XGI Volari V8 Duo Review
- Thread starter VIAN
- Start date
This review shows it in a slightly more competitive light than that Deutsch review from last week. In the benchmarks where the drivers actually work, the Volari does a decent job of keeping up with the 9800 Pro. And it wasn't even operating at the expected retail speeds! The core was slower by 25MHz and the RAM was under by 50. Once the driver issues are worked out this could be a good high-end alternative if Tom's is right about that $300 MSRP.
Ah, finally an *english* review. THG is better than nothing.
Some ugly results here, like check out Wolfenstein ET . Remind anyone of the Rage Fury MAXX?
I don't think in 1-2 months XGI will be able to iron out most of these kinks and optomize everything for their 2-GPU design. As suspected, their drivers are still seemingly in infancy.
At least the price point has been dropped to ~$300 US, which should make it more competitive (but still dead in the water compared to the 9800np and 5900nu).
Plus we still don't know how the single Volari will perform. My guess is that compared to ATI and Nvidia's single GPU designs, it won't be pretty.
I think the Volari Duo will be an "interesting" option at $300, but ultimately drivers will be the noose around their neck. Hopefully they prove me wrong...
Some ugly results here, like check out Wolfenstein ET . Remind anyone of the Rage Fury MAXX?
I don't think in 1-2 months XGI will be able to iron out most of these kinks and optomize everything for their 2-GPU design. As suspected, their drivers are still seemingly in infancy.
At least the price point has been dropped to ~$300 US, which should make it more competitive (but still dead in the water compared to the 9800np and 5900nu).
Plus we still don't know how the single Volari will perform. My guess is that compared to ATI and Nvidia's single GPU designs, it won't be pretty.
I think the Volari Duo will be an "interesting" option at $300, but ultimately drivers will be the noose around their neck. Hopefully they prove me wrong...
How can the price have "dropped" on a product that isn't for sale yet? Like I said in the other thread, your prediction of $399 was just too high for SIS/Trident.
Anyway, I don't think we can really speculate anything with the drivers used in that review. It's pretty obvious the card is barely working at this point. Starting to remind me more of the S3 Savage launch than the MAXX. The MAXX was the second fastest card you could buy when it arrived, and had better IQ the the first. (GF1)
I think it's good to have another card in the mix, it means we can have something else to look at when it comes to comparing how different cards render scenes, it might make it easier to spot "optimisations" when there are 3 different screens to look at (assuming that all 3 cards manage to work in games, as ATi and nVidia even now have issues sometimesof cards not working at all)
the preformece is in some cases pretty decent, that shows that the card is not a compleate pos, @ full speed and with some decent drivers this could get relly good, not saying it will just saying it might... i highly doubt i will buy one no matter how good/bad it gets but another player in the industry is a relly good idea if u ask me.
any1 else noticed XGI's plan: world domination by 2007 (in the grapchics industry that is
) shows that they aint gonna give up that easly. in a few generations they will probably be "up there" with NV and ATI if they dont run out of green that is
any1 else noticed XGI's plan: world domination by 2007 (in the grapchics industry that is
) shows that they aint gonna give up that easly. in a few generations they will probably be "up there" with NV and ATI if they dont run out of green that is I would'nt touch that thing with TWO ten foot poles.
However, it's fairly new and maybe they can work out their driver woes, though I really doubt it.
However, it's fairly new and maybe they can work out their driver woes, though I really doubt it.
Funny that once you stray from the default benchmarks the numbers tank.
UT - default bench = FAST
UT - custom bench = SLOW
AquaMark = FAST
Halo = SLOW
Q3 = FAST
RTCW:ET = SLOW
X2 = SLOW
3DMARK = FAST
This remindes me of the Savage 2000 "preview" on THG.
UT - default bench = FAST
UT - custom bench = SLOW
AquaMark = FAST
Halo = SLOW
Q3 = FAST
RTCW:ET = SLOW
X2 = SLOW
3DMARK = FAST
This remindes me of the Savage 2000 "preview" on THG.
They don't even have AA/AF in the drivers yet, they're nowhere near finished.
And according to the ATi commentary, more videocard reviews use RTCW than Aquamark, so optimising for Aquamark over RTCW seems silly.
Oh, and you notice the "beta drivers" notices under all the graphs? They're beta drivers, not finished.
And the card wasn't even clocked at "proper" speeds.
You're already saying a beta product with beta drivers is cheating, c'mon, get some sense into your head.
The fact that they are beta tells us nothing as to why the scores tank when you suddenly use a different timedemo
Interesting. Been waiting for this as well. I have a few concerns.
1) Image quality is NOT up to par. When/if Trilinear is fixed, does this mean performance will drop quite a bit? Will it still be able to compete in the current bracket when trilinear filtering, or will it seem more of a midrange solution?
2) The card appears to be at the complete mercy of the drivers right now....it performs well in some areas and tanks in others. Driver support will have to be solid for this card to fight its way into the market.
3) It's promising, at least. If the company can pull off some driver magic and there's more performance in the silicon to be tapped, we could have a third GPU option overnight. Time will tell, but here's hoping
1) Image quality is NOT up to par. When/if Trilinear is fixed, does this mean performance will drop quite a bit? Will it still be able to compete in the current bracket when trilinear filtering, or will it seem more of a midrange solution?
2) The card appears to be at the complete mercy of the drivers right now....it performs well in some areas and tanks in others. Driver support will have to be solid for this card to fight its way into the market.
3) It's promising, at least. If the company can pull off some driver magic and there's more performance in the silicon to be tapped, we could have a third GPU option overnight. Time will tell, but here's hoping
Sheesh BFG, a $299 launch price and performance potentially in 9700 Pro range with decent drivers and you're mad about price?
Its sooo hard for a new graphics chip designer to make it in today's oligopoly of Ati and Nvidia. You really gotta beat these two to exist, like providing a graphics card on par and at much lower prices. But I admit that the two's competition, makes good products and drives prices down. Although another piranha in the tank could help the consumer even more !
!I was referring to the review over at Toms that showed the card actually performing on par with a 9700 at some standard benchmarks, then mysteriously losing 4/5 of its performance at non standard.
My point was that if they can get all the performance at 9700 level, $299 seems like a fair launch price to me.
My point was that if they can get all the performance at 9700 level, $299 seems like a fair launch price to me.
Those results stink of heavy (and most likley questionable) optimizations. All the default normal benchmarks fair well, but run anything new and the results fall apart. My guess is that if you ran this card thru a ton of real world apps you'd be better off spending your $299 on a mid-range nVidia or ATI product.
I also find it funny that Tom's review seems to imply that XGI started the design with dual chips in mind. I don't know of any company that "wants" to make a dual chip board just to compete with main-stream parts. Dual GPU's is often the results of a meeting that goes like, "oh crap, our performance sucks, we don't have the HW to compete". Then someone says "Throw another chip on there". Their dies are huge and that board has to be extremely expensive to make (1lb of copper, a billion traces, two power connectors, 2 fans, god know's how many layers, etc).
I also find it funny that Tom's review seems to imply that XGI started the design with dual chips in mind. I don't know of any company that "wants" to make a dual chip board just to compete with main-stream parts. Dual GPU's is often the results of a meeting that goes like, "oh crap, our performance sucks, we don't have the HW to compete". Then someone says "Throw another chip on there". Their dies are huge and that board has to be extremely expensive to make (1lb of copper, a billion traces, two power connectors, 2 fans, god know's how many layers, etc).
or they have a list of optimizations and are going down the list, starting with well known benchmarks, and they aren't done yet.
OK the guy at toms is an idiot. packaging and dies are not the same thing. if memory servers this chip uses over 1100 pins. that is more than an opteron with 3 HT links and 2 memory contollers. that is why the PACKAGE is large. this says nothing about the actual die size.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 24K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 20K
-
-
-
Facebook
Twitter