Do people want xfire results between a 4870x2 and a 4890?

[chrisk]

If you have Vista there is a utility in "Accessories" called "Snipping tool" which is for screenshots. Saves the file in a decent size as well.

Also, you can just try the 'print screen' button, and paste and crop in your favourite image software. Again, in Vista, Paint can save as a jpeg.

[sheepdogexpress]

I am biting into quadfire 4870 range now.

HT is lowering my score because it lowering my gpu score. I think my power supply is not powerful enough and running HT is causing instability in the videocards. I notice the overall fps during the program are lower when HT is enabled. You will noticed this when you compare my first score to my lastest score.

Results in first post.

I can definitely break 30,000 if I get my x2 on water and install xp and get the HT to not lower my overall score.

I can get the 4890 card to run at 990 with lots of fans and with air the 4870x2 at 800.

[IRQ Conflict]

Very interesting sheepdog. I know I'll kick myself for this but what is 'HT'? Hyper Transport?

I am shocked that the clocks were able to be run differently. Did you try to down clock the 4890 to see if it's higher clocks had any effect on the outcome?

[sheepdogexpress]

Originally Posted by IRQ Conflict

Very interesting sheepdog. I know I'll kick myself for this but what is 'HT'? Hyper Transport?

I am shocked that the clocks were able to be run differently. Did you try to down clock the 4890 to see if it's higher clocks had any effect on the outcome?

Hyperthreading or running 8 threads from a 4 core processor. This will always increase cpu score in 3dmark.

To answer your other question. I simply overclocked the 4890 and I changed it's frequency a couple different times and it always increased the score. I never clocked it any lower than stock. I would imagine downclocking the 4890 would simply just decrease the score.

[IRQ Conflict]

Good to know. Your Orb links aren't working for me however. I thought i'd log in but it just takes me to my results, go figure.

At any rate, thanks for the tests! Good job.

I didn't know there was an option to enable/disable HT in 3DMark. I thought that was more of a P4 thing from yesteryear.


Edit: How is this scaling? very well? I did read an article with 'tri-fire' but I can't remember the source or the scores.

[sheepdogexpress]

They get deleted right away because catalyst 9.4 is not futuremark approved, even those it is a final release driver; so the links get deleted right away. That why as of late, I now take a screen shot.

[IRQ Conflict]

How is the PSU holding up?

[sheepdogexpress]

I got a bit higher score very close to 29k

I think the links get taken down because I am using catalyst 9.4. They only support up to 9.2 which is ridiculous.

I updated the bios and they allowed me to squeek out a bit more mhz.

The PS is still holding up, it still very new, but I am pretty sure I am pushing hit pretty hard those. The fans on these things is deafening.

new score 06

28736

[IRQ Conflict]

Sorry to res this thread (not really ). But, I was wondering how you made out with the frankenfire after seven CCC releases or did you change your hardware?

I was thinking of picking up an X2 for cheap if it was worth it or just another 4890 if it scales better with that.

[Chilly]

Originally Posted by IRQ Conflict

Sorry to res this thread (not really ). But, I was wondering how you made out with the frankenfire after seven CCC releases or did you change your hardware?

I was thinking of picking up an X2 for cheap if it was worth it or just another 4890 if it scales better with that.

I did Tri-Fire for about a week with a 4890 back in Sept, worked fine, very close performance to Quad-Fire if I remember correctly.

Let me see if I can't pull up some numbers/screenshots I took.