Computer Chess Club Archives


Search

Terms

Messages

Subject: You are amusing

Author: Aaron Gordon

Date: 16:16:46 10/03/02

Go up one level in this thread


On October 03, 2002 at 17:28:54, Rajen Gupta wrote:

>where does intel bullshit people? their fastest processor the 2800 is faster
>than the fastest AMD has on offer ie XP2200 for 99% of applications-even for our
>beloved chess programmes- on the contrary it has been AMD which has been
>bullshitting people with periodic announcements and launches of their  virtual
>processors(ie AMD 2400-2800 which do not exist).
>
>They first bullshitted about the SOI on Barton. now they claim that SOI will be
>absent from Barton.
>
>http://www.theinquirer.net/?article=5599
>
>They are also completely clueles about when the hammer will be out-this year?
>next year? never?
>
>http://www.theinquirer.net/?article=5696
>
>look at waht intel ahs on offer
>
>http://www.xbitlabs.com/news/story.html?id=993470728
>
>
>
>>Intel will of course find new ways to BS people. Whether it's with them trying
>>to make people think a new CPU is going to increase their internet performance
>>as they've tried in the past or ramping up clock speeds while providing poor
>>performance.
>
>why do you consider intels performance poor? they have always admitted that
>instuctions per cycle of a P4 is less than competing or even its own P3 but it
>has the potential to ramp up to "Infinity and Beyond" (refer to previous link)it
>currently thrashes the living daylights out of AMD in 99.99% of applications,
>including chess programmes-ask John Merlino why did he decide to chose the P4
>rather than an AMD for his match against Larry.
>
>I'm positive Intel will continue to lie, buy people off, influence
>>review pages so they provide favorable reviews, etc. Intel will be around for
>>some time no doubt but as I've said in the past on many occasions... you can
>>only BS people for so long.
>
>Facts are Facts-however unpalatable
>
>rajen

Where does Intel bullshit people? Well, lets see.. "With the new Pentium3
processor your internet experience will be faster than ever!" was a line that
was even on their main page. Take a 486dx2-66 on a T3 and a P4-2.8GHz with a
56k. Gee, I wonder which will download faster. They talked up MMX like it was
going to increase performance 2 fold on many occasions. This too is also crap.
Now they make highly clocked cpus with poor performance. Poor performance in
what you might ask? Well, when a 1.2GHz Celeron (based on a P3 core) beats the
absolute hell out of a Celeron 1.7 @ 2.26GHz (P4 based) you know there are
problems. Also, Celeron-1GHz beats a P4-1.5Ghz in Crafty. THAT is sad.
Just take a look at my Crafty benchmark graph:
http://speedycpu.dyndns.org/crafty/craftybench4.jpg

If you think the 2200+ loses in 99% of applications you are sadly mistaken.
Many, many review pages are Intel biased and so are some benchmarks. Take
SysMark 2002 for example. The 8 tests where the Athlon beat the P4 in were
replaced by one of the few tests where the P4 wins. All eight tests do the same
thing. Even AMD got on Bapco about this.

Lets take a look at Tomshardware now. There was a faked P4 review on there not
too long ago. Also sites like Anandtech.com, Tomshardware.com, Hardocp.com, etc
all cripple the AMD systems. Look at a little bit of proof...

Here is the graph in question:
ftp://speedycpu.dyndns.org/pub/misc/q3-tom.gif

I ran the same test with the *EXACT* same Quake3 settings and version as Tom.

You see the XP 1900+ with Geforce4 ti4600 get 62.7 fps.. I clock down to
1.6ghz/1900+, I have an overclocked GF3... I get 61.7 fps. Seems reasonable, no?
Well, let me tell you what settings I used. *1x* AGP, 4mb AGP aperature size,
lowest ram settings possible (3-3-3-6), disabled write combining on the
videocard, ALSO I ran the ram at 100mhz instead of 133mhz

Now, when running 1900+ (1.6ghz), 133mhz fsb, 133mhz ram, full ram tweaks, full
vid tweaks, 4x agp, 256mb agp ap. size I got *80.1fps*... That is *NOT*
overclocked AND with a Gf3(they had a GF4 Ti4600).. That means that puts my PC
(stock) between a AthlonXP 3400+ (2666mhz/166fsb/166mem) on toms hardware and a
P4-2.66ghz. This is at 1.6ghz w/ full bios tweaks

At 1800MHz, 150fsb, 150mhz memory I get 90.1 fps, right inbetween a p4-3.1ghz
(590mhz memory) and a P4-3GHz w/ 580mhz memory

Tom definitely cripples his AMD systems. All of his scores are ridiculously low
and this goes for other pages as well. AMD shipped this reference system to me
for development on optimized Quake3 dlls. During that time AMD also shipped the
same reference systems to many, many review pages. Guess what? I ran the SAME
benchmarks they did at the lowest possible settings I could in the bios and got
identical scores. This in my opinion is horrible testing on their part.

http://www.vanshardware.com/reviews/2002/01/020107_AthlonXP2000_2/020107_AthlonXP2000_2.htm

Vanshardware also got the exact same system as I did but with a 2000+ instead of
a 1900+. I checked his scores against mine and his system WAS optimized. This
surprised me as most review pages have Intel firmly crammed in their colon. Just
take a look at the results of a 2000+(1.67ghz) vs a P4-2GHz.




This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.