Reverend
15-Apr-2003, 04:59
The v1.2 patch for the US version of Splinter Cell is now out. You should visit www.splintercell.com/us to get the patch. It is preferred that, if you had previously installed the v1.1 patch, that you uninstall/re-install the retail version of the game, and then apply this v1.2 patch.
How this is relevant to this site :
- v1.2 introduces benchmarking feature
- Splinter Cell is a DX8.1 game, utilizing pixel programs v1.1 for various effects, including but not limited to, ocean effects and night vision modes. There are no pixel programs v1.4 codepaths for ps_1_4 capable hardware to take advantage of ps_1_4's feature of collapsing multiple ps_1_1 passes into a single pass
- the patch has three benchmarking demos included; B3D recommends ignoring these demos in favour of our own (http://www.beyond3d.com/downloads/splintercelldemo/index.php) which is more video-card-dependent - B3D's demo had its introduction in our Fairway Radeon 9000PRO review (http://www.beyond3d.com/reviews/fairway/r9000pro/index.php?p=2)
- B3D's EIC Dave Baumann has determined that Splinter Cell v1.2 would be B3D's "de-facto" Direct3D game benchmark at the moment, supplanting the "hacked" Max Payne
NOTES:
- anti-aliasing will incur a performance penalty and can be used for benchmarking -- however, due to Splinter Cell's "glow" algorithm, it introduces sometimes-unacceptable artifacts (http://www.beyond3d.com/forum/viewtopic.php?t=4595) and hence AA should not be used for image quality analysis/comparisons for multisampling-AA video cards in scenes wher SC's glow algorithm is used. At least, and not until if/when Ubisoft introduces another patch (that would also require the installation of DX9.0a) that overcomes this "limitation"
- settings pertaining to "Shadow Quality" differ when we are talking about 64MB and 128MB video cards -- be careful about this if you are to do comparisons
As usual, please read the v1.2 patch's readme.rtf for very important notes pertaining to benchmarking Splinter Cell v1.2 . It should be noted that one of the bundled demos (1_1_2TbilisiDemo) can appear to be completely unreliable due to issues with AI-lockdown (AI appears to react differently, which should not happen, depending on what I currently suspect to be down to framerate issues) -- again, we encourage the use of B3D's own benchmarking demo.
How this is relevant to this site :
- v1.2 introduces benchmarking feature
- Splinter Cell is a DX8.1 game, utilizing pixel programs v1.1 for various effects, including but not limited to, ocean effects and night vision modes. There are no pixel programs v1.4 codepaths for ps_1_4 capable hardware to take advantage of ps_1_4's feature of collapsing multiple ps_1_1 passes into a single pass
- the patch has three benchmarking demos included; B3D recommends ignoring these demos in favour of our own (http://www.beyond3d.com/downloads/splintercelldemo/index.php) which is more video-card-dependent - B3D's demo had its introduction in our Fairway Radeon 9000PRO review (http://www.beyond3d.com/reviews/fairway/r9000pro/index.php?p=2)
- B3D's EIC Dave Baumann has determined that Splinter Cell v1.2 would be B3D's "de-facto" Direct3D game benchmark at the moment, supplanting the "hacked" Max Payne
NOTES:
- anti-aliasing will incur a performance penalty and can be used for benchmarking -- however, due to Splinter Cell's "glow" algorithm, it introduces sometimes-unacceptable artifacts (http://www.beyond3d.com/forum/viewtopic.php?t=4595) and hence AA should not be used for image quality analysis/comparisons for multisampling-AA video cards in scenes wher SC's glow algorithm is used. At least, and not until if/when Ubisoft introduces another patch (that would also require the installation of DX9.0a) that overcomes this "limitation"
- settings pertaining to "Shadow Quality" differ when we are talking about 64MB and 128MB video cards -- be careful about this if you are to do comparisons
As usual, please read the v1.2 patch's readme.rtf for very important notes pertaining to benchmarking Splinter Cell v1.2 . It should be noted that one of the bundled demos (1_1_2TbilisiDemo) can appear to be completely unreliable due to issues with AI-lockdown (AI appears to react differently, which should not happen, depending on what I currently suspect to be down to framerate issues) -- again, we encourage the use of B3D's own benchmarking demo.