Re: faa
Actually, 100fps is not a "lot", because it's Quake3 - one of the least demanding games on the market. It may be a "lot" with decent AA compared to what we've seen currently, but certainly not enough to justify the high price tag. The point of FAA was not only that it was better looking that 4XFSAA, but it provided a minimal performance hit - right now, this is not entirely the case. Yes, it's better looking than 4XFSAA - enough to justify the price and the fact it won't work with all games, as 4XFSAA will? And I doubt most Quake3 players would consider just over 30fps "playable". You're not paying $400 for a Quake3 card, as Matrox has mentioned themselves. While it may be in the 4600's ballpark, so is the 4200 (and would probably still beat it with these drivers in many games) - and costs less than
half the price. Comparing it to a 4600 is the best case scenario for the Parhelia, actually – it’s the only way Matrox’s offering isn’t completely embarrassed on the price/performance scale.
Folks, when's the last time you saw a videocard release that actually
exceeded your expectations? It seems we go through this every year, hype builds for a product, people state "It's on my purchase list!", then the benchmarks come out, and people are dismayed. Drivers, drivers, drivers – a lot of us seem to forget this whenever we hear the specs of a new card. I can understand those that want this as a workstation card for Matrox’s excellent 2D display and multi-monitor support, but that’s an incredibly small fraction of the market.
It's obvious that driver tweaks could/should bring about a large performance boost, but that's quite a variable. How much? When? Is Matrox's driver team up to the task? Are Nvidia's the "standard", or is their driver development process unmatched, and Matrox will never catch up? I have to side with those who just don't see a market for this card. Even if the R300 and NV30 aren't orgasm-inducing (well, any card can be provided you go to the right sites
), they would have to be a screw-up of gargantuan proportions not to trounce these benchmarks. Chances are both will offer far more DX9 capability as well, even if they aren’t screamers with today’s games.
I'm particularly surprised at the 1600*1200 scores - even with immature drivers I would have expected the raw bandwidth available would let Matrox distance themselves in that area, but unfortunately not.
I'll confess that I was never really interested in Parhelia, I’m more interested in graphics products that can advance the PC gaming market (or at least keep it afloat!). $399 cards simply don’t do that, especially in light of increased competition from consoles. What they do however, is provide a platform to launch more economically feasible (from a consumers perspective) models, which might be interesting. But again, that’s another huge variable – the drivers have to mature
quickly, and the price has to plummet – I can’t see a 64meg version of this card having a snowballs chance in hell of succeeding unless it’s priced in the ~$250 range, max.
