GF FX 5800 Ultra Vs. Radeon 9700 Pro

[Rogozhin]

HM

Well I've been running benchmarks over at nvnews.net with a guy that owns an fx5800ultra. WE used raven shield and UT2003 both at 1024x768 resolutions.

Used fraps to bench ravenshield and the built in benchmark of ut2003.

We tried with no aa and no af and I was about 30-40% faster in both games.

Then We maxed both aA and AF and I was 40-60% faster in both games.

This was with him using quality instead of application.

It was just sad but I did not say anything negative to him because he spent some good money on his rig and I just didn't want to bum him out.

My system

athlon xp 2100+ @ 1.9
768mb pc 2700
radeon 9700 non pro @ 330/315
soyo kt 333 lite
2 ibm 60gxps
fortissimo 2


His system

P4 2.4ghz 533FSB | MSI 648 Max, AGP8x | 512MB DDR400 Corsair CAS 2 | 40gb Maxtor 7200rpm | Liteon 16x DVD | Liteon 52x24x52 CDRW | Soundblaster Live 5.1 | Creative Inspire 4.1 | Enermax 430 Watts | EVGA GeforceFX 5800 Ultra 128MB


Rogo

 

[Rogozhin]

Well

My conclusion regarding the FX based on online reviews AND this real time benchmark is that the FX is not worth the money and is not even comparable to a 9700 non pro let alone a 9800 pro.

Rogo

Please tell me why I'm stating a logical fallacy, please

rogo

 

[blahblah]

Originally posted by: Rogozhin
Well

My conclusion regarding the FX based on online reviews AND this real time benchmark is that the FX is not worth the money and is not even comparable to a 9700 non pro let alone a 9800 pro.

Rogo

Please tell me why I'm stating a logical fallacy, please

rogo

Careful, you're gonna get the

from Rollo.

Anyways, let's just hope NV35 turns out to be a kick ass product.

We really need 2 or more competetive companies in this market. (Like Intel & AMD)

 

[Rogozhin]

I sure do hope the NV35 can pull out some big guns

rogo

 

[OpStar]

if your 9700 non pro can beat my FX Ultra I'll buy you a 9800 Pro.

I compared (if you wanna say that) to a 9700 Pro, cuz I owned one. I haven't had the enjoyment of a 9800 Pro as of yet.

3D Mark 2003 is looked upon differently (sorta like the FX). As it stands it is a DX 9 bench. Its used by 9800 Pro and FX owners
alike, and I have beaten quite a few 9800 pros (and lost to quite a few as well).

I use the 43.51s. The ONLY set that were <<cheating>> were the 42.68s, which you can't even find anymore.

Fan noise = personal opinion. My friend has 14 vantac tornadoes. I assure you my FX can't touch that. I even have it 12v modded
to run the fan 24/7.

nVdia did not say that FX was a failure. They said it "didn't live up to their expectations." That is true of everyone. If its such a flop
and horrible card as so many of you state (who don't own it) why can it keep up with a 9800 Pro?

What amazes me about this stupid IQ arguement is who in the hell do you know who plays at 16x12 with 6x aa and 8x ansio? I'd like to
meet them. Ansio and AA becomes less of a factor at high resolutions anyways.

I play at 10x7. That is rez I play at regardless of video card. At that rez I run 4x aa and 8x ansio in EVERY game at VERY playable frame rates.

The 9800 is a good, fast card. It isn't w/o problems. The FX is a good, fast card. It isn't without problems.

I agree the cooling solution is lackluster, I also hate that it runs so hot <gg ddr> but it is NOT a piece of junk it DOES NOT have HORRIBLE IQ.
(I'd like to meet the ppl who have actually seen the FX and 9800 pro side by side. Only thing I think would be noticeable really is the Radeons 6xaa)

All in all, you guys argue so much about something that doesn't matter. You advise other people how to spend the money they work for, and then
you ridicule for them if they don't agree with the masses (communism anyone?) Some of you ppl (and other forum monkeys) promote ATI products
without ever owning on.

What amazes me at most is how much people say OMFG THE FAN!! OMFG THE NOISE!! And you don't say DAMN I CAN'T PLAY <INSERT GAME HERE>
ON MY RADEON.

fanATics over at Rage3D sit around waiting for driver updates with a smile on their face cuz they own the fastest card you can't use.

"Hey I have a counter-strike problem" "Wait for Dets 3.3 that should fix it". When I owned my 9700 Pro (when it first came out) they were supposed
to fix it every driver revision.

All in all both cards have pluses and minuses.

You guys just generally try to make the FX into a huge flop that sucks so bad and is a waste of money blah blah.

Yeah. I paid a lot for it. 50 more than 9800 Pro. I play games everyday. Every one I own. No issues. Maybe a degree or less AA or Ansio. But
at least I'm playing while the guy that saved 50 bucks on his silent, cool running, 9800 Pro (worlds fastest video card :/) is waiting for driver revisions.

Have an opinion. You are entitled. I guess I just prefer educated ones.

Thanks

/end rant

 

[NYHoustonman]

Nice.

 

[Rogozhin]

"Have an opinion. You are entitled. I guess I just prefer educated ones."

Then what reviews are you reading?

the rest of your post response

I never once said anything about NOISE, HEAT, or SIZE.

I said that after comparing to a fellow (who could have lied and didn't, as well as myself) who owned an FX on a system that's better than mine, there is no way I'd spend $400 dollars on an FX (and that is the minimum I've seen them for, $450 is middle of the road for the fx).

My little $230 non pro is more than enough (it smoked the fx in the two benches I've run against real time users).

If you really did use a 9700 could you please post your benchmarks of both (did you use fraps txt log? on any) PLUS your 3dmark results (even though most people believe they are irrelevant it's still valid when proving you own a piece of hardware).

Thanks

Rogo

PS

and as rollo would say.

touting the FX as a great card will NOT

1.get you laid
2.into heaven
3.or high

rogo

 

[OpStar]

I tout it cuz a great card cuz it is. If the 9800 Pro is, so is the FX, those are the facts.

 

[Pete]

Originally posted by: chizow

Edit: An example of this is game's that have the "nVidia, the way its meant...." logo. They typically run better on nVidia cards relative to the respective part from the competition. Why's that? Because the developers used those cards as target hardware and used them throughout development to optimize performance and game code to take advantage of specific features for that part. Any bugs would be squashed through the course of development. Again, this is why I feel that nVidia has better drivers than ATi. I think the gap has closed significantly for game support, but I still don't think IHVs are fully testing for ATi boards yet. I haven't had a problem with any recently purchased games (6 months) on my 9700pro other than PlanetSide (which is still Beta, so no biggy), but I have had some issues with the hardware. They've all been corrected over the last 5 months, but still shows ATi isn't the priority when testing for hardware compatibility IMO.

Chiz

Quite a few developers in B3D have said the "Meant to Be Played" logo is added by the game's producer, not developer, so it doesn't necessarily reflect dev. platforms. In terms of opimizing performance: UT2K3 has that logo, and ATi cards spank nV ones pretty handily. IMO, you're confusing driver quality with developer support in the first half of the quoted passage. They're both part of end-user experience, of course, and that's really what counts; but if we're debating the specifics of game bugs, I'm sure quite a few R300 problems are due to it being a new architecture, as you said in the second half of the quoted passage. I'm guessing ATi's current dominance will lead to many more developers taking ATi's cards into primary (rather than secondary) consideration during QA.

 

[Pete]

Originally posted by: OpStar
I use the 43.51s. The ONLY set that were <<cheating>> were the 42.68s, which you can't even find anymore.

Wrong; in fact, the reverse may be true, in that I think the only drivers that didn't cheat were 43.00 and 43.03 (and they produced low 3DM03 scores). 42.xx cheated, and 43.xx resumed the cheating (and high scores).

nVdia did not say that FX was a failure. They said it "didn't live up to their expectations." That is true of everyone. If its such a flop
and horrible card as so many of you state (who don't own it) why can it keep up with a 9800 Pro?

I don't want to nitpick every one of your points, but I think it's clear the FX 5800U can't keep up with a 9800P with high levels of AA & AF, in both performance and IQ.

What amazes me about this stupid IQ arguement is who in the hell do you know who plays at 16x12 with 6x aa and 8x ansio? I'd like to
meet them. Ansio and AA becomes less of a factor at high resolutions anyways.

9700P & 9800P owners. ;P And though AA may become less important at high res (AF remains so, IMO), it's still a factor in larger displays. And ATi's bandwidth advantage, as well as its superior clock-for-clock architecture (and dollar-for-dollar cards), will provide a better ROI for ppl who buy $300+ video cards for high-quality images (which should now be considered high res and AF & AA).

I don't think everyone is trying to prove conclusively that the FX sucks, just that it's overall inferior to the 9700/9800 series. Some people, who aren't as concerned with saving $50 as they are with spending their time enjoying games rather than researching cards and tweaking driver settings, or who want better Linux drvers, or who want to stick with what's familiar to them (a known good quantity), will prefer an NV30 to an R3x0 solution, and they can certainly enjoy their FX (if the fan noise doesn't bother them ); others may have read a review or may be willing to take a chance with ATi, and it looks like they'll be rewarded with the currently overall better card (with its fair share of niggles, of course). Hey, nV can't win 'em all, and I'm sure NV35 will fix NV30's quirks, as nV did moving from the GF to the GF2.

 

[oldfart]

Fan noise = personal opinion. My friend has 14 vantac tornadoes. I assure you my FX can't touch that. I even have it 12v modded to run the fan 24/7

Ouch. So a 757 is not as loud as a 747? Having 14 high speed fans in a PC is just ridiculous as it hot wiring the dustbuster so it is always in high speed maximum noise mode. I guess you guys must really enjoy lots of noise. I'll pass on that. I kind of like my near silent system.

play at 10x7. That is rez I play at regardless of video card

That is a shame. Why buy a high performance card if you aren't even going to use its capabilities?

 

[OpStar]

Splinter Cell. That and like any enthusiast, I like to have the latest and greatest.

4600: 8x6 no aa no ansio maybe 30fps

FX Ultra: 10x7 4aa 8x ansio 30~50fps

9700 Pro: Wouldn't run

 

[CraigRT]

I like the 9700 Pro better... the stupid vacuum card just doesn't do it for me.

 

[oldfart]

Originally posted by: OpStar
Splinter Cell. That and like any enthusiast, I like to have the latest and greatest.

4600: 8x6 no aa no ansio maybe 30fps

FX Ultra: 10x7 4aa 8x ansio 30~50fps

9700 Pro: Wouldn't run

I downloaded the Splinter Cell demo (the newer one). It runs just fine on my 9700P using Cat 3.2. Yeah, framerate is pretty slow 30 - 50 FPS. It looks like a pretty poor port over to the PC platform to be performing this badly.

Third person Yuck.

I wont waste my $$ on this game.

 

[chsh1ca]

I also saw a review (HardOCP's Visiontek XTasy9800Pro IIRC) that used the Splinter Cell demo as a comparison, and the R9x00s performed about 5-10fps better on average, which surprised me, given Splinter Cell's history with the XBox.

 

[chizow]

Originally posted by: Pete
In terms of opimizing performance: UT2K3 has that logo, and ATi cards spank nV ones pretty handily. IMO, you're confusing driver quality with developer support in the first half of the quoted passage. They're both part of end-user experience, of course, and that's really what counts; but if we're debating the specifics of game bugs, I'm sure quite a few R300 problems are due to it being a new architecture, as you said in the second half of the quoted passage. I'm guessing ATi's current dominance will lead to many more developers taking ATi's cards into primary (rather than secondary) consideration during QA.

I'm not confusing driver quality with developer support, since they are interdependent. Game development and driver quality requires feedback and collaboration from both parties in order to fix any issues that come up during Q&A. If there's a known limitation due to hardware and the game is developed using that hardware, that will have a direct impact on game development in the form of a workaround or optimization or a driver revision. The same is true for a particular feature of a product. If that hardware is ignored during development, chances are the features and benefits of that feature won't be implemented in that game.

Going back to the nV logo in games, of course the R300 is going to outperform the GF4, but the question is by how much? In some games, the % difference is significant, in others, not so much. Look at UT2K3 and Splinter Cell for instance. In situations where the GF4 doesn't run into architectural limitations, it performs extremely well compared to an R300.

I don't think game developers will be prioritizing ATi as target hardware anytime soon; the best they can hope for is equal-footing with nVidia. Its hard to ignore 2/3 (64% from latest Mercury reports) of the AIB GPU market that nVidia enjoys, which also happens to be the overwhelming majority of the PC gaming base. There's always those that argue that ATi has 19% of the GPU market and Intel has a close second wtih 29%, but if you want to do some quick research on what the target hardware is for game developers, just look at the supported cards on the box of any game you've bought in the last 6 months. Or better yet, open up the read.me file and look at the supported chipsets. Chances are good that integrated intel GPUs will be unsupported and listed specifically as such.

nVidia's latest deal with EA just puts into writing what has been the case for some time with game developers. Driver compatibility and game support for nVidia products has been better in my experiences, as it seems ATi is always playing catch-up with their drivers.

Chiz

 

[BenSkywalker]

I also saw a review (HardOCP's Visiontek XTasy9800Pro IIRC) that used the Splinter Cell demo as a comparison, and the R9x00s performed about 5-10fps better on average, which surprised me, given Splinter Cell's history with the XBox.

If H was running both boards at the highest setting then the results are invalid. The R200 and R300 core boards lack the features to render the game in the same fashion that the NV(NV20 and up) based boards do.

ATi

nVidia

The XBox advantage for SC carried over to the PC, it is the lone game that I have seen that lives up to nV's "The way it's meant to be played" PR campaign. BTW- nV and ATi boards can be forced to render in the same fashion, although you have to lower the settings for nVidia based parts.

 

[chsh1ca]

Ben, I'm surprised you didn't bother to look at the review before opening your mouth.

Here is a link for the motivation impaired.

And IMHO, I don't care how they render, if an image in the end looks cleaner and crisper for me, that's all that matters. Whether or not it's going 'by the book' (or whatever) to achieve it doesn't really matter to me, as long as the end result is better.

 

[BenSkywalker]

Ben, I'm surprised you didn't bother to look at the review before opening your mouth.

Are you familiar with the word 'if'.... curious Reading reviews at Kyle's is akin to slamming my fingers repeatedly with a hammer. It is something I'm fully capable of doing, but see no purpose in it as it both wastes my time and is overall quite painful

Whether or not it's going 'by the book' (or whatever) to achieve it doesn't really matter to me, as long as the end result is better.

Look at the screenshots. That's a 2x4 between the eyes

 

[chsh1ca]

Originally posted by: BenSkywalker
Are you familiar with the word 'if'.... curious Reading reviews at Kyle's is akin to slamming my fingers repeatedly with a hammer. It is something I'm fully capable of doing, but see no purpose in it as it both wastes my time and is overall quite painful

Very familiar, but then, if you don't know, and aren't interested in finding out, why open your mouth (or put your fingers to use as the case may be) at all?

Look at the screenshots. That's a 2x4 between the eyes

I have, and the end result given by the Radeon 9700 Pro seems much better to me than the end result put out by the FX5800Ultra -- again, from the screenshots. Unlike all you guys, I don't make enough money to just go spend $400 on a video card just to see it in action. If you believe the overall IQ the FX5800U offers is so superior, and have tested it (it seems like you have both the FX5800Ultra and the R9800Pro to play with), then why would you bother with owning the inferior card in the R9{7|8}00 Pro series? Why not just sell it if the end result is just a two by four to the face?

 

[Rogozhin]

"The R200 and R300 core boards lack the features to render the game in the same fashion that the NV(NV20 and up) based boards do."

IN what way? The r300s with 16x quality is using trilinear 128tap AF, while the NV30 and NV20 are using 8x trilinear 64 tap AF.

And I suppose if you ramp the 5800ss AA all the way up to 16x it MIGHT look better than ati's 6x but from the apples to apples reviews it seems like nvidias 8x is closer to ati's 4x.

OPstar

I would be willing to use fraps to bench splinter cell with you just to see how we compare. We can use ICQ to talk back and forth between runs. Just let me know what level with what settings and resolution and what time you'd like to do it (I"m off today).

Let me know!

Rogo

 

[nRollo]

"touting the FX as a great card will NOT
1.get you laid
2.into heaven
3.or high"

I think I said that about the 9700, but I suppose it applies either way.

Oh no! A 5800 Ultra user who LIKES his card! It's probably only 50x better than the cards 90% of the people who slam it have too!
LOL

No matter what anyone here says, the 5800/9700/9800 are all probably fairly difficult to tell the difference between in use, unless you're going to get really specific about AA/AF settings. The 5800 wins a lot of non AA/AF, the 9X00 win a lot of AA/AF. I'd say this is definitely a preference call.

(but my head isn't so far up my a$$ I can see this morning's donuts, so I don't try to tell others what benchmark settings are relevant and which aren't)

 

[Rogozhin]

I quoted you rollo.

touting the FX as a great card will NOT
1.get you laid
2.into heaven
3.or high

Rogo

 

[Rogozhin]

The real argument should be that you're effectivly CPU limited unless you're running 3.06ghz+ and the only reason to have one of those cards on a sub 3ghz system would be for AA and AF.

Rogo

 

[gururu]

Rollo writes:

Oh no! A 5800 Ultra user who LIKES his card! It's probably only 50x better than the cards 90% of the people who slam it have too!

LOL me too.

No matter what anyone here says, the 5800/9700/9800 are all probably fairly difficult to tell the difference between in use, unless you're going to get really specific about AA/AF settings. The 5800 wins a lot of non AA/AF, the 9X00 win a lot of AA/AF. I'd say this is definitely a preference call.

YES!

It amazes me in the end how many times people feel they must re-iterate their previous posts over and over and over and over and over. I was so happy to see a fresh new opinion (OpStar). Why can't people NOT NEED to have the last post?