[MW2-3D] While
searching around for info on the series I came
across your IGDA.ORG profile and noticed that
you mention an nVidia NV1 version that was not
released. Any details on why it was not released?
Weak performance, politics, and/or the NV1 becoming
obsolete? And what becomes of something like
that after it is decided against? Is it destroyed or
locked up in a vault with the SC?
[Chacko
Sonny] Nvidia was
actually the first company to come to us suggesting a
3D accelerated version of Mech 2. They came in to our
offices and demoed Panzer Dragoon and Virtua
Fighter on the NV1. We, in
production, were drooling at the possibilities.
However, at the time, no one knew much about 3D
accelerator cards, so while management wasn't opposed
to the idea, they weren't initially very willing
to spend money on creating these versions.
We offered them some support from the original
programming team and a small production team (the
producer and an assistant), but the actual conversion
was to be done by an Nvidia engineer. The guy
from Nvidia was a brilliant programmer, and a great
guy to work with, but in the end, it was just too big
of a task to be done by one engineer, who was also
working on other projects. Between simply
understanding the massive Mech 2 code-base and then
figuring out how to convert it to run on the NV1
(pre-Direct3D days), there was simply too much to
do. The NV1 version was the first
3D-accelerated Mech 2 version to start production,
and yet it was never completed. I very vaguely
recall problems with alpha blending, and z-buffering,
but I can't be sure. By the time the version
got close to being finished, we were so busy with
other accelerated versions that it simply slipped
under the radar. And, as you know, the NV1 never
caught on, so there was no pressure to have the
version finished. As for the state of that
version, I'm sure all of the builds Nvidia provided
us with were archived, documented and put into
storage somewhere at Activision. (think big warehouse
like at the end of Raiders of the Lost Ark)
A
quick note on Nvidia: I think their experience with
the NV1 and ports to that chipset was really an
eye-opener for Nvidia, and in some ways, the turning
point. Months after the Mech 2 NV1 experience,
Nvidia sent their engineers down to our offices for a
virtually day-long session in which they sat with our
production teams to figure out what features were
important to us in designing our 3D games, and what
we would like supported in the next generation of
chip-sets and 3D APIs. They were the only ones of all
the 3D card companies who did that. It seems
failing at first teaches you excellent lessons on how
to succeed.
[MW2-3D] Did
you guys ever discuss a Rendition Verite port?
I'd guessed that since all the ports were Win95
native, Rendition was the odd man out not having a
native 3D API for Win95 until RRedline was
released in early 1997. The Rendition people also had
VQuake, so maybe they felt they didn't need MW2?
[Chacko
Sonny] I recall working
with the Rendition people quite a bit for some
reason, and I think I have a RRedline shirt lying
around somewhere. :) Seriously though...I think we
might have been in discussions with them for a
version and we might even have done some initial work
on them, but I don't recall finishing one. Until
RRedline, the lack of a 3D API was definitely a
barrier, but I don't recall why we didn't eventually
do one. Most likely it boiled down to politics or
business terms or something. We were keen to have
Mech 2 on every single 3D card out there. It was nice
being on the cutting edge.
[MW2-3D] How was the
work load split on these things between the
Activision Studios and the hardware venders? IIRC, A
long time ago Dan Kegel mentioned [in
alt.games.mechwarrior2] that some of the work was
done by vendors.
[Chacko
Sonny] Mostly, this was
determined by the terms of the business
deal. For some OEM bundling deals, the machine
manufacturer (IBM, Compaq, etc.) would be offering a
guarantee of so many dollars or so many
units sold with a particular percentage
royalty. In those cases, we would field a
team of programmers, and production staff. In other
cases, the hardware manufacturer just really
wanted to have the game on their hardware, or there
wasn't an OEM bundling deal immediately tied to
the project, so they would volunteer programmers from
their staff to do the work. In fairness to our
programmers though, the bulk of the work (all but 1
of the versions, excepting the NV1) were done
entirely by our programmers.
[MW2-3D] Any
particular notable memories that stand out from
working on these versions? Anything particularly
hard? I'm guessing it was hell trying to get
acceptable performance out of the S3 ViRGE.
[Chacko
Sonny] I do
recall encountering substantial technical
hurdles while working on the S3 ViRGE version.To
focus instead of on positive notable memories, I
distinctly remember when we were going to get our
first build of the 3DFX version of the game.
The programmer (one of their original guys...I can't
remember his name...I think it was Scott
something...) had mentioned nonchalantly that they
had gotten the game running and were seeing around 50
fps on our test level. I mentioned it to a few of the
original Mech 2 programmers...we all snickered,
thinking that Mech 2 was the game that could bring
any 3D card to its knees. (On some of the versions,
we had even jacked up poly counts on the Mechs and
some mission geometry). We FTP'd the latest
build and ran it on a test system we had set up
earlier in the day. And there it was, running smooth
as silk at about 50 fps. There was a small
problem with alpha rendering that slowed it down a
bit during explosions, but that was fixed before
release. It was our first glimpse of how
quickly technology was progressing in this arena.
[MW2-3D] One
technical question: Did you guys ever create a
way of running a pre determined demo sequence to
compare the performance of the versions? Ala Quake's
timedemo, but (without console) requiring a command
line switch to activate or if not within the existing
game, created a separate *.EXE to do so? I'm guessing
the answer is ~no~ because of this feature seems to
be added to games that were released much later
than MW2 engine and tend to be FPS types.
[Chacko
Sonny] Unfortunately,
no, we never created a pre-determined demo sequence
to compare performance. Instead, we just ran through
the various levels playing the game and
compared frame rates. In addition, we would
compare performance in certain situations (in the
middle of a firefight...with lots of alpha, etc.)
Comparing level vs level for each card gave us a
rough idea of comparative performance.
Plus, this was early enough in the life-cycle of 3D
cards that differences in performance could
be pretty substantial.