Basically today I wasn't doing much and was a little curious how some older hardware would work when paired with newer hardware. This was only going to be a short test, nothing lengthy, and simply measure by 3D Mark 06, I know it's a bit outdated now but I thought it would be a good way of comparing everything. So here goes...
The test system
Intel E8400 @ 3.5GHz
Giel 4GB Black Dragon @ 933MHz
MSI Neo2-FR
Sapphire 4850 512MB Rev 1.0
CoolerMaster Real Power M620
2 x 750GB Samsung F1's
1 x 500GB Samsung Spinpoint
Antec 900
Acer P223W 1680*1050
Additional hardware to be tested
1 x Asus X800XT 256MB
1 x Sapphire 4850 512MB Rev 2.0
Pictures
Sapphire 4850 512MB Rev 1.0 - Image1 | Image2
Asus X800XT 256MB - Image1 | Image2
Sapphire 4850 512MB Rev 2.0 - Image1 | Image2
Pre-test notes
I am aware that the motherboard I plan to be using doesn't support true 2 x PCI-E x16 slots, one will be running at x16 and the other at x4. So I know that there will be slightly more to push out of the system. I also will not be making notes of temps and power consumption, purely scores and maybe some FPS. I will be using ATi's 9.8 Driver set. I will not be testing it in some low RAM usage state, it will be from a normal boot. For these tests though, I have disconnected my 2nd display
Predicted outcomes
As I'm sure most of you are aware I'm not hoping for that greater performance increase from the X800XT, and obviously I won't be able to crossfire it either so that will once again limit it. From the additional 4850 I'm hoping to see at least a 25% increase over a single 4850.
Let the testing begin!
3D Mark 06 - stock test settings
GTA IV Benchmark - Settings
Crysis Benchmark - 1680*1050, DX9, No-AA, High
Thoughts during testing
I personally wasn't too sure if the system was going to accept the X800XT when I put it in, there was a lot of restarts before it even posted, but after setting my clock back to what it was, it then proceeded to boot into windows and install the drivers correctly for it. Shown correctly installed here.
During testing I decided that it would prob be best if I added a few more tests to help determine an outcome, so additionally I choose GTA IV bench and Crysis Bench.
Whilst I wasn't recording tempratures I realised the 4850 with just S1 on getting a bit warm and so I decided to put a fan blowing over that 4850. Picture here.
One final note, at the beginning I said it would be a short test, I've been doing this for around 2hrs now and I'm still writing it up, if you plan on testing systems, give yourself plenty of time!
Conclusion
Whilst the X800XT seemed to give a slight dip in performance, the 4850's in xfire gave a rather nice increase in the 3D Mark score, and again in the Crysis bench. I was suprised to see a decrease in the Xfire for GTA IV, but I guess a hatrick would have been too much to ask for
I'm not really sure what I thought would happen with just having a 2nd GPU in there. Maybe I should have looked into it a bit more before testing, I may have over looked something completely like how to set it up for physics or something??
I'll have to post these findings and complete the single card at another date. Thoughts and opinions so far?
The test system
Intel E8400 @ 3.5GHz
Giel 4GB Black Dragon @ 933MHz
MSI Neo2-FR
Sapphire 4850 512MB Rev 1.0
CoolerMaster Real Power M620
2 x 750GB Samsung F1's
1 x 500GB Samsung Spinpoint
Antec 900
Acer P223W 1680*1050
Additional hardware to be tested
1 x Asus X800XT 256MB
1 x Sapphire 4850 512MB Rev 2.0
Pictures
Sapphire 4850 512MB Rev 1.0 - Image1 | Image2
Asus X800XT 256MB - Image1 | Image2
Sapphire 4850 512MB Rev 2.0 - Image1 | Image2
Pre-test notes
I am aware that the motherboard I plan to be using doesn't support true 2 x PCI-E x16 slots, one will be running at x16 and the other at x4. So I know that there will be slightly more to push out of the system. I also will not be making notes of temps and power consumption, purely scores and maybe some FPS. I will be using ATi's 9.8 Driver set. I will not be testing it in some low RAM usage state, it will be from a normal boot. For these tests though, I have disconnected my 2nd display
Predicted outcomes
As I'm sure most of you are aware I'm not hoping for that greater performance increase from the X800XT, and obviously I won't be able to crossfire it either so that will once again limit it. From the additional 4850 I'm hoping to see at least a 25% increase over a single 4850.
Let the testing begin!
3D Mark 06 - stock test settings
- 4850 - 12362 *May need to be re-tested
- 4850 + X800XT - 12124
- 2 x 4850 - 12451
- 2 x 4850 (Xfire) - 16124**
GTA IV Benchmark - Settings
- 4850 - figure
- 4850 + X800XT - 45.91
- 2 x 4850 - 46.52
- 2 x 4850 (Xfire) - 43.26
Crysis Benchmark - 1680*1050, DX9, No-AA, High
- 4850 - figure
- 4850 + X800XT - 35.515
- 2 x 4850 - 36.53
- 2 x 4850 (Xfire) - 45.675
Thoughts during testing
I personally wasn't too sure if the system was going to accept the X800XT when I put it in, there was a lot of restarts before it even posted, but after setting my clock back to what it was, it then proceeded to boot into windows and install the drivers correctly for it. Shown correctly installed here.
During testing I decided that it would prob be best if I added a few more tests to help determine an outcome, so additionally I choose GTA IV bench and Crysis Bench.
Whilst I wasn't recording tempratures I realised the 4850 with just S1 on getting a bit warm and so I decided to put a fan blowing over that 4850. Picture here.
One final note, at the beginning I said it would be a short test, I've been doing this for around 2hrs now and I'm still writing it up, if you plan on testing systems, give yourself plenty of time!
Conclusion
Whilst the X800XT seemed to give a slight dip in performance, the 4850's in xfire gave a rather nice increase in the 3D Mark score, and again in the Crysis bench. I was suprised to see a decrease in the Xfire for GTA IV, but I guess a hatrick would have been too much to ask for
I'm not really sure what I thought would happen with just having a 2nd GPU in there. Maybe I should have looked into it a bit more before testing, I may have over looked something completely like how to set it up for physics or something??
I'll have to post these findings and complete the single card at another date. Thoughts and opinions so far?
