HD4870X2 possibly not using both GPUs? FPS seems low.

[MichaelD]

Originally posted by: jaredpace

Originally posted by: MichaelD
I was VERY CPU/FSB limited. I put the 4870X2 into my Q6600@3.2Ghz/400FSB box. Cat8.8s, latest DirectX, etc. Just for starters, 18,300 3DMark06. That's an increase of 4,100 points over my 8-core system (14,200 3DM06 score). That's a freaking HUGE increase. The combination of a 900MHz faster core and 25% faster FSB makes a huge diff.

You should put it with a 4.1Ghz+ qx6850, q9650, e8xxx.

Well, any of the Extreme Q series are out of my budget; not spending $900+ on a CPU. What's the "easy" (like the G0 Q6600 was a year ago) quad core Intel OC these days?

 

[Tempered81]

q6600 g0 at 9 x 400

or any of the 45nm quads with the highest available multi. they all do about 440-470FSB if you know what you're doing. The duo's E8400,8500,8600 almost always do over 4.0.

 

[MichaelD]

Quad core's where it's at for gaming these days. Pretty much all games released recently are SMP-aware and it's only going to get better.

What kind of RAM do you need to run a 470MHz FSB? That must be expensive stuff!

 

[apoppin]

Originally posted by: MichaelD
Quad core's where it's at for gaming these days. Pretty much all games released recently are SMP-aware and it's only going to get better.

What kind of RAM do you need to run a 470MHz FSB? That must be expensive stuff!

i don't agree for now

i'd prefer a dually at 4.6Ghz to a quad below 4.0 Ghz .. until i go i7

If your RAM cannot do 470FSB, you can usually use the CPU:Memory divider
- it does not kill performance

 

[betasub]

Thanks for the update MichaelD. Interesting that your CPU score in 3dm06 has actually decreased by ~10% (from the Harpertown score), whereas the video scores have shot up. Is it that the poor scores in the video tests are a result of them not utilising all eight cores, or that the Harpertown set-up has a limitation for hardware-accelerated video (e.g. PCIe, latency or drivers)?

 

[myocardia]

Originally posted by: betasub
Thanks for the update MichaelD. Interesting that your CPU score in 3dm06 has actually decreased by ~10% (from the Harpertown score), whereas the video scores have shot up. Is it that the poor scores in the video tests are a result of them not utilising all eight cores, or that the Harpertown set-up has a limitation for hardware-accelerated video (e.g. PCIe, latency or drivers)?

The reason for his lower CPU scores in 3dMark are because, even though the CPU section of that test is multithreaded (will use as many cores as you have installed), he has twice as many cores in the Harpertown system. zThe reason that the Harpertown is only 10% faster is because of how slow his Harpertown cores are running, along with the higher latency from the FB-DIMMS, and most likely even higher latency from the chipset, I would suspect.