FireGL2 Review

[Kodiak]

Regarding Tom's latest review of two cards:
I'm not a professional graphics card user. However, it seems to me that for a person who makes his/her living off his/her videocard, or for a company that has to install cards in 100s or 1000s of computers, the most important factor is not the raw performance, but rather reliability.
This was not mentioned at all in the review.
The Geforce2 GTS card has in some tests greatly surpassed the FireGL card, while being one fifth the price. So why don't companies buy it? Because they're not willing to mess around with mobos, drivers and setup as much as we are. They want to stick it in and have it work until they decide to stick something else in

It seems to me therefore that review has completely missed the requirements of its market audience. Just listing the features (available in the brochure) and running a few pre-made benchmarks is hardly an effort or a conclusive review of a $1200 investment...

Again though, I'm not a professional user, so I'm eager to be educated on how professional CAD designers etc use their cards and how useful was this review to them

 

[Guest]

I agree. From a business standpoint, stability is worth the extra cost in the long term.

 

[Guest]

i think you should read that article one more time carefully.
as the review pointed out the card IS specially engineered for the most CAD designers habbit/workflow.
creating/editing with wireframe model > texturing on final output.

designers working on frame models all the time.

FireGL series are designed for OpenGL apps. (for making games)
GF2 GTS aimmed to the game market. (for playing games)

you mentioned test should be some game test.
will you invest on a game card in your workplace ?
pls. tell me if you say yes to this. i'll be the first one to send you a application form.

<font color=orange>What do you think? </font color=orange>

 

[Guest]

The most important thing for people in the market for these cards is the performance in view of what they are doing. At the college I work at, they bought a batch of 50 Fire GL1's in one of the labs for some CAD workstations. The GL1's were the cheapest card that would meet their needs (at around $1000 Canadian at the time). I thought, "they're only 32MB video cards, why not just get any other card?" They explained to me that their needs are very specific. The typical way we measure performance for our gaming cards (ie. Geforce 2, etc.) is meaningless. They don't care about Quake performance or MDK2.

Incidentally, if you know of anyone thinking of getting a GL1, tell them NOOOO!!! we've replaced more than half of those 50 because of a "fan is blocked" error message. Diamond finally did a bios and driver update, which worked for a couple, but not all. They wouldn't replace all of them, now SonicBlue is making us send the defective ones back before they'll ship the replacements, so we have to wait for the new ones to come in before we can use the computers. Notice that the new Fire GL2 doesn't have a fan?!?!?! Coincidence?

 

[Kodiak]

re alexick post... actually, I think you should read my post again, not me read the article (or maybe I just didn't explain my point well

I fully agree that FireGL2 is not a gaming card, and again I fully agree that I wouldn't buy a gaming card for the business place -- so no need to correct me there

>> you mentioned test should be some game test<<
don't know where that came from... I wanted to suggest the opposite.

My point, in those benchmarks, a GAMING card has as often as not produced as good results as PRO card - at least on the benchmarks they used. So, since people are paying 4 more times for the Professional card, and not the gaming card, following a logic approach, it means they are not buying the Professional card because of benchmarks ('cause then they WOULD buy the gaming card. They are buying it for other reasons as well, and those reasons weren't covered by the review...

 

[Guest]

thank you for making points clear.

i think reasons not included in an review coz they try to make the article as neutral as possible.

<font color=orange>What do you think? </font color=orange>

 

[Guest]

Well I read it, and the emphasis is wrong. Rendering, and rendered models become more and more important with model complexity, and huge wireframes are unusable. You just can't figure out what you are looking at.

Also large amounts of polygons are unneeded until the model is done, and if you use NURBS based modeling, which is all any sane person would use, then adjusting the poly count is trivial and done at the end one a project. During construction the task is a repetitive handling of small, rendered, low poly portions of the model, rendered so you can see what the heck you are doing. Rendering is usually done using curves, so it looks/acts like the high poly finished model. So tons of polys aren't need to the end, either.

Which is why three expensive Elsa Gloria's are sitting on my desk, and my CAD kids are off using GF2's, happily. They threatened me if I took them back. This was a happy discovery when one of the Gloria's died, and we through a GF in one of the boxes as a stop-gap measure. Now it's all we use.

The benches were interesting, but off the mark. Almost crap, if I may say so.

Find out what really happening, don't just let the graphics card magazines tell you what you need. Walk into a design studio and ask the guy on the tube what he does all day. You might be amazed to know, he doesn't do what the largest advertiser in the graphics magazine says he does, or what the company that makes the software and demo's it at trade shows says he does.

 

[Guest]

thank you for your comments.
however, these comments are based on your own working enviroment and/or you own some GF2-powered box...
hm......


<font color=orange>What do you think? </font color=orange>