Tuesday, September 21, 1999
A while ago Bubba "Masterfung"
Wolford posted a interview with Scott Sellers from 3dfx. This
interview is too good not to be posted and i wanted to share it with
you.
Our questions are in red. Our
comments are in italics. Scott Sellers responses are in black and
normal font. We have requested an interview with Nvidia so that we can
grant "fair time." Up to now, they have chosen not to call
us back. When we hear back from them, we will post an interview as
quickly as possible.
These questions were sent to 3dfx last week and we sincerely
appreciate Scott Sellers, founder and CTO of 3dfx, taking the time to
answer them.
Q) You've stated that
your next part will have 32-bit color depth output. Didn't 3dfx just
earlier this year state that gamers don't need more than 16-bit color,
or is this merely a continuation of your 60 FPS first policy and that
we can expect to see games running in 32-bit color at such speeds?
A: We have never stated that gamers
"don't need more than 16-bit color". Instead, what we've
always been very consistent in saying is that maintaining 60 fps is
the number one priority, and that we did not want to release a product
until we could sustain that framerate with 32-bit color enabled. Our
next product will certainly be able to sustain 60 fps while running at
high resolutions and 32-bit color.
Well, we do seem to recall a time when 3dfx stated that we did not
need more than 16-bit color because 32-bit caused too heavy a
framerate hit. Now that it appears that everyone (including 3dfx) has
moved to 32-bit, they are pushing that 32-bit color should be a
standard. Some gamers think 32-bit color was a standard a year ago
when TNT hit the shelves. Clearly, there is a framerate hit when
running 32-bit color the amount however, seems to be indicative of the
resolution the game is running and what each person considers an
"acceptable" framerate…
Q) Since Glide has been
written in 16-bit color for so many years, what process is involved in
getting Glide standard to 32-bit color while remaining backwards
compatible with 16-bit Glide?
A: This is actually very
straightforward. With Glide 3, we have an extension mechanism which
allows us to add new features while maintaining complete backwards
compatibility. So, it's really not a problem to add something like
32-bit color and also keep compatibility.
This is certainly great news! Some were concerned that games might
suffer from "backwards incompatibility" when the move was
made to Glide in 32-bit color. It seems that 3dfx has made provisions
for this issue and are not concerned by it. Of course, time will tell
if this hold true.
Q) Current memory speed
and interfaces seem to be limiting fill rates from achieving their
peak levels. Since rumors abound that the Next Generation Chip will be
a fill rate monster, can you give us any hints at how you expect to
overcome this limitation?
A: 3dfx has always been the king of
fill-rate, and we expect to remain there with our next generation
product. I can't get into the details right now of how we intend to do
that, but maintaining fill-rate dominance was a big factor in the
design of the next generation product and we believe we have succeeded
in that goal.
Here we go again! Another hint toward an awesome fill-rate but no
numbers! This is exactly the reason Kenn Hwang from Thresh's
Firingsquad and I have labeled 3dfx, " 3TeaseFX". So what we
are left with (with regards to the fillrate) is speculation about what
the product might "be" and since that is not exactly fair to
do without some numbers, let's not go there. We can however say that
3dfx expects to be able to reach 60 FPS at 1024x768 in 32-bit color.
That alone would be outstanding not to mention the effects that
T-Buffer and FXT1 will bring!
Q) Since the
introduction of the GeForce 256, many have questioned how the next
generation 3dfx chip will stack up. However, given the specs for the
Savage 2000, it appears that S3's chipset is the one that is really
leading the pack right now. You have mentioned that you believe the
next 3dfx card will be better than GF256 in most games. . .do you feel
the same about the Savage 2000?
A: Actually, we try not to judge a
potential competitor's product simply based on spec sheets. We can
only really comment on our ability to compete against the GeForce 256
because we've seen it in action and Nvidia has pretty much fully
disclosed that part.
S3, on the other hand, has not
demonstrated their product publicly (at least to my knowledge) and so
we really don't know whether their product lives up to the hype or
not. All we can go on is S3's track record of delivering products
which are compelling for the OEM markets because they are feature
rich, but are not particularly successful in the gaming market because
of their lackluster performance in real-world games and applications.
As a result, we'll have to wait and see
some early reviews on Savage 2000 before we can go in depth of how we
intend to dominate that product in the gaming sector. S3 has the
experience behind them, so we are keeping a close eye on them.
It is definitely ironic that S3 is riding the high life on their spec
sheet. Of course, everything is speculation until we see real world
performance. It is troubling, however, that even though Nvidia and S3
have announced their chipset, they have not released any actual FPS
numbers. Could it be that they want to ride the hype all the way to
product release hoping that 3dfx is late?
[ Next Page ]
|
