tue, Aug 27, 2002 at 12:53:34AM -0700, Peter Jay Salzman wrote:
> the article claimed that the "release" of the 2600+ came right on the
> heels of intel's announcement.  a marketing ploy to keep up with the
> main competition.

Right, but intel makes quite a few announcements, and the 2600+ announcement
came BEFORE the p4-2.8 announcement and WELL after the p4-2.53 announcement.
So actually intel announced right on the heels of the AMD (today for intel
vs last friday for amd).

> but you're arguing with hardware journalists, not me!

Well the hardware reviewers are a dime a dozen, I've yet to see error
bars, or any other indication I should look at 1 maybe 2 significant
digits of their results. 

> > Umm, sounds like unscientific bull.  Fact is according to spec they are
> > just about identical, in the 60-75 watt expected power usage range.
>  
> i guess you didn't read the article?

I read a couple.

> the reviewer had you sentiments but verified the temperature to be near
> accurate.  see my references below.

Hrm, okay how about:
http://www.amdzone.com/articleview.cfm?articleid=1198&page=2
Typical Thermal Power: 62.0 W

I couldn't find the table I was viewing, but found the Thermal design
power for the p4-2.53 is 59.3 watts, as close to a comparable number
as I could find.  IMO in the noise compared to fan rpm, heat sink
differences, heat sink compound, application pressure, case ambient temp,
case airflow etc.

As far as I'm concerned it's in the noise, even things like the angle
of an ide cable can affect airflow across a head sinke


> > Actually the benchmarks I've seen show the xp2600 winning most benchmarks
> > against the 2.53 Ghz p4.

> 
> ok, i gave my references.  here they are again:
> 
> http://www.hothardware.com/hh_files/CCAM/p4_28g.shtml
> 
> and if you don't like that, you can saunter over to:
> 
> http://www.gamersdepot.com/hardware/cpus/intel/p4_2.8/001.htm
> 
> where are YOUR references?  or would you like us to take your word for
> it?   ;-)


I'm asserting that a XP2600 ~= p4-2.6.  For example:

http://www.amdmb.com/article-display.php?ArticleID=202&PageID=9

The Athlon XP 2600+ processor is the fastest processor on the market
without a doubt; even overtaking the Intel Pentium 4 2.53 GHz setup that
we have here.



http://www.tomshardware.com/cpu/02q3/020826/p4_2800-16.html#conclusion_change_in_power

Conclusion p4-2.8 is faster (no surprise), my claim is xp2600 != p4-2.53:
As always, price-conscious buyers who are interested in getting the best
price/ performance ratio are a bit better off with an AMD Athlon XP than
with a P4..


http://www.tech-report.com/reviews/2002q3/athlonxp-2600/index.x?pg=10

Heck, Intel has a fight on its hands right now. The Athlon XP 2600+ isn't
the outright fastest x86 processor in every way, and it didn't even win
on a majority of our tests. Still, the 2600+ took top honors in Business
Winstone, flouted its FPU in Serious Sam, burned through our test render
in POV-Ray, and absolutely laid down the smack in ScienceMark.

> well, yeah.  anyone who doesn't take this for granted prolly shouldn't
> be worrying about benchmarks in the first place!   :*)

Results vary, the athlon xp 2600 is much faster then a poor p4-2.53
(i.e. ddr), but somewhat slower on some benchmarks when compared to
a dual rambus channel system.

> don't you think that depends on who "you" are?  i already said "major
> drag".  but i'm living on a grad student salary.  however, if i were
> making 80k a year, i don't think i'd be worrying about a few hundred
> dollars here and there for bragging rights for a couple of months.

Even if I made $80k a year, I'd probably still buy the best
price/perf system as long as it was 80-90% as fast.  I'd rather
save the $500 (1 GB ddr is a fair amount cheaper then rambus, 
athlons/266 boards are much cheaper then the dual rambus boards,
the cpu's are a fair amount cheaper as well) and apply towards
upgrading the system 6-12 months sooner, or *gasp* spend it on
something else fun.

Thus I wouldn't buy a radeon 9700 today for $400, I'd spend $140 on
a gf4-ti-4200 then wait 6-12 months and get something much faster
or at least much cheaper then the radeon 9700, say a NV30, or maybe
a future radeon that has a dramatically better price/perf ratio.

In the last years 3d cards have been increasing in speed even faster
then the cpu's.  Alas this also means that power and heat budgets are
being increasingly taken up by the video cards.  I think the 9700
actually requires an additional power connector *gross*.

> speaking of which, i _just_ got my tax refund check, believe it or not.
> the IRS kept nagging me that i owed $15 more than i calculated that i
> owed them.  i tried and tried to contact them to talk about it, but
> their lines were always busy.
> 
> finally, they sent a letter that mentioned "legal action" so i just
> coughed up the extra $15 and didn't bother worrying about whether i
> really owed it or not.
> 
> along with my refund check (which is what, half a year late?) came a
> letter saying that they're sending me the $15 back.  and it's considered
> taxable income for next year, on top of that.  holy cow.  i think this
> is a fight i'm going to walk away from.
> 
> anyway, i think part of that is earmarked for a new machine.  :)

Well, I'd hold on unless your desperate.  Future upgrades to the
FSB are close, the nforce 2 is close, and even the hammer isn't to far
away.

-- 
Bill Broadley
Mathematics
UC Davis