View Full Version : Tom's Does XGI
Originally posted by $E Mun$
Another Parahelia...<sigh>
No, the Parhelia was an interesting card with some amazing features, nice AA, and decent performence, but at a high price. The Volari appears to be a card with no new features, sporatic performence ranging from good to unplayable, slow AA, and a medium price-point. I really don't see why everyone was hating on the Parhelia, the performence was fine for the time (around GF3 speeds) but the AA was superb and the multi-monitor support was great, it was just expensive and had a few small flaws (limited Aniso).
I really don't know how the Volari is going to be even moderatly compedative when they're relying on things like Super-Sample AA. I mean the 8500 BARELY got away with using that back when it was released, how do they expect a card priced higher then the mid-range to even come close to the offerings by ATI and Nvidia? If they don't do anything that improves the SSAA on the thing, 2x AA on the Volari will have the same performence hit as 4xS AA on the GFFX, and the FX's AA modes are considered to be lackluster and slow compaired to ATI's. I'm really hoping the Volari will pull ahead and suprise me, but when it relies on such outdated technology for some things I really can't see it becoming a choice for anyone that's seen a benchmark on it and knows about it's limitations. Oh well, at least it supports DX9, so that's a plus. Though it's DX9 speed seems to be below that of the FX5200...
ChrisRay
11-07-03, 02:24 PM
I hope I see some improvements performance wise. I mean they still look interesting, But this preview provides plenty of FUD
3dfxfan
11-07-03, 02:46 PM
The card will be available soon, no need to be alarmed.
digitalwanderer, I haven't got one either, so there's no need to panic. A source in XGI told me that these sites actually got a board as they went to Systems 2003 in GERMANY (that's why the first reviews were from GERMAN SITES).
Did you notice Tom used the latest drivers for everything except ATI's cards? Not only that but he used two different driver sets (Cat 3.7 & 3.8) so he could pick and choose the results he wanted to print. Why does he keep getting away with these tacktics?
The XGI cards are dead. Too little, too late. Who wants to have to use two power supply rails just to get a card that underperforms? Who would actually replace their current ATI or nVidia based card with this? Unless they have some really cool features they are holding back on us, this card will be a flop.
Originally posted by TheOneKEA
Seriously, though, there are some problems, IMO, with that review:
1. THG wrote it
2. The drivers are still beta and unfinished
3. THG wrote it
4. It was an engineering sample
5. THG wrote it
6. THG wrote it
7. THG wrote it
When the [H] tells me that the card sUx0rzZ, then I'll join in the moaning and groaning.
Has little to do with THG.
Other sites have posted news bits saying that the XGI Duo V8's targeted performance is between an FX5700 and a FX5900U (Or something similar). XGI isn't expecting to take the performance crown, which is a reasonable goal for them.
Regardless of THG's article (Which I agree, is too preliminary)... you CAN'T sell that for $300, when you can get a better performing 9800 Pro for $300.... and you DEFINETLY CAN'T sell it for $500.
Who the hell is going to buy a 9600XT+ or FX5700U+ for $500???
(Besides outright haters of both nVidia and ATi)
Regards,
Taz
goofer456
11-07-03, 04:30 PM
Originally posted by TheOneKEA
Seriously, though, there are some problems, IMO, with that review:
1. THG wrote it
2. The drivers are still beta and unfinished
3. THG wrote it
4. It was an engineering sample
5. THG wrote it
6. THG wrote it
7. THG wrote it
When the [H] tells me that the card sUx0rzZ, then I'll join in the moaning and groaning.
LOL:p
TheOneKEA
11-07-03, 04:52 PM
Originally posted by TheTaz
Has little to do with THG.
I haven't visited THG in a long time; some of their articles put me off a while ago. Not only that, but I have read in quite a few places that the reviews are somewhat shoddy compared to the reviews done by other sites. So THG is not on my short list of sites to visit.
Originally posted by TheTaz
Other sites have posted news bits saying that the XGI Duo V8's targeted performance is between an FX5700 and a FX5900U (Or something similar). XGI isn't expecting to take the performance crown, which is a reasonable goal for them.
Right. Even if they only put pressue on ATi and NVIDIA in the budget/mainstream end, that's still better than only having two really worthy vendors to pick from.
Originally posted by TheTaz
Regardless of THG's article (Which I agree, is too preliminary)... you CAN'T sell that for $300, when you can get a better performing 9800 Pro for $300.... and you DEFINETLY CAN'T sell it for $500.
Who the hell is going to buy a 9600XT+ or FX5700U+ for $500???
(Besides outright haters of both nVidia and ATi)
Regards,
Taz
I doubt those pricepoints are stable. Let's just wait for someone like $FAVSITE to review thr card and issue a verdict. Having only one review to go on isn't really a good idea, IMO.
Hellbinder
11-07-03, 05:14 PM
Guys, If you really look at their game performance.. Its BAAAAD. Forget the decent score in the Flyby. That matters not a hill of beans. Look at their actual in game scores. They are a JOKE. I am honestly a little upset. This chip according to their stated specs should be performing way better than this. I mean Seriously.. Who do they think they are kidding here? 500$ for a card that loses to a FX5600U most of the time? With 50 FPS in games that the other 500$ cards are getting 100 FPS??
These cards are a complete WASTE Of everyones time if these numbers are even close to accurate. Imagine what their single Chip solutions will perform like. Its irritating becuase it is obvious what they should have targeted, what they would need under the hood to deliver the goods etc..
Its a sad sad joke and the people at XGI have to know that themselves.
Hellbinder
11-07-03, 05:23 PM
I guess i would also like to say that some of Toms (lars) Numbers are a complete JOKE. I am not going to say he is a Biased Stooge, but i will say that if his numbers are not "Fixed" Ill eat my dog.
5600U beating a 9700pro in Q3 is a JOKE Notice its lars "custom" level. 50$ says Nvidia provided him with this "Custom" level. This is also one of the only sites i have seen where the fx59xx's Beat a 9800 in Aquamark03. In fact you can follow this pattern all the way through his entire set of numbers. They have unusually high results on Nvidia boards. This is nearly ALWAYS the case with a Lars review. I would not be supprised if he is tweaking the texture details down a little with riva Tuner or something for the Nvidia cards.
Seriously..
Originally posted by TheOneKEA
I doubt those pricepoints are stable. Let's just wait for someone like $FAVSITE to review thr card and issue a verdict. Having only one review to go on isn't really a good idea, IMO.
Well... that's what I'm b!tching about... heh... IF those prices are accurate... the peeps at XGI are fricken wacked! :D
Regards,
Taz
ChrisRay
11-07-03, 06:48 PM
Originally posted by Hellbinder
I guess i would also like to say that some of Toms (lars) Numbers are a complete JOKE. I am not going to say he is a Biased Stooge, but i will say that if his numbers are not "Fixed" Ill eat my dog.
5600U beating a 9700pro in Q3 is a JOKE Notice its lars "custom" level. 50$ says Nvidia provided him with this "Custom" level. This is also one of the only sites i have seen where the fx59xx's Beat a 9800 in Aquamark03. In fact you can follow this pattern all the way through his entire set of numbers. They have unusually high results on Nvidia boards. This is nearly ALWAYS the case with a Lars review. I would not be supprised if he is tweaking the texture details down a little with riva Tuner or something for the Nvidia cards.
Seriously..
Hehe good Ole HB :) You're conspiracy theories are great you know :) :cool:
But I can agree with you. Judging by these results, All I can say is they dont look too promising. Unless XGI takes some Nvidia tactics in hand and aggressively optimises its software.
ChrisRay
11-07-03, 06:50 PM
Jesus, I just read this review, Why, oh WHY on earth did they let Anyone have their card to review it? If they knew their drivers were this bad, Something has got to be awry, It's specs simply cant produce performance that bad in UT2003..
Johnmcl7
11-07-03, 07:22 PM
I'm surprised they gave a card to THG for a review as well, I do appreciate they need to pull some stopgap measures while they try and get their retail boards ready but this sort of review could do a lot more damage than good. As others have pointed out their driver team are going to have their work cut out, I was surprised at the sheer amount of problems, no AA, AF, and bugs everywhere with games and benchmarks.
As to the review in general, I can't believe anyone is still using Quake3 as a benchmark. It's had its day, and Q3 performance clearly does not give any sort of indication to Q3 based games (which there are still quite a few of).
John
Looking at 3DMark03 score the card should be at the level of Radeon 9700Pro which is good. I think we should wait the final drivers, before any conclusion...
Originally posted by lukar
Looking at 3DMark03 score the card should be at the level of Radeon 9700Pro which is good. I think we should wait the final drivers, before any conclusion...
ROFLMAO
CaNn 3Ye pL4y3 Tr3e-de3 M4aarK to0, PlZ? :D
Seriously tho,
Going by the major difference between the 3DMark scores and the actual games... Either XGI is *cough* "optimizing" *cough* for 3DMark, nVidia style.... OR... it proves overwhelmingly that 3Dmark is totally invalid, as a benchmark.
/shrug
Taz
digitalwanderer
11-07-03, 09:49 PM
Originally posted by TheOneKEA
Seriously, though, there are some problems, IMO, with that review:
1. THG wrote it
2. The drivers are still beta and unfinished
3. THG wrote it
4. It was an engineering sample
5. THG wrote it
6. THG wrote it
7. THG wrote it
When the [H] tells me that the card sUx0rzZ, then I'll join in the moaning and groaning.
Yeah, but it was wrote by THG. :rolleyes:
ChrisRay
11-07-03, 10:22 PM
Im not too concerned about the Unreal Tournament 2003 results. Even if the card isnt a big hit. a Card of that spec simply cant perform so slowly in such a game :)
Hellbinder
11-07-03, 11:37 PM
Originally posted by ChrisRay
Im not too concerned about the Unreal Tournament 2003 results. Even if the card isnt a big hit. a Card of that spec simply cant perform so slowly in such a game :)
Thats the Problem.
How Truthfull are the Specs? This is supposed to be a 16 pipeline card.
Also Consider the IQ shots. They are REALLY bad on the Xabre oops i mean Volari. They look like they are using some "tex-Turbo" Trick just to get the results they are. That is not a good sign.
I honestly really Expected them to deliver a lot better than this. Not that they would win everything.. but come on.
ChrisRay
11-08-03, 12:18 AM
Originally posted by Hellbinder
Thats the Problem.
How Truthfull are the Specs? This is supposed to be a 16 pipeline card.
Also Consider the IQ shots. They are REALLY bad on the Xabre oops i mean Volari. They look like they are using some "tex-Turbo" Trick just to get the results they are. That is not a good sign.
I honestly really Expected them to deliver a lot better than this. Not that they would win everything.. but come on.
Well I can only assume on this. I would assume under worse case scenerios this could also be an 8x2 card (Virtual Pipeline Nvidia crap) you get the jist.
It's really gonna be hard to determine the specs. I noted the texture problems. For now I have to give them the benefit of the doubt. That they may deliver, But who knows. XGI may even become a bigger evil than Nvidia around here if these scores keep up. But really. If it's bandwith is even remotely correct (Could they be decieving us with their bandwith numbers?)
That some of these tests have to do better, Also, Are to we to assume it's forcing Pixel Shader 2.0 precision? Is it using Pixel Shader 1.3 in demos that require it? Just Imagine it. UT2003 uses Pixel shaders to speed up rendering.. Imagine if it were forcing Pixel Shader 2.0?
Blah So much speculating at this point. And alot of FUD, We can only wait and see ;)
Originally posted by AthlonXP1800
Tomshardware is mistaked about the price, the top model is not going to cost $300, it will cost near $500, see my attachment for the picture.
Well with all that benchmarks, the top model look alot worsen for $500, it will gonna flop plus FSAA not working, maybe it will never working with retail drivers.
Some individuals here just like to give crap to non-NV companies.
Remember the first price speculation on NV30? 700€ and that was pre-launch availability prices. With a site that actually had that price.
But it's NV so of course it can't be serious :rolleyes:
But now it's XGI and we have the same hill-billy **** going again and since it's not NV it's totally ok and everyone will take it seriously :rolleyes:
And as mentioned, some individuals here really like to toss crap at non-nv companies within the graphics card industry.
3dfxfan
11-08-03, 07:39 AM
Furthermore some of Tom's benchmark results go against what Tecchanel and PCWelt say, and I rather trust these two than Tom...
ChrisRay
11-08-03, 11:40 AM
Originally posted by Unit01
Some individuals here just like to give crap to non-NV companies.
Remember the first price speculation on NV30? 700€ and that was pre-launch availability prices. With a site that actually had that price.
But it's NV so of course it can't be serious :rolleyes:
But now it's XGI and we have the same hill-billy **** going again and since it's not NV it's totally ok and everyone will take it seriously :rolleyes:
And as mentioned, some individuals here really like to toss crap at non-nv companies within the graphics card industry.
Why is this turning into an anti Nvidia thread? I mean the product performs lousy compared to ATI as well you know. There are plenty of ATI users who are unsatisfied with what they have seen.
This has absolutely nothing to do with Nvidia really.
Hellbinder
11-08-03, 04:12 PM
Originally posted by ChrisRay
Why is this turning into an anti Nvidia thread? I mean the product performs lousy compared to ATI as well you know. There are plenty of ATI users who are unsatisfied with what they have seen.
This has absolutely nothing to do with Nvidia really.
Th only connection this has to Nvidia or ATi is thats who they XGI are Competing for our cash. The comparrison simply has to be made.
I suppose it is possible that XGi might need another respin to solve some hardware problems. Persoanlly I cant see a Driver at this stage in the game taking a score form 7 FPS to 40 FPS etc.. We are talking about the most basic of Direct X and OpenGL compatability here. If the card is really 16 pipelines then it damn well should get better than 7 FPS in any game period.
I am not closing the door here on these guys just yet. However.. If they dont get their act together until february then it is a TOTAL lost cause worse than any release in recent history if not ever. I should note that I am specifically talking about their Duo Ultra card. They may still salvage something if they can get their Volari V8 out the door with decent mid range performance and undercut everyone else.
emailthatguy
11-08-03, 08:45 PM
ok im trying to think of something positive here for these new guys... umm..
umm.....
they doing a lot better than bitboys? =p
StealthHawk
11-08-03, 10:27 PM
Originally posted by digitalwanderer
Close, but I think you should make that "to keep ppl interested till a now much delayed launch date?".
I think XGI has realized that their drivers ain't ready for primetime and they need more time to work on 'em...I don't expect them to launch this month anymore and I'm getting kind of doubtful they'll make it this year. :(
Well, Tom's Hardware agrees with you about a delay. From the conclusion:
http://www.tomshardware.com/graphic/20031107/xgi-volari-21.html
XGI's driver team definitely has its work cut out for it. We can expect to see final boards with shipping status and WHQL drivers within the next one or two months, or so. That gives XGI's driver programmers a bit of time to iron out the kinks.
vBulletin® v3.7.1, Copyright ©2000-2010, Jelsoft Enterprises Ltd.