I was reading the QuakeIII piece in this month's PCGamer and there was a
quote from Brian Hook (ex-3dfx employee) in there I found interesting. It
was particularly interesting because of what he says about the Hercules...
"The best consumer-level card for GL is the Rendition v2200 in the Hercules
Thriller3D, or possible the Intel i740. They are both comprehensive 2D/3D
boards, they have a lot of RAM. Now, in the next two months
(like..now -j.), we're going to see the advent of Banshee, Savage3D and Riva
TNI. Out of those, I think Riva TNT is going to be the best, hands down. I
think the Savage3D is going to be a great low cost board."
-Brian Hook
id Software
>On Wed, 7 Oct 1998 21:01:38 -0700,
>>> Well. I guess Tom can forget about playing GPL, RB2, LB2 and F15 then.
>>That's true...but...we all know that both Sierra and Papyrus have VERY
>>poor track records in regards to the decisions they've made with 3D.
>Well, Phil, I beg to differ. *Without* D3D, Nascar2 is, by far, one
>of the all-time top-selling sims/sports titles. Currently, GPL is
backordered
>on lots of sites. EBWorld sold their initial allotment prior to ship. I
picked up
>my copy from the local EB on release day, and ALL copies were allocated
>within 1 hour of store opening. I live in the South. This is Nascar
country.
>That didn't matter in this case, since Papyrus' reputation is well-known
>and lots of people here are thrashing thru the Eiffel forest, as I write
>this.
>Doesn't sound, to me, as if they are making lots of bad decisions at
>Papyrus. I wish I had a nickel for every copy of Nascar* they sold,
>how about you?
>>Sierra will lose many sales of GPL due to the lack of up-to-date D3D
>>support and the extended DX6 feature sets.
>Why would they want to worsen the graphics in GPL, and place more
>work on a D3D *software* graphics solution when the AI is occupying so
>much of the CPU already? I, for one, am REAL glad they chose to write
>DIRECTLY to the 3DFX and Rendition chips, to unburden the CPU.
>I like lots of AI goodness and I wouldn't want the cpu spinning its
>tiny wheels in a swamp of D3D .DLLS.
>Thanks PAPY!
>>It's a shame that Sierra can
>>not tap into the huge market that it is failing to address with such a
>>great product...and don't tell me they "can't" program it because "D3D is
>>too hard" to program for. I'm testing software now that in both cases the
>>D3D performance is superior to GLIDE, using D3D/DX6.
>Phil, since you're a tester of some upcoming D3D sims, I assume that you
know
>that individual software titles can be optimized for *anything*. I could
write a
>piece of software that would shriek on Voodoo, at say 100fps and get 10fps
on D3D,
>*or* vice-versa. Can you dig it? It's all a matter of where you put your
priorities
>in development.
>Believe me, I understand the goodness of having an all-encompassing API,
like
>D3D. The problem is, like anything in life that is general in nature,
rather than
>specialized, the results can be lessened vs. a tailored solution.
>>I have a V2 and a
>>TNT and test the software in as many combinations of API's and chipsets
>>as possible.
>>I'm convinced that for now, owning a 12 meg V2 card such in a PCI slot,
>>and a 16 meg TNT AGP_2 card, is the way to go. You're covered for D3D and
>>GLIDE.
>>-Surfer
>I agree *totally*, but to imply that somehow Glide (direct to metal) API is
>somehow inferior to a software based, general purpose API, like D3D, is a
>bit misleading, dontcha think? What if tomorrow, 3DFX announced a <$200
>card that was, say, 5X the speed of a VoodooII? Would you still tout
>the value of D3D vs. Glide? I bet they are working on one. How much you
>want to bet?
>--
>// rrevved at some ISP which calls itself mindspring dot com //