DDR2 results in faster sound benchmark compared to DDR1

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View
Can someone explain why I get better results in 3D Mark 03's sound
benchmark when I use DDR2 RAM instead of DDR1?

The mainboard can take both DDR1 and DDR2 (not at the same time though).
  I've run the test with DDR1 (@200 MHz, timings 3-3-3-7 2T) a few
times.  Then with DDR2 (@266 MHz, timings 4-4-4-12 2T).  Every single
time the sound test was much faster with DDR2.

FSB was the same each time (271MHz).  The FSB/DDR ratio is 1:1 in the
case of DDR2, and 4:3 for DDR1.  RAM timings weren't "pushed"; the
modules are rated for those timings.

CPU: Intel Core 2 Duo E6600 @ 2.439MHz
Mainboard: Asrock 4CoreDual-VSTA (Chipset: VIA 880 Ultra)
DDR1 RAM: 2x1 GB Kingston Value RAM PC-3200 (400MHZ)
DDR2 RAM: 2x1 GB Corsair XMS2 DHX PC2-6400 (800MHZ) CL4 Dual Channel Kit
(Part number: TWIN2X2048-6400C4DHX)
Sound card: Sound Blaster Live 24-bit PCI

Screenshot of DDR1 results:

Screenshot of DDR2 results:

I'm scratching my head why I get better results in a test that has
nothing to do with RAM whatsoever ?:/

Re: DDR2 results in faster sound benchmark compared to DDR1

On Sat, 29 Mar 2008 11:40:46 +0200, Nikos Chantziaras

Quoted text here. Click to load it

Quoted text here. Click to load it

It does have something to do with ram.  All parts of the
system work together, but where does the CPU get the data it
processes?  What happens if it can't get the data fast
enough?  Memory is always a "potential" bottleneck but often
fast enough it doesn't reveal itself as one before some
other part is a more significant bottleneck.

Sound benchmarking is about processing sound, what your
system has to do before sound data gets to the sound card
output, not about what the sound card outputs to the
speakers.  Memory is bottlenecking the CPU's performance and
you can also see the result in the CPU test on your linked

Site Timeline