Matrox Parhelia Benchmarks and Review
Crankshaft writes "Matrox Parhelia boards are due to hit retail shelves by the end of the month.
Cards have been sent out and
the
benchmarks are up, showing the card's strengths and weaknesses. You
want great looking Anti-Aliasing? The Parhelia is for you. You want
killer frame rates? You might have to look elsewhere."
this card is not woth the asking price. It's nice to see Matrox trying to get back into the game, but the technology in the card is well... so last year.
:P.
I predict that this card may eventually be popular in high-end workstations, with matrox fans (if there are any), and with people who for one reason or another just don't like nVidia.
It may also take some market share away from ATi, but I don't suspect it will cause a huge dent.
The stats really just don't impress me. Then again, I'm a heavy Windows gamer, and from reading their white papers on this card they must not be trying for the gamer market.
I'm just gonna sit back and wait for the GeForce 5, just like I waited for the Voodoo 5500. Hopefully nVidia doesn't go out of business
Linux is dead.
LU
It's an astronomy term given to the effect which occurs when water crystals in the upper atmosphere make it look like there is halo around the sun i believe, but it could also be when it makes it look like there are two suns.
I'm just bored, look on google. I might be right tho.
Linux is dead.
LU
yet they benchmark a card built around it.
i mean, play a game written for the gf3 on a gf4 system. its slower then on a gf4 because the gf4 is lacking the features the gf3 has so it has to take a non optimal path.
the specs from the new matrox card suggest a more general aproach to the rendering pipeline, not that fixed hardware nvidia vertex/pixelshader crap.
just wait until its actualy USED.
However, there are always problems:
1. The hardware never lives up.
2. Three words: drivers, drivers, drivers. Or, should I have just said "software, software, software"? Putting out beta drivers and leaving them out there for over a year without a final release is par for the course at Matrox. OpenGL promises, you say? Yeah, been there, done that.
I won't trust Matrox ever again. I was screwed by two generations of cards. Yeah, yeah, shame on me for trusting them a second time.
You're right - there's not much point in having FPS exceed the refresh rate of your monitor. I think that quick spin-arounds may be slightly smoother, but it's pretty subjective - that's just the way it feels to me.
However, those refresh rates are an average over many different sorts of scenes and can drop much lower during intensive scenes - e.g. if you just walked into a rocket arena with 30 guys throwing rockets at each other with particles nice lighting and everything, that's when you probably won't be getting 150 FPS anymore. That's when you really need those FPS.
Also, most people want a bit of headroom for future games.
As well, some people want more quality - better lighting, more polys, etc. If your card can do 150 FPS without anti-aliasing, maybe it can do 85 FPS with funky lighting and AA on (just an example).
However, your original point is correct - excess frame rate beyond your monitor's refresh is not really visible, but the extra power comes in handy for other things.
MeepMeep
How else would you like these benchmarks done, they don't exactly hand out Doom 3 Alpha binaries to just anyone. Besides, if the Parhelia doesn't generate massive framerates on 'old tech' like Quake 3, what do you think it's going to do on Doom 3?
Who cares what Carmack thinks, he's not about to say "The Parhelia is the only card for Doom 3" theres no way he'd shoot himself in the foot like that. Think of how many people get software bundles with their new videocards, and often these include games like Quake and it's brethren, but if Carmack makes it known that he's optimized or somehow skewed it so that it's better on the Parhelia, you're not going to find a lot of people getting special versions of Doom 3 with their cards. Unreal Tournament 2003 maybe?
As for waking up and finding life after Quake 3.. what would you suggest these people play? Most everything else is just a clone, or a licensed clone. If you want a fucking beautiful slideshow, theres this game called Myst you might enjoy, and if you want to show off your nice new card, RealMyst.
"You want great looking Anti-Aliasing? The Parhelia is for you. You want killer frame rates? You might have to look elsewhere."
Why does this so vividly remind me of 3Dfx vs NVidia a little over two years ago ? 3Dfx had their uber-AA system, but it would drag Quake3 to about 8 frames per second while the butt-ugly TNT2 just cruised along at a clean 40 fps (which was remarkable back then). 3Dfx collapsed months later when they learned one of the golden rules of computing : quantity over quality.
Granted, Matrox' prime market isn't the gaming sector, they've truly carved their throne in the business sector, filling in the gap left behind by Number Nine, but now they're trying to market at the gamers with this feature-packed chipset, yet I fear they're going to fall flat on their ass just like last time. If we've learned anything from NVidia, it's that people are willing to buy gobs of GPU power at insane prices. Your Geforce2 is too slow to play UT2 in 1600x1280 ? Then get the Geforce4, with two GPUs this time for more power.
If Matrox wants a share of my gaming budget, they'll have to start putting more raw goodness into their boards. Heck, just figure out how to link two or four Matrox GPUs and make then spew pixels like there's no tomorrow.
-Billco, Fnarg.com
I am going to wait to see what the overclocking results are like.... I have a feeling the P-512 is loosing due to its lower clock speed rahter than its hardware features/setup. Matrox are rumoured to release a FULL DX 9.0 hardware compliant card next year on a 0.13u process with much higher clock speeds.
Going by what I have seen so far, the P-512 does well on possible "future" applications (aka games)but poorly compared cards with higher clock speeds when using "current" applications. Think about how the GF3 was slower than a GF2Ultra on DX7 games due due to the GF3 being 200MHz while the GF2U was 250MHz. Once DX8 applications came along, the GF3 spanked the GF2U thanks to the features unique to GF3 (at that time)
The Parhelia is Matrox's first attempt at a competitive 3D card. As the process shrinks, the speeds will go up. And the drivers will mature over time.
How much better it'll get is a valid speculative point. Did they hire any of the old 3Dfx crew?
I seriously doubt that people are buying these cards to play Quake or Serious Sam, so why is it always these games that get benchmarked?
Because various 3D engines use different technologies, and these are in turn supported differently by the card manufacturers, it's not possible to simply extrapolate from Quake results to NWN.