View Full Version : Gainward 5950 Ultra 1800XP Review
Post your comments/questions here for this review. (http://www.nvnews.net/reviews/gainward_fx_5950_golden_sample/index.shtml) Thanks!
Dude me just loves them dual fan dual LED flame livery HSF :D
Great work :D
the difference between 2.2 and 2 ns
calculating speed rating in MHz
1000/2.2=454.5 MHz or 909mhz DDR
1000/2=500MHz or 1000mhz DDR
so the lower rated memory for more stable overclockabilty?
Thanks, yeah that is a cool HSF and the fan leds are a nice touch.
Hey, didn't you test with 52.16 vs. 53.03 in 3DMark03 and found that difference was negligible? Or did you see that tested somewhere? Or am I just out of my mind? :D
Tim informed me that:
53.03 reenables the NVIDIA cheats in 3DM03 build 340, which is why they're not approved. I would highly recommend pulling that bit ASAP. you get about a 20% performance increase with 53.03 over 52.16.
So, I've edited that paragraph. I didn't have time to test with 52.16 drivers on this but I will and will update that section of the review.
Originally posted by MUYA
Dude me just loves them dual fan dual LED flame livery HSF :D
Great work :D
the difference between 2.2 and 2 ns
calculating speed rating in MHz
1000/2.2=454.5 MHz or 909mhz DDR
1000/2=500MHz or 1000mhz DDR
so the lower rated memory for more stable overclockabilty?
Yes, that's probably the case but haven't confirmed it. I know that's typical but didn't want to assume.
Nice review!
Yields on Hynix 2.2 ram must really be good now with all the 5900nu's and Ultra's getting clocks over 1000. My 5900U hits the wall at 960. I have used different type sinks, thermal compounds, thermal tape, you name it but it still hits the wall at the same spot.
Yes, I was pretty impressed to see it run stable at 1050. That was with me hammering it over and over again and in long loops.
Gainward makes nice boards. :)
Good review though personally I'd like some reference numbers (standard 5950 Ultra and a 9800 XT for an example).
Here in Sweden Gainward has a con above all else though. They're often MUCH more expensive than other. Sometime (when it comes to the high end boards) they're up to (and on some rare occasions) 100 Euros more expensive than other brands.
I seacrhed the prices of 5950's on the link in the conclusions and it seems as if this is not the case elsewhere which makes me wonder why they're treating us swedes this way. :)
For an example:
Right now the Leadtek 5950 Ultra is roughly 3800 SEK (~420 Euros) on avarage while the Gainward 5950 Ultra is roughly 5000 (~550 Euros) on avarage.
(And mind you the leadtek board also has ViVo, good overclocking margin and a massive software bundle so it's not like it's an inferiour product or anything)
Still, the funny thing is: Gainward are doing great in Sweden. hehe
Originally posted by AnteP
Gainward makes nice boards. :)
Good review though personally I'd like some reference numbers (standard 5950 Ultra and a 9800 XT for an example).
Thanks AnteP. I would have liked that too but I have neither a standard 5950 nor a 9800XT.
Originally posted by MaxPower
Thanks AnteP. I would have liked that too but I have neither a standard 5950 nor a 9800XT.
what, you're not a millionaire? :afro2:
LOL :D Don't I wish! :cry: :)
Nice one.
And sure, Gainward is saweet!
UT2003 sure looks crisp at 16x12 with some AA/AF on top. :p
But the only thing DM-IceTomb is good for is benching. Yikes, aweful map, zero gameplay. :)
The Baron
02-02-04, 09:39 AM
Icetomb's one of the better Invasion maps, IMO, especially when you're playing low-grav instagib at 135 gamespeed...
ATI_Dude
02-02-04, 04:52 PM
I was a little disappointed to read that the memory maxes out at 1050 MHz. I expected it to to come close to 1100 MHz especially since it's more expensive than the other 5950 Ultras. My Leadtek A380 Ultra easily matches those clockspeeds and it costs about $ 50-100 less.
One question:
How is 2D image quality? Is it superior to the other 5950 Ultras and how does it stack up to the 9800 XT?
I actually remained what I thought was a bit conservative until I could confirm whether or not the memory timings were, in fact 2.2ns or 2.0ns. I just received word from Gainward that it is a typo on that chart so this card does have 2.0ns rated memory. Since Halo was a bit finicky (as was KotoR) I decided not to push past 1050. Now that I know the speed is 2.0ns I will re-run my CoD and UT2k3 benchmark tests at higher speeds. Look for an update within the next few days.
To your last comment, see my post two posts back ;) Thanks for the feedback and I will see how much higher I can push it past 1100MHz.
Well done MAX . :)
just a question : In Halo you mentioned PS1.4 as DX8.0 mode , isnt it a DX8.1's ?
i thought that PS1.1-1.3 were in DX8.0 , 1.4 DX8.1 , 2.0 DX9.0 right ?
Originally posted by SH64
Well one MAX . :)
just a question : In Halo you mentioned PS1.4 as DX8.0 mode , isnt it a DX8.1's ?
i thought that PS1.1-1.3 were in DX8.0 , 1.4 DX8.1 , 2.0 DX9.0 right ?
it's not uncommon to simply refer to DX8.1 and 8.0 as "dx 8" the difference is quite slim
in any case:
DX8 is PS1.0 and 1.1
DX8.1 is PS1.2-PS1.4
DX9 is PS2.0 and 3.0
Originally posted by AnteP
it's not uncommon to simply refer to DX8.1 and 8.0 as "dx 8" the difference is quite slim
in any case:
DX8 is PS1.0 and 1.1
DX8.1 is PS1.2-PS1.4
DX9 is PS2.0 and 3.0
i see ...
so how come The Inquirer site says : DX 8.1 introduced new shader models and upgraded them from PS 1.3 to 1.4 -- a huge boost for ATI at that time since they could render some shaders in less passes with PS 1.4 than Nvidia’s 1.1 or 1.3 shader models. ??
Originally posted by SH64
i see ...
so how come The Inquirer site says : ??
I'd assume the reason is simple: they don't know what they're talking about. :)
The GeForce 4 Ti cores supported PS 1.3 I assume that's what's causing the confusion.
Originally posted by SH64
Well done MAX . :)
just a question : In Halo you mentioned PS1.4 as DX8.0 mode , isnt it a DX8.1's ?
i thought that PS1.1-1.3 were in DX8.0 , 1.4 DX8.1 , 2.0 DX9.0 right ?
Thanks SH64. :) And thanks AnteP for fielding the questions. Yes, I just used DX8 as an abbreviation (I saved two characters with ".1") :D
Neutronbeam
02-03-04, 10:51 AM
Great review and helpful.
I researched 5950 cards since October 2003 and finally decided on the Gainward 1800/XP 5950 Golden Sample and got it last Friday from Directron---best price online I've found---but they raised the price AFTER I bought my card. Good shipping cost too.
Very happy with the card and the overclocking utility. Noise is not relevant as I just put in a louder case fan so the Gainward's noise is not a probelm.
Hey Neutronbeam, glad you liked the review and found it helpful. I appreciate the feedback. I like the EXPERTool overclocking utility as well. One minor gripe that I failed to mention in the review is that the slider bar/dial control options, while easy to use, don't give you much precision...you really have to tinker with them. I'd like to see either a spinner control that would allow you to increment by 1 or a textbox where you could key in the values.
Neutronbeam
02-03-04, 12:32 PM
Yeah, good idea---one reason I almost used Powerstrip instead of EXPERTool. Now if you want to get REALLY granular then you could have a sensor displaying temperature readout as clock speeds are increased.
One idea for a column---online buying tools. The tools--pricwatch, dealtime, pricegrabber, mysimon, etc.---have some big gaps. though pricewatch is quite good. I had to spend a lot of time going into sites to find or confirm prices, while is how I got the price I wanted on the Gainward 5950.
That is a cool idea to see realtime temp increases as you tweak the rates. Of course, the those temps would only reflect the current (basically idle state) and not how they would rise/fall within actual games. It could still give you a good rough idea though.
That would be an interesting article. Maybe someone around here would like to write one up on that topic.
Question, why png's for the not-so-important pictures?
Originally posted by volt
Question, why png's for the not-so-important pictures?
No real pressing reason. Could use GIFs but PNGs of graphs and charts are sometimes smaller than (256 color) GIFs (16 colors are really small but look nasty)...I didn't really compare the two in this case, just went with PNGs on non-photo type images since JPGs even at 10% compression muddy the graph/chart images up a bit.
vBulletin® v3.7.1, Copyright ©2000-2012, Jelsoft Enterprises Ltd.