Explain This...

[CompSci]

I have two Athlon xp boxes.
1. XP1700+ @ 12*166, Asus A7nxx mobo, ATiVid, WD40C 7200...
2. XP2500+ @ 12.5*166, Epox 8rda+ mobo, ATiVid, WD40C 7200...
Both exactly the same except cpu and mobo... and for this test i give em both exactly the same RAM - 1g - 2 * 512pc2700

I do lots a sql server work, and have DB load and query scripts that load up about a 3 Gig database and then a complete set of qurey reports...
Tbe load script runs about 33 minutes, query scripts ~12 minutes.

Runnin comparisons on both machines, same conditions, right after clean cold boot, and then start the scripts - and I get LESS THAN 1% DIFFERENCE REPORTED SQL QUERY EXECUTION TIME! Amazing that the two different boxes get almost exactly the same(<1%) real world preformance and thruput!!!
Like...
Sql Load script: a7n9x/1700+ = 33:38 Ep8rda+/2500+ = 33:31 !!!
Sql Query script: a7n9x/1700+ = 12:15 Ep8rda+/2500+ = 12:22 !!!
And these results are repeatable, less that 1% dirrerence on all processes that takes any signifigant time, loads, builds, big querys scripts with mulitple queries in em... Side by side - almost Exactly the Same!

How can these two different cpu/mobo boxes be commin up with almost exactly the same real world preformance???
(one time i got EXACTLY the same on BOTH - coldstart 33:32 for load script!)

 

[Flinx]

Guessing!

Deleted by me, didn't make sense


The loving are the daring!<P ID="edit"><FONT SIZE=-1><EM>Edited by Flinx on 07/22/03 01:54 PM.</EM></FONT></P>

 

[smitbret]

Lower clock frequency of the Barton core of the 2500+ means it isn't that much faster than a 1700+. I would think a process like this would not have much need for a larger cache. Believe it or not AMD fans (I am one), sometimes there's no substitute for raw power (MHz). That's why Pentium can stick with the less efficient processors they put out, because on some tasks they've just got more uuumph. The other argument would be that, in real life application, the difference just isn't as big as all the tech heads in the world would have you believe. Just my guess.
-Brett

 

[shadus]

Because the processors are running at aproxomately the same mhz and motherboards rarely alter benchmark performance much...

1700+ = 1992mhz, 2500+ = 2075mhz

The difference 83mhz... that's miniscule. Insignificant really.

Shadus

 

[slvr_phoenix]

Not just are their clock speeds almost identical, but SQL being database software and all, it's probably affected significantly by hard drive speeds, not just by processor speeds. Since both systems have the exact same hard drive...

"<i>Yeah, if you treat them like equals, it'll only encourage them to think they <b>ARE</b> your equals.</i>" - Thief from <A HREF="http://www.nuklearpower.com/daily.php?date=030603" target="_new">8-Bit Theater</A>

 

[CompSci]

Yea, must be, with clock speeds about the same, FSB & disk THE same...
but
I've recently upgraded from 2 PIII 1000's basicly the same setup, one an Asus mobo and one Abit mobo, except both with exactly same cpu, and the Asus P3v4x ran away from the Abit VH6T by 10-12%... Here the setup had cpu, chipset & disk the same and one mobo was signifigantly faseer than the other. Oh well...

I must say I am very pleasently pleased with the preformance increase of these athlonXP's for the money (both 1700+ and 2500+ which suprizingly preform the same). Compared to a PIII 1000, the same database load takes less than 1/3 the time - about 114 minutes to load compared to athlon's 33 minutes... I like em, literally SMOKIN compared to a PIII - and BIG increase for the Bucks!

But i sure would have thought the XP2500+ with it's much larger cache would have greatly outpreformed the 1700+, expecially on the 12 minutes of read only report queries where indexes can get loaded up in memory and cache...

Oh well - don't underestimate a little $42 XP1700+ 'B" core - they'll run right up there with a 2500+ @3000+ speeds... Just gotta love that 1700+...