March 13 99
A version of the IHV (independent hardware vendor) test has been leaked . The leaked version is used for the sole purpose of internal testing and contains no real levels. It serves no purpose other than for hardware vendors to test and optimize features such as multitexture and compiled vertex array implementation. It contains no playable maps. Some maps are specific for testing oversixed textures, extreme texture aspect ratios, and high polygon outputs (much higher than normal maps). We hope the public is aware that this leaked test is useless to them and illegal to have in nor is it the test that will be out in about a month. their possession. This material was given under NDA and its sole purpose for distribution was for hardware vendors to test out various features. We're flattered that so many people are interested in getting their hands on the game, but what is out now is not a game, Trying to play the IHV test will only lead to frustration.
March 21 99
Brian Hook Creates FAQ - [Kamarov] Ok I lied, but he should considering all these questions he is constantly answering on a daily basis. He updated his .plan (he must be eating his wheaties) with more answers for you all:
Seems like every product I have to go through the same Q&A with benchmarks and deal with the same stuff. I think I'm going to save this update for an FAQ next year :-)
1. "Why don't you test on a Pentium/PentiumMMX/K6-2/P2/Celeron/Linux/WinNT, is it because your engine sucks?"
No, it's because John hired me as a programmer, not a professional benchmarker. I could easily spend weeks doing nothing but benchmarking in every conceivable configuration. The last round of benchmarks was attempting to show what a high end configuration is capable of. We will likely run some benchmarks on a lower end CPU at some point, but this type of stuff pretty much eats an entire day, and Q3TEST is a bigger priority right now.
2. "Dude, Q2 benchmarks shows XYZ running faster than ABC, so obviously you're screwed up"
Q3 is not Q2, and has a different set of performance criteria than Q2. Just like Q2 was different than Q1.
3. "It's not fair that you were comparing an overclocked XYZ vs. an underclocked ABC"
The benchmark numbers were on a preliminary build of an unreleased game running on pre-production hardware using beta drivers. In other words, this stuff is subject to change. That said, both the TNT2 and V3 were running at their lowest common clockspeeds (125/150 on the TNT2 and 143 on the V3).
Apparently... - [Kamarov] "These are not my pants"
