OK, I'm confused, which is the better CPU?

  • Thread starter Thread starter Mortimer Schnurd
  • Start date Start date
M

Mortimer Schnurd

I just installed my new A7N8X MB with an XP +2500 Barton. My old MB is
a SOYO KT400 Dragon with an XP +2400 T'bred. I used the same Corsair
PC3200 512M memory in each. No overclocking. I use the Passmark
Performance Test that comes with Norton 2004.

According to Passmark, the XP +2500 runs at 1833 and the +2400 runs at
1966. As I have just recently found out, these are the correct speeds
for the CPU's.

Now my question: according to Passmark, the +2400 beat out the +2500
in every category involving CPU and memory by a goodly margin. Since
the +2500 has a faster FSB (166 vs 133) and twice the L1 Cache (512
vs. 256), how can it be so much slower? Or, is it that Passmark is
really not a valid test to use? How can I determine, realistically,
which is the better chip to have in my system? I don't mind OC'ng a
bit to obtain better results. I just need to know all the factors to
help me assess if one is better than the other.
 
I just installed my new A7N8X MB with an XP +2500 Barton. My old MB is
a SOYO KT400 Dragon with an XP +2400 T'bred. I used the same Corsair
PC3200 512M memory in each. No overclocking. I use the Passmark
Performance Test that comes with Norton 2004.

According to Passmark, the XP +2500 runs at 1833 and the +2400 runs at
1966. As I have just recently found out, these are the correct speeds
for the CPU's.

Now my question: according to Passmark, the +2400 beat out the +2500
in every category involving CPU and memory by a goodly margin. Since
the +2500 has a faster FSB (166 vs 133) and twice the L1 Cache (512
vs. 256), how can it be so much slower? Or, is it that Passmark is
really not a valid test to use? How can I determine, realistically,
which is the better chip to have in my system? I don't mind OC'ng a
bit to obtain better results. I just need to know all the factors to
help me assess if one is better than the other.
Click to expand...

Hi John,

First, you have not given details of your OS.

Second, you have not said if you are using the latest NVidia drivers.

Third, if you are running Win XP and did not reinstall (I didn't on
two partitions without probs) windows will still assume that you have
256 L2 cache. You will need to modify the registry to set it to 512.

Fourth, I would have thought that one of the Futuremark benchmarks
would be better. Try several of them.

Fifth, some 2500+ Bartons can be successfully O/Cd to 3200+.

Cheers,

Guy

"If you want to find out what is wrong
with democracy, spend five minutes with
the average voter." - Winston Churchill
 
| On Mon, 05 Jan 2004 17:10:36 -0500, Mortimer Schnurd
|
| >I just installed my new A7N8X MB with an XP +2500 Barton. My old MB
is
| >a SOYO KT400 Dragon with an XP +2400 T'bred. I used the same
Corsair
| >PC3200 512M memory in each. No overclocking. I use the Passmark
| >Performance Test that comes with Norton 2004.
| >
| >According to Passmark, the XP +2500 runs at 1833 and the +2400 runs
at
| >1966. As I have just recently found out, these are the correct
speeds
| >for the CPU's.
| >
| >Now my question: according to Passmark, the +2400 beat out the
+2500
| >in every category involving CPU and memory by a goodly margin.
Since
| >the +2500 has a faster FSB (166 vs 133) and twice the L1 Cache (512
| >vs. 256), how can it be so much slower? Or, is it that Passmark is
| >really not a valid test to use? How can I determine, realistically,
| >which is the better chip to have in my system? I don't mind OC'ng a
| >bit to obtain better results. I just need to know all the factors
to
| >help me assess if one is better than the other.
|
| Hi John,
|
| First, you have not given details of your OS.
|
| Second, you have not said if you are using the latest NVidia
drivers.
|
| Third, if you are running Win XP and did not reinstall (I didn't on
| two partitions without probs) windows will still assume that you
have
| 256 L2 cache. You will need to modify the registry to set it to 512.
|
| Fourth, I would have thought that one of the Futuremark benchmarks
| would be better. Try several of them.
|
| Fifth, some 2500+ Bartons can be successfully O/Cd to 3200+.
|

Can you provide further info on how XP is in control of the use/size
of the CPU's L2 cache? My understanding is the L2 cache is an
"invisible" hardware system component operating beneath the BIOS at a
hardware layer level, and the only way an OS might change it's
operation would be to disable the L2 cache by directly accessing the
CPU's control registers.
 
Regarding <[email protected]>
Dr Teeth wrote:

Hi John,

First, you have not given details of your OS. MS Windows XP Pro

Second, you have not said if you are using the latest NVidia drivers. Latest drivers

Third, if you are running Win XP and did not reinstall (I didn't on
two partitions without probs) windows will still assume that you have
256 L2 cache. You will need to modify the registry to set it to 512. Complete, new install

Fourth, I would have thought that one of the Futuremark benchmarks
would be better. Try several of them.

Fifth, some 2500+ Bartons can be successfully O/Cd to 3200+
Click to expand...
is that 3200Mhz or an XP +3200 (not being facetious, just want to be
sure).
 
is that 3200Mhz or an XP +3200 (not being facetious, just want to be
sure).
Click to expand...

XP 3200+ by setting the FSB to 400 MHz.

Cheers,

Guy

"If you want to find out what is wrong
with democracy, spend five minutes with
the average voter." - Winston Churchill
 
Can you provide further info on how XP is in control of the use/size
of the CPU's L2 cache? My understanding is the L2 cache is an
"invisible" hardware system component operating beneath the BIOS at a
hardware layer level, and the only way an OS might change it's
operation would be to disable the L2 cache by directly accessing the
CPU's control registers.
Click to expand...

The following registry key is where it stores the amount of L2 cache:-
HKLM\SYSTEM\CurrentControlSet\Control\Session Manager\Memory
Management\SecondLevelDataCache


Cheers,

Guy

"If you want to find out what is wrong
with democracy, spend five minutes with
the average voter." - Winston Churchill
 
Back
Top