"DB" said:
this is what came out.........
CPU-Z version 1.30.
CPU(s)
Number of CPUs 1
Name AMD Athlon XP
Code Name Barton
Specification AMD Athlon(TM) XP 3000+
Family / Model / Stepping 6 A 0
Extended Family / Model 7 A
Package Socket A
Technology 0.13 µ
MMX, Extended MMX, 3DNow!, Extended 3DNow!, SSE
CPU Clock Speed 2153.9 MHz
Clock multiplier x 13.0
Front Side Bus Frequency 165.7 MHz
Bus Speed 331.4 MHz
L1 Data Cache 64 KBytes, 2-way, 64 Bytes line size
L1 Instruction Cache 64 KBytes, 2-way, 64 Bytes line size
L2 Cache 512 KBytes, 16-way set associative, 64 Bytes line size
L2 Speed 2153.9 MHz (Full)
L2 Location On Chip
L2 Data Prefetch Logic yes
L2 Bus Width 64 bits
Processor looks good. You can check the voltage in the hardware monitor - it
won't be exactly 1.65, but it will be somewhere in that neighbourhood. The
processor speed is the number one factor, followed by the RAM setup. The
AthlonXP Barton likes all the FSB it can get, but the chipsets and the
nature of the bus on the AthlonXP, seems to preclude giant overclocks of
the FSB. In any case, I wouldn't mess with the CPU clock setting, unless
your RAM was a bit more capable.
Mainboard and chipset
Motherboard manufacturer ASUSTeK Computer INC.
Motherboard model A7V8X-X, REV 1.xx
BIOS vendor Award Software, Inc.
BIOS revision ASUS A7V8X-X ACPI BIOS Revision 1013
BIOS release date 09/02/2004
Chipset VIA KT400A rev. 80
Southbridge VIA VT8235 rev. 00
Sensor chip IT 8712
Graphic Interface AGP
AGP Status enabled, rev. 3.5
AGP Data Transfer Rate 8x
AGP Max Rate 8x
AGP Side Band Addressing supported, enabled
AGP Aperture Size 64 MBytes
You can also check out the video card settings, either in the SmartGART
tab of the Display control panel, or you can download a copy of Powerstrip
from entechtaiwan.com and look in the O)ptions item in the toolbar popup.
It looks like you are running at AGP 8X, which is good. With Powerstrip,
you've be able to tell whether everything is enabled or not - virtually
every time I've set up a video card, not everything is healthy in Powerstrip,
so it takes me multiple install/uninstall attempts, with different
driver versions, until everything is working properly. There is nothing
more embarrassing than discovering six months from now, that AGP
texture acceleration was never working.
Memory
DRAM Type DDR-SDRAM
DRAM Size 1024 MBytes
DRAM Frequency 165.7 MHz
FSB
RAM 1:1
DRAM Interleave 4-way
CAS# Latency 2.5 clocks
RAS# to CAS# 5 clocks
RAS# Precharge 4 clocks
Cycle Time (TRAS) 9 clocks
# of memory modules 3
Module 0 DDR-SDRAM PC3200 - 256 MBytes
Module 1 DDR-SDRAM PC3200 - 512 MBytes
Module 2 Samsung DDR-SDRAM PC2100 - 256 MBytes
That is quite the mixture of RAM. I guess the Samsung is running overclocked ?
The claim is, the DRAM Frequency is 166, which is DDR333 transfer rate, and
the Samsung is rated at DDR266. It could be that the RAM timings are loose
like they are, due to the Samsung stick. You might pull the Samsung stick and
see how much better the timings become. Performance is related to CAS and
to tRCD (which is 5 clocks).
The best way to see what is happening, is to get a copy of memtest86+ from
memtest.org . It has two functions. It is certainly a memory test program, and
with your eclectic mix of RAM, I would definitely want to test the RAM for
errors. Do at least two full passes error free, before booting into Windows.
In the upper left hand corner of the memtest86+ operating screen,
there are three bandwidth indicators. The slowest of these is a measurement
of the memory bandwidth. The bandwidth is measured by a sequential read of
a region which is larger than the processor cache - that makes the accesses
basically uncached.
On my A7N8X-E, I get 1485MB/sec in dual channel mode, and 916MB/sec in single
channel mode. In my case the memory is running 2-2-2-6 timing at DDR400. You
can use memtest86+ to tweak the memory timings, based on whatever spec you
have for each stick (CPUZ SPD dump might tell you something about each stick).
The slowest stick of memory sets the timing for the whole bus. Since your
memory will be running at DDR333, you won't get to the 916MB/sec
level, but you can still use memtest86+ to compare the impact that Samsung
stick is having.
From an application perspective, take the percent improvement and divide by 3,
to get the percentage effect on your applications. In other words, if you
speed up the memory by 15%, expect your applications to be 5% faster than they
were before (as a rough average figure).
If this was my machine, I would probably be using 2x512MB, installed in slot 1
and slot 3, with either PC2700 or PC3200 RAM. Since PC3200 RAM is backward
compatible with everything, it makes sense to only be buying PC3200 now. As for
the CAS of any RAM you use, there is CAS2, CAS2.5, and CAS3, and it probably
doesn't make sense to fork out a lot of cash for some CAS2, for the amount of
improvement you would get. Chances are your existing stick is CAS3 (as that
is kind of an industry standard - CAS2 is more of an overclocker RAM, selected
by speed binning). A second 512MB stick of PC3200 CAS3 can be as cheap as $50
or so, and where you buy it, really depends on how many times you've been burned
by cheap generic RAM (I've been burned twice now).
This article compares the impact of CAS on performance:
http://www6.tomshardware.com/motherboard/20040119/index-09.html#application_benchmarks
So, work on your RAM a bit, and see what you think.
HTH,
Paul
