> I remember the days of the hobbiest magazines when they did a lot of
> benchmarking ... many compilers and video drivers were altered to recognize
> the benchmarking and would return higher numbers by such cheats ...
I remember those days as well.
Another memory is from the "DOS isn't done until Lotus won't run" days.
> Did they really do that??? my ... how childish :D
Yep. I remember following the execution of a video benchmark program in a debugger years ago. There was a place in the code that tested for an Tseng Labs ET-4000 then a BIOS check for a manufacturer name (Damnit! I forget which!). The pixels-per-second variable was incremented slightly if both tests were true.
This was not uncommon at all and I suspect it *still* isn't.
Benchmarks are next to meaningless on a level playing field and even less so with tricks like that being played.
Well there was the Quake quack issue with nVidia, and ATI got caught with it's pants down in one of the benchmarks (3D Mark I think it was) and placing a detection routine in it's drivers. When it detected this executable, things ran really fast (but the image quality was lowered, and noticeably).
Someone detected this, when they tried re-naming the .exe file and lo and behold, the IQ looked better (as it did on the nVidia card with nVidia's drivers) but the score was also lower. I remember that being discussed on the Internet (in various video card forums) after some review site discovered it while benchmarking...
In fact, I had that in mind (along with some of what Microsoft has done) when I posted my reply :D
AKA, I seem to remember hearing how some win3.1 beta testers were stating that if they used a non-M$ DOS, such as DR DOS when installing the win3.1 beta...they were presented with a rather questionable, and meant to be frightening to the end user error message...to try to persuade the person to use M$ DOS with Windows instead... I think that indication was brought up in Caldera's lawsuite (after they acquired DR DOS) against Microsoft, along with indication of how they were able to replace the bundled DOS in Windows 95 with DR DOS, with a simple 23 or so KB file they added to represent the extentions in MS DOS 7.0. Also they had a logging facility which showed all the DOS level calls Windows 95 made, and showed that it didn't just use DOS to load win95, but used it for general operations after one was at the desktop also... They showed this demonstration to the court...
> I remember the days of the
)
> I remember the days of the hobbiest magazines when they did a lot of
> benchmarking ... many compilers and video drivers were altered to recognize
> the benchmarking and would return higher numbers by such cheats ...
I remember those days as well.
Another memory is from the "DOS isn't done until Lotus won't run" days.
A lot of monkey-business in the computer field.
> Did they really do that???
)
> Did they really do that??? my ... how childish :D
Yep. I remember following the execution of a video benchmark program in a debugger years ago. There was a place in the code that tested for an Tseng Labs ET-4000 then a BIOS check for a manufacturer name (Damnit! I forget which!). The pixels-per-second variable was incremented slightly if both tests were true.
This was not uncommon at all and I suspect it *still* isn't.
Benchmarks are next to meaningless on a level playing field and even less so with tricks like that being played.
Well there was the Quake
)
Well there was the Quake quack issue with nVidia, and ATI got caught with it's pants down in one of the benchmarks (3D Mark I think it was) and placing a detection routine in it's drivers. When it detected this executable, things ran really fast (but the image quality was lowered, and noticeably).
Someone detected this, when they tried re-naming the .exe file and lo and behold, the IQ looked better (as it did on the nVidia card with nVidia's drivers) but the score was also lower. I remember that being discussed on the Internet (in various video card forums) after some review site discovered it while benchmarking...
In fact, I had that in mind (along with some of what Microsoft has done) when I posted my reply :D
AKA, I seem to remember hearing how some win3.1 beta testers were stating that if they used a non-M$ DOS, such as DR DOS when installing the win3.1 beta...they were presented with a rather questionable, and meant to be frightening to the end user error message...to try to persuade the person to use M$ DOS with Windows instead... I think that indication was brought up in Caldera's lawsuite (after they acquired DR DOS) against Microsoft, along with indication of how they were able to replace the bundled DOS in Windows 95 with DR DOS, with a simple 23 or so KB file they added to represent the extentions in MS DOS 7.0. Also they had a logging facility which showed all the DOS level calls Windows 95 made, and showed that it didn't just use DOS to load win95, but used it for general operations after one was at the desktop also... They showed this demonstration to the court...
(edit)
)
(edit)