Hence at IDF, NVIDIA's focus for the new GeForce 256 was on WindowsNT-based software and the visual splendor that transformation and lighting creates under ideal circumstances.

We witnessed several demonstrations of the GeForce 256 under these conditions, and it does appear to be a much stronger part than the TNT2 Ultra in the limited apps we were witness too. However the jury is still out on just how well the part will perform with most of the current DirectX 6 and 6.1 games that have been and will continue to be released over the next six to eight months.

Some show attendees we spoke with offered their fresh opinions on NVIDIA's new monster, and were principally concerned about the somewhat unimpressive pixel fill rate estimate of just 480 megapixels per second for a chip that has been in research and development for over 18 months. At a core clock speed of just 120MHz, it's easy to see how these issues could be brought to the surface so quickly.

Another question we heard mentioned frequently was why the GeForce 256 doesn't support texture compression routines in hardware, only offering a software level of support currently. This would seemingly be a no-brainer now that the impressive technology, originally developed by S3, has been officially adopted by Microsoft in DirectX 7.0.

These questions will likely be answered as more information is released by NVIDIA on the new part over the next few weeks.

NVIDIA wasn't publicly (or privately) producing game benchmark results on the GeForce rig they had at the show, they cited driver immaturity as the reason for the unwillingness to commit to specific frame rates, which is understandable.

Until we're able to use one in an environment controlled by Sharky Extreme, we won't really know just how good the GeForce 256 can be either with current software technology, or with future apps.