ukgamer 5900 review - 3 5900/5900u's

[Sazar]

http://www.ukgamer.com/article.php4?id=5&page=1

just saw this... got link off of [H]

I read through the whole thing and the same thing struck me again and again... this card can't possibly be as slow as it is being shown to be in the tests...

I can understand a flipflopping of results with the 9800pro in some situations but the 3 nv35 iterations were having tremendous difficulty keeping up with the 9700pro...

I can't for the life of me figure this one out... is there something in the benchmarks they did wrong or what ? these are retail boards after all...

also this was rather interesting...

General conception is that the difference between a FX 5900 and a FX 5900 Ultra are just clock speeds. If that were the case then it's surprising and shocking to see when we overclocked the 5900s past the 450MHz/425MHz (850MHz DDR) of the Ultra, we saw very little performance gain, let alone reach or surpass 5900 Ultra speeds. We will investigate this phenomenon further in the coming days and try and give you our results.


I refuse to comment on the shadermark issue :) since obviously ukgamer is unaware of such a thing.. or the splinter cell issue which many have called a genuine bug...

please post relevant comments w/o troll baiting/flaming posts...

cheers..

[StealthHawk]

I will say that the reviwers have poor graph making skills. Other than that, I will not comment.

[aapo]

Originally posted by StealthHawk
I will say that the reviwers have poor graph making skills.

Yep, I certainly agree with you. Other than that, I liked the review. They didn't bash the cards, and actually played the games and told how they did like them. It seems that they had fun and enjoyed testing the cards, I liked especially this remark:

Therefore when we loaded up Unreal Tournament 2003 and played a few hours of deathmatch, (on company time? - Ed.) and although this is fairly subjective; we weren't able to see any real difference whilst playing.

Regardless of the we-don't-now-or-care-about-technical-issues attitude, they reached the same conclusion as I did: the FX5900 series is generally ok, but a little too expensive considering the competition. Maybe the reviewers should generally do some fragging "on the company time" and report how they feel about a card. :)

[BigFish7]

it's a really bad review... :(

[aapo]

Originally posted by BigFish7
it's a really bad review... :(

I'm surprised. It's obvious that about half of the graphs are surprisingly poorly formatted and very hard to read, but the data and the conclusions from the data seem to be ok. Would you care to elaborate your opinion a little?

[Sazar]

Originally posted by StealthHawk
I will say that the reviwers have poor graph making skills. Other than that, I will not comment.

I would agree with that assessment.. its obvious that the graphs could have been split up or designed better :)

too much clutter

[G6-200]

They could have just skipped all the "performance" modes completely and just tested max quality, no one in their right mind paying $500 for a graphics card would use them.

G6

[IISquintsII]

Man I got to sell this 9500 pro to my brother and get me a 9700 pro!

Edit: I said "give" instead of "sell". I like "sell' better :)

[Dark Jedi]

I thought it was a nice round up... will agree graphs could have been laid out better, but they were honest and fair in their opinions, plus it gives us a chance to look at how various manufacturers implimented their 5900 designs.

Gainward card looks sweet... if a little pricey. (£400? Yikes) ;)