Jetway X1950 Pro 256MB PCI-E Video Card 

Testing:

     To get a solid vision of the performance you can expect from the Jetway X1950 Pro I will be using a combination of synthetic benchmarks and gaming benchmarks.  I'll be using 2 of today's most popular games; Battlefield 2142 and Team Fortress 2. Benchmarking games is the easiest and probably the best way to compare video cards in the real world. For synthetic benchmarking programs, we will be using 3DMark06 and a relatively new benchmark; Lightsmark Version 1.3. 

     Futuremark's 3dMark06 has become the go-to benchmarking tool when it comes to synthetic benchmarking.  Love it or hate it, there is no question that 3DMark06 is the measuring stick of video card performance industry wide.

     We will be using Version 1.1.0 professional.  The free downloadable version can be found here.

     To get a good baseline for performance I ran a test using the default settings in 3DMark06.  While a score of 4785 3DMarks won't set a record anytime soon, it is a solid score coming close to the 5602 score we measured on our last X1900 XTX testing.  Not bad for a $127 entry level X1950 Pro.

     With the same settings my 7800 GTX turns in a score of 5003 3DMarks.  The only major performance jump is in the SM2.0 Score with a score of 1972 for the 7800.

     The Lightsmark benchmark is designed by Stepan Hrbek.  Like 3DMark06, Lightsmark runs a series of synthetic tests designed to measure video performance with an emphasis on real-time generated lighting.  Lightsmark generates a score measured in average FPS that can prove useful as another tool to compare video performance.  I don't know how Lightsmark will catch on but I thought I would include it as a possible rising star! It looks pretty cool too.

"Before Lightsmark, real-time global illumination was limited to small scenes, small resolutions, small speeds, specially crafted scenes with handmade optimizations. Lightsmark breaks all limits at once, running in reasonably sized scene (220000 triangles) in high resolution (1680x1050) at excellent speed (100-400fps). Lighting is computed fully automatically in original unmodified scene from 2007 game World of Padman, not tweaked for Lightsmark and with all sorts of geometrical difficulties, with extra rooms hidden below floor etc." ~ Lightsmark.

     With the X1950 Pro turning in a score of 45.2 I went on to test the 7800 GTX to put this number into perspective.  The 7800 clocks in at a much higher 108.6 showing the overall speed advantage of the 7800 GTX edging out the newer architecture.  With such a new tool for benchmarking, more test results will help to ferret out the strengths and weaknesses of this new software.

     ATI Tray Tools offers a simple benchmark as part of their suite of tools.  To get another baseline, I fired up the test and recorded this 5109.52 score.  This number will only be useful for other ATI users to see how they measure up.  Well, its a start!

     Without the ATI Overdrive tool to help us monitor the card performance for overclocking, I quickly turned to my other software overclocking utilities for help.  First lets fire up the overclocking utility included with ATI Tray Tools.  This utility allows you to manually set your clocks or use the automated tool to find the maxes for you.  Its definitely a slick tool.  Sadly any attempt to use this utility to up the clocks met with an immediate crash requiring a reboot at the default clocks.  Next!

     Next I loaded up RivaTuner and navigated to the overclocking tab shown here.  Despite my best efforts, I was unable to get the memory to move even a single setting without a crash.  I was able to get the card to hold stable at a core of 674 MHz on the desktop but any attempt to benchmark the card met with failure.  Bummer as 674 MHz is a pretty respectable score that had me eyeing the tiny heatsink with hope.

     For my next attempt, I fired up the ATITool.  Like the other tools this setup allows both manual and automated overclocking.  Once again, any memory tweaking means its reboot time.  Call it luck or just the new tool, but I was able to get the automated Core setting back to the 674 MHz number before a crash this time out!  After yet another reboot I manually backed things up to the 674 MHz number and attempted a benchmark.  No dice, but at least I didn't have to reboot.  I gradually brought the clocks down until the included test ran clean for over 30 minutes.

     As you can see, the end result is a much more conservative Core speed of 621 MHz.  With out any way of measuring the temperatures, I am forced to turn up the case fans and cross my fingers for a completed benchmark.

     Back to ATI Tray Tools for another bench.  The above picture confirms my max overclock and provides a view of the new 5157.02 score.  That's right a whopping increase of just 47.5 points or about .99%.  Hummm...

     OK so maybe the 3DMark06 numbers will show more of a performance boost.  As you can see here the much longer test duration of the 3DMark 06 test coupled with the heat generated by the overclock resulted in a drop in score.  While many products peak and then decrease as the overclock increases, this was unexpected after the slight improvements above.