For the purpose of this review I will be installing the Sapphire X1650XT into the following system.
Intel Core 2 Duo 6400 @ 3Ghz
Western Digital 80gb Sata2
Duke Multimedia case
I will be comparing the Sapphire X1650XT to the current graphics card I have fitted, a 7600gt, this should be a reasonable comparison in my opinion as both are priced similarly, the X1650XT is slightly more expensive by around £10 to £15, though the X1650 has better specs so I fully expect it to come out on top in most, if not all, tests.
Installing the Sapphire X1650xt was an easy task, all I had to do was uninstall the drivers for the 7600gt, remove it from the motherboard. Then replace the 7600gt with the X1650xt reinstall the correct drivers, reboot and the job was done. I have to say it looks quite nice when fitted as the copper of the heat sink goes quite nicely with the copper of the heat sinks on the motherboard. The copper heat sink is also the reason the X1650xt seems to be quite heavy for its size.
I am pleased to say the fan is variable speed and like most, if not all, variable speed fans it’s quite noisy when first powered up but soon quietens down.
First of all I will run the following tests with the 7600gt: - 3dMark 06, FEAR and Counter Strike Source built in test and a PREY demo test. Then I will rerun them for the Sapphire X1650XT. I have chosen these, as the tests will be identical in all cases for both cards.
Once I had got the stock results from both the 7600gt and the Sapphire X1650xt (stock for my 7600gt is 560Mhz Core and 1400Mhz memory as it comes pre overclocked) I decided to see how well the X1650xt over clocks. To do this I used Ati’s Overdrive feature from within the Catalyst Control Centre to auto detect the highest clocks possible. I mean who is going to leave things stock these days when the drivers come with a built in Overclocking tool?
Using Overdrive increased the clocks from 594Mhz core 700Mhz memory (1400 effective) to 621Mhz core 790Mhz memory (1580 effective) a nice increase on the memory speed but a bit disappointing on the core speed.
First the results for 3dMark06, this I ran at the default settings.
Pretty much the results I was expecting, I thought the X1650xt would have been slightly better at stock but the over clocking made a nice improvement of approaching 10% so was definitely worth doing.
Results for F.E.A.R. From experience this game really needs 2gb of memory to run at high detail settings at 1280x1024 I only ran this at 1024x768 and even this caused the 7600gt to struggle with all the graphical settings set to max.
Again pretty much what I was expecting and again I was hoping the X1650xt would fair a little better particularly where the minimum frame rate was concerned, though this might be due to the system memory, or lack of to be precise or possibly due to its 128mbit memory bus. Turning down the detail should fix that though. Also we see the 10% increase in average frames per second from over clocking which equates to 25% increase over a stock 7600gt.
Counter Strike Source
Results for Counter Strike Source’s video stress test, which I ran at 1280x1024 with the texture details all set to max, Antialiasing set to 4x and Anisotropic Filtering set to 16x and HDR set to Full. So basically the game looks as good as it possibly can.
Again no real surprises here a good increase in performance over the stock 7600gt from the stock X1650xt and a very good jump in frames per second when over clocked.
Results for Prey, again I originally ran this at 1280x1024 with 2x AA and 2x AF but as the results for the 7600gt were what you might consider “barely acceptable” for normal playing conditions so I reran them at 1024x768 with 2x AA and 2xAF.
The Prey benchmark was the only real surprise of all the tests due to the fact the 7600gt was better on the lower resolution test than the stock X1650xt. Though over clocking the X1650xt bought it back into line at the lower resolution and widened the gap at the higher resolution to frame rates that would be more playable.
Next Page - Conclusion