Search results
Ross Smith, Scott Sellers, Gary Tarolli: Defunct: 2002; 22 years ago () Fate: Bankrupt; Acquired by NVIDIA Corporation: Headquarters: San Jose, California, U.S. Products: Voodoo Graphics series: Website: 3dfx.com at the Wayback Machine (archived February 1, 2001)
- 2002; 21 years ago
- Ross Smith, Scott Sellers, Gary Tarolli
- August 24, 1994; 29 years ago
- Semiconductors
- Ross Smith: Ross Smith.
- Smith: Value.
- Smith: Yeah.
- Sellers: Yeah.
- Tarolli: Yup.
- Sellers: USVP.
- Smith: Yeah.
- Campbell: Yeah.
- Tarolli: Right.
- Tarolli: Yeah.
- Sellers: Yeah.
- Campbell: Too long.
- Sellers: No.
- Sellers: Yeah.
- Sellers: Taito.
- Sellers: Yeah.
- Tarolli: Sure.
- Sellers: Yeah.
- Sellers: '95.
- Sellers: Yeah.
- Smith: Yeah.
- Tarolli: Yeah.
- Tarolli: Yes.
- Sellers: Yeah.
- Tarolli: Yeah.
- Smith: Yeah.
- Tarolli: Yes.
Gordon Campbell: And Gordon Campbell. Hodge: And so why don't each of you take about a minute or two and describe your lives roughly up to the point where you need to say 3dfx to continue describing them. Tarolli: All right. Where do you want us to start?
Sellers: We definitely were the classic unfocused start-up I think. We did way too many projects for the people that we had. But ultimately, we found something that worked in the Windows Accelerator. And that's what caused the exit to Media Vision. Hodge: So how did you guys go from, I believe you described it as, the worst interview you ever did...
Sellers: That's the one you were talking about, before or after the Media Vision. But I do remember we had proposed a project within Media Vision to build a 3D game engine, a 3D chipset specifically for games. And Media Vision was seemingly a great company to do this for, because Media Vision was a consumer company. They sold these multimedia upgra...
Campbell: And the reason it kind of took six months was that there was a mindset that this was going to be a motherboard. Smith: See, he always tries to hang that on me. Campbell: It was kind of a mindset of three against one. And I kept seeing these proposals to do this motherboard. And I kept saying, no motherboard [because motherboards were ...
Smith: And at the time, a Pentium 90 was a big deal-- with all the memory we could find. Sellers: And expensive [INAUDIBLE]. Smith: And Scott and I lugged that thing across Japan on the subway. And Gordon's is saying, we have to be here, here, and here. You had to go to Japan in those days. Hodge: What was the reason for that? Smith: It's w...
Campbell: Yeah, USVP was the guy we picked. Smith: US Venture. Hodge: US Venture Partners?
Sellers: They were the only round A guy, right? Campbell: No, I think we had a little bit in there from Chase at the time, too. Sellers: That's right. Campbell: Because the introduction to USVP ultimately had come through Chase. Sellers: We brought Tony Sun from Venrock [INAUDIBLE]. Tarolli: Venrock was the other one. Sellers: Is that ...
Smith: Maybe that was our first pitch was to Mitsui. Campbell: But it was kind of interesting. Because when we went out and we picked a small number of VCs, we went out and did the pitch and came back pretty favorable. It was a pretty short process actually. Hodge: So what year was this now? Was this '94? Campbell: About that. And what was in...
Sellers: I remember it had all these different-- Tarolli: Oh, yeah. Do it the right way. Sellers: --flags, where you could do full floating point calculations and do everything kind of the SGI right way. And then Gary would use that as a kind of apples to apples comparison against OK, here's the cheap way. And since this was all about gaming a...
Sellers: --where we tried to make it look good enough. Everything had its limits. And so even in some of the early chipsets, developers couldn't do anything that they could do on an SGI workstation. For example, the depth of the scene, because of the Z-buffer limitations and other things. It was constrained, but it was perfectly acceptable for ga...
Tarolli: Memory speed, at that time, was very slow. And that was one of the keys that totally efficiently used memory. Sellers: That's right, yeah. Hodge: So were these custom fabricated or were they FPGAs [field-programmable gate array]? Sellers: Fully custom. Hodge: OK. Sellers: Yeah. Hodge: Did you have your own fab? Sellers: What'...
Sellers: You mean, in terms of when you release it to fab and when you get samples back? Hodge: Hm-hmm [Yes]. Sellers: I think at that time, it was about four to six weeks. Hodge: What is it now? Sellers: It's actually about the same. Tarolli: Still four weeks I think. Smith: Costs a little more now. Ultimately, it depends on how much y...
Tarolli: Not between the chips. Just between the Pentium CPU and our card. And then on the chip, it was all [a] custom [interface]. Sellers: So there was this frame buffer controller chip, the FBI chip as we call it, the frame buffer interface. That had the PCI connection, so that the CPU would calculate the triangle coordinates and send down co...
Smith: It became very possible for us to be in the consumer market much earlier than we anticipated. And so by the time we went public, our revenues were dominated by consumer business not arcade business. The arcade was a very important component, because it legitimized what we were doing. Hodge: What were your first customers in the arcade spac...
Smith: Konami. Hodge: Do you remember what the first game was out that had 3dfx? Smith: Was it San Francisco Rush? Tarolli: SF Rush was one. Smith: San Francisco Rush. Tarolli: Gretzky Hockey, was that an early one? I think so. Smith: NFL Blitz. Tarolli: NFL Blitz. Sellers: Midway did a lot of games.
Smith: Hydro Thunder. Tarolli: Hydro Thunder. Hodge: Arcades were still fairly viable at this time, correct? Sellers: It was a $6 billion market when we started. Smith: At the time. Sellers: But it declined quickly. Smith: It became a $0 billion market by the time I had a business focused on that-- note to self. Sellers: We used to s...
Smith: The point I'll make here is about scalability, which is we took it to new levels. Hodge: All right, we're reconvening after a break with some props this time. The panelists will describe them at the appropriate points: one of the original 3dfx boxes, [a] quad board system, another board in the back I can't see, and apparently an iPhone, wh...
Hodge: So how did these buses impact you guys? Smith: Given that we had a software geometry pipeline, we had to take the geometry work that the processor had done, and move it over to the frame buffer chip, which meant we had bus traffic. And the PCI bus at the time was a great inflection point for us, because it was significantly faster and lowe...
Smith: The chip was probably a little behind schedule. And we had a COMDEX event. Whose booth were we in? We were in Orchid's booth? Sellers: Yeah. That's right. Smith: I think it was Orchid's booth.
Smith: And Orchid was next door to Diamond. And it turns out, several of our colleagues from Pellucid were at Diamond at the time. And we didn't have silicon. So what are we going to do? So we said, well, let's put an RE [RealityEngine], and it'll be the Voodoo graphics simulator. And we were very clear that it was a simulator. And we showed an N...
Sellers: It was perfect for them. Because they didn't like the asbestos falling down from the ceiling. Hodge: Is this the building that had birds nesting on the inside?
Smith: Yeah, yeah. I-- We had a swallow's nest, a barn swallows. It would come every year. Tarolli: Back at the Sunnyvale golf course. Sellers: The eighth hole or something like that. Tarolli: The eighth hole. Smith: We'd sit out there and drink beer. And incoming golf balls you'd have to dodge a bit. But the culture was great. Sellers:...
Hodge: Very, very loud. Tarolli: Well, we're getting them quieter now. Smith: Some things don't change. Tarolli: Yes. We're now working on a quiet and low power. Hodge: So we've sort of gone through the history from beginning to end and have skipped over probably quite a lot in doing so. So what haven't we talked about that we ought to, p...
Tarolli: We had old video games. Smith: We tried to collect as many as we could. I'm not sure what happened to that though. It's kind of a tragedy.
Sellers: We had a bunch of old consoles and such. Hodge: What prompted you to do that? Smith: We were in the game business. And so knowing the history of games was really important to us. And so I think Alma brought in like a Neo Geo or something. I can't remember. It started a trend. We got a Sega Genesis. And we had this whole display case fu...
Tarolli: To answer that question, which was what would you do differently, I guess I would transition to being a more schedule focused company earlier. Early it didn't suit us, because you just got to get done what you've got to get done. But eventually, as you make that transition to what Scott said, to like the OEM company, you have to meet thei...
Sellers: Thank you. Smith: Thanks. Sellers: Thanks. END OF INTERVIEW
Feb 16, 2018 · While SGI wasn’t well-suited to take advantage of the market shift toward commodity graphics, its alumni were. And in 1994, three of those alumni—Scott Sellers, Ross Smith, and Gary...
- Ernie Smith
Jun 5, 2019 · Prior to the announcement of the Voodoo chipset, the founders Ross Smith, Gary Tarolli and Scott Sellers were showing their plans for their SST-1 graphics engine for 3D acceleration to various arcade and AIB companies.
Apr 4, 2019 · Ross Smith, Scott Sellers, and Gary Tarolli originally met while working at SGI [4]. After a short stint at Pellucid where they tried to sell IrisVision boards for PC (at 1994 $4,000/piece), they started their own company with backing from Gordie Campbell's TechFarm. Headquartered in San Jose, California, 3dfx Interactive was founded in 1994.
Mar 10, 2019 · 3dfx Interactive started back in 1994 by Ross Smith, Gary Tarolli and Scott Sellers, who were all engineers at Silicon Graphics (SGI) before breaking off to form their own graphics chipset company. Two years after starting, 3dfx released the original Voodoo Graphics chipset, the first of a short but eventful reign of graphical prowess.