R420 Shows up at IDF

[zephyrprime]

Yeah it was the worst thing ever, only the second best gaming card on the planet at the time. I can see why you mock it, I imagine all of your work is always the best on the planet, and we'd all know your real name.

Yeah, it was so great that they had to PULL IT OFF THE MARKET. No doubt so they could put it in the Smithsonian where it really belonged.

 

[DerekBaker]

Seem's like we might be able to get some baseline benchmarks soon to compare these claims to, assuming, of course, that this is the same demo.

http://www.bluesnews.com/cgi-bin/board.pl?action=postmessage&boardid=1&id=0&threadid=47219


Derek

 

[bunnyfubbles]

Originally posted by: Rollo

Remember how great the fx30 was supposed to be according to nvidia?

The nV30 offered double the AA/AF performance of it's predecessor, better non AA/AF performance than it's competitor, and only slightly worse AA/AF performance than it's competitor.

Yeah it was the worst thing ever, only the second best gaming card on the planet at the time. I can see why you mock it, I imagine all of your work is always the best on the planet, and we'd all know your real name.

The point here is that someone claimed there was a bunch of hype surrounded nVida's next GPU (of which I countered with "what hype?"), trying to prove that hype needs to be taken with caution. Your "defense" of the NV30 only helped out with the hype argument, as the NV30 was majorly hyped up to take back the performance crown, some predicted that it would take it back with vengence. Even it performed slightly better in all areas it would still be a failure because of where the hype had placed it to be, that and it wouldn't be much of an improvement over what was already available for months.

 

[Insomniak]

Frankly I'm with Rollo on this one. People seem to think that anything less than best is suck. Actually, NV30 was only about 85% as good as R300 unless the AA/AF was cranked to max. Still quite capable, but no, not best.

Everything, IMO, from Nvidia and ATi since the Riva TNT and Radeon 7xxx series respectively has been good. Stuff from XGI, Intel Integrated, that's crap.

 

[nRollo]

People seem to think that anything less than best is suck

Exactly.

Rollo, you forgot to mention that the 5800U also had the best--er, biggest HSF ever

There were brands (Gainward, ABIT) with quieter HSFs, and water cooling was possible.

To this day, over a year after it's release, the 5800U is still running with the high end.
http://www.tomshardware.com/graphic/20031229/index.html

 

[BenSkywalker]

To this day, over a year after it's release, the 5800U is still running with the high end.

Yeah right. My R9800Pro is getting 1.07FPS faster performance in the most demanding game out right now(Halo). How the hell can you say that the 5800U is competitive when it is more then an entire frame per second behind today's high end parts? You are seriously screwed in the head.

 

[BugsBunny1078]

2-3 times faster my butt!
Why no benchmarks.
Also please do not link articles which have no content in them but merely refer to another article in a link.

 

[nRollo]

Yeah right. My R9800Pro is getting 1.07FPS faster performance in the most demanding game out right now(Halo). How the hell can you say that the 5800U is competitive when it is more then an entire frame per second behind today's high end parts? You are seriously screwed in the head.

LOL

When I'm wrong, I'm wrong Ben. That 1.07 fps would be worth the upgrade to me if I had one of those horrible mistake cards.

Seriously, if nothing else, the fact that year old 5800Us and year and a half old 9700Pros aren't functionally that different than today's best is illustrative of how incredibly boring this market has become. We quibble over 10% framerate differences, a few more or less jags in AA, and whether AF is straight trilinear or "brilinear" as if it makes a difference.

In the days when this stuff was "fun", new cards meant new technologies and real differences in framerate.

When it was fun to buy hardware

Back then, when you bought a next gen card, it changed you gaming life. The 9700Pro was as close as I've seen to that in a while, and all that really did was make AA/AF more usable.

 

[reever]

There were brands (Gainward, ABIT) with quieter HSFs, and water cooling was possible.

Water cooling was possible, but the card then costs one thousand dollars and offered about zero additional overclocking, the only benefit was less sound, and at a >500 dollar permium, it's not worth it

 

[Pete]

Sure, the 5800U competes with AA+AF. But those THG benches show the 5800U dropping way behind the 9700P with AA+AF, and the big deal over those two cards was the ability to run high res with AA+AF. So the reasons for choosing a 5800 were less for the hardware's technical merits than for the IHV's (namely, nV's better drivers).

I'm still puzzled by the Halo benchmark. Those of you with both a 9700+ and a 5800+, do the two cards actually play similarly with PS2.0? I mean, the 5900 has more than twice the theoretical PS2.0 performance of the 5800, yet the two score practically the same in that THG Halo benchmark. IMO, the default Halo bench doesn't seem to be a good test of PS2.0 performance considering what most other PS2.0 benchmarks show.

 

[BoomAM]

Thats cos Halo doesnt exclusivly use PS2.0 in PS2.0 mode. Its a mixed mode. PS2.0 is only used on some surfaces. PS1.1/1.4 is used for the most part.

 

[nRollo]

But those THG benches show the 5800U dropping way behind the 9700P with AA+AF, and the big deal over those two cards was the ability to run high res with AA+AF.

Wise Axl Rose once said, "Tell me who you're gonna believe?"
5800 doing much better at AA/AF here on AT
On the two playable resolutions, 10X7 and 12X10, the 9700P and the 5800U are virtually tied. Only at 16X12, where both are well below 60fps average, is there a noticeable difference.

Same thing at Splinter Cell

Jedi Knight is the same, you wouldn't notice the difference

Neck and neck at Commanche

No clear winner at Doom 3

So I guess I stand by what I said earlier: You would have been no worse off with a 5800U than a 9700Pro for the last 13 months. I said it back then, I say it still. I would guess that 99.99% of those who complained about it's noise never heard one. (myself included)
And like I said, there were models that had quieter cooling, only the reference design was loud.

DX9 PS2.0? Pfffftttt. What's that? Last I checked, I can count the games on one hand that have it, and some of them don't even show much advantage for ATI.

 

[Genx87]

IMO, the default Halo bench doesn't seem to be a good test of PS2.0 performance considering what most other PS2.0 benchmarks show.

Of course it doesnt. I mean it is after all a playable game. Only benchmarks and unreleased hype that shows Nvidia PS2 performance like 70% slower can possibly be considered "good" tests. ::rollseyes::

 

[Blastman]

If the 5800 is keeping up with the 5900 -- I ?might? be concerned about the 5800?s IQ/precision being rolled back to keep up.

Regarding DX9, the 5600 already defaults to DX8 and in the Farcry demo all NV cards defaulted to DX8. In that news item apparently Colin McRae Rally 3 is a ? ?Direct X9 heavy? ? title. I linked to some Rally 3 benches showing the ATI cards a lot faster in that game. At 12x10 4AA/8AF the 9800XT was 50% faster than the 5950.

 

[Blastman]

This Freedom Fighters game gamepc was labeled as a DX9 benchmark. 9600XT is faster than a 5900 in that game.

 

[jim1976]

:music:
Don't. Don't. Don't Believe the hype Don't Don't Don't believe the hype ....

 

[reever]

Originally posted by: Genx87
IMO, the default Halo bench doesn't seem to be a good test of PS2.0 performance considering what most other PS2.0 benchmarks show.

Of course it doesnt. I mean it is after all a playable game. Only benchmarks and unreleased hype that shows Nvidia PS2 performance like 70% slower can possibly be considered "good" tests. ::rollseyes::

And then future games come out....

 

[BFG10K]

To this day, over a year after it's release, the 5800U is still running with the high end.

Yeah, it's just a shame that 1024 x 768 isn't terribly useful at testing GPU performance.

Wise Axl Rose once said, "Tell me who you're gonna believe?"

I really wish Guns 'N Roses would get back together but that ain't gonna happen. They're one of the greatest bands ever.

 

[nRollo]

Originally posted by: BFG10K

To this day, over a year after it's release, the 5800U is still running with the high end.

Yeah, it's just a shame that 1024 x 768 isn't terribly useful at testing GPU performance.

It's fortunate that Anand used 12X10 and 16 X12 in his testing.

Wise Axl Rose once said, "Tell me who you're gonna believe?"

I really wish Guns 'N Roses would get back together but that ain't gonna happen. They're one of the greatest bands ever.

Saw them on the "Use Your Illusion" tour at a big outdoor amphitheatre. They really rocked, a great show. Personal favorites from concert: "Mr. Brownstone", "Paradise City", and "Live and Let Die".

 

[Sylvanas]

to the comments further up the page, im sure ati isnt going to sit around looking at their competitors stats on their new cards and do nothing about it! ati will own this year as they did last.

 

[Pete]

Of course it doesnt. I mean it is after all a playable game. Only benchmarks and unreleased hype that shows Nvidia PS2 performance like 70% slower can possibly be considered "good" tests. ::rollseyes::
Sure, TR:AoD and Far Cry are insignificant compared to the might of that DX8 PC port. So mighty a port that the dev has already created several patches to significantly improve performance, yet MS is holding them back. We can ignore the HL2 benches, too, because we all know that Valve is a bunch of sell-out frauds.

Of course, "synthetic" benchmarks are nothing like "real" games, even if "real" developers like Tim Sweeney recommend 3D hardware reviewers test a card's performance by creating their own "synthetic" test apps.

 

[nRollo]

Sure, TR:AoD and Far Cry are insignificant compared to the might of that DX8 PC port. So mighty a port that the dev has already created several patches to significantly improve performance, yet MS is holding them back. We can ignore the HL2 benches, too, because we all know that Valve is a bunch of sell-out frauds.

Pete, I'm surprised to see you sink to this level.
A.TR:AoD is basically a joke that few people own
B. FarCry is a demo of a game not out yet
C. And yes, we can ignore the ancient HL2 benches because we don't really have any idea when HL2 will come out or how it will perform on nVidia cards when it does?

Personally, I wouldn't put too much stock in a joke game, a demo, and some vaporware when making my hardware choices, but hey, that's just me.

 

[BenSkywalker]

IMO, the default Halo bench doesn't seem to be a good test of PS2.0 performance considering what most other PS2.0 benchmarks show.

IMO Halo is the only worth while PS 2.0 bench right now. We know FM tried to make their bench perform as slow as possible, using rendering techniques that intentionally break several optimizations that both IHVs have(rendering back to front, absurd). Halo is out, it's a great game, and it is a year after the launch of DX9.

Sure, TR:AoD and Far Cry are insignificant compared to the might of that DX8 PC port.

You realize that TRAoD is a port of a DX6(PS2) title right? Also, have you seen the game running in DX9 mode? It's pretty ugly(likely the worse aliasing I've seen in any game), not to mention performs horribly even on a R9800Pro(came with my Audigy2- no I didn't spend money on it). As far as FarCry, I would say that would make a good bench, if only it had a bench built in.

We can ignore the HL2 benches, too, because we all know that Valve is a bunch of sell-out frauds.

Well, on September 30th 2003 they are going to release the bench to the public so we can all verify their numbers so on that date we will be able to tell how honest they are

So mighty a port that the dev has already created several patches to significantly improve performance, yet MS is holding them back.

Maybe, maybe not. If they really have a patch that significantly improves performance that is ready to be released then I can't see why MS wouldn't allow it. I've been following their statements, none of them seem conrete enough in terms of their confidence of the patch.

Of course, "synthetic" benchmarks are nothing like "real" games, even if "real" developers like Tim Sweeney recommend 3D hardware reviewers test a card's performance by creating their own "synthetic" test apps.

Or we could quote Carmack saying that first generations of new features are for developers, not consumers. With all the BS sites crammed down everyone's throat about how fast DX9 was going to be adopted and how many titles we would be seeing it is currently comical to have the only two games out that use the features in any real way at all be ports from consoles of a DX8 and a DX6 title.

BFG

Yeah, it's just a shame that 1024 x 768 isn't terribly useful at testing GPU performance.

You very clearly don't own Halo

 

[ronnn]

Originally posted by: Rollo

Pete, I'm surprised to see you sink to this level.
______________________________________________________________________________________________

Now that the normal flame baiting statements are being made....... about hardware that is almost extinct (in terms of new sales). Having read the link at the top and the forum stuff. It seems that some people think that the new hardware may actually be twice as fast. Find it hard to believe myself. If it is true, I will wait a couple of months for the kinks to be worked out and break down and buy one.

 

[nRollo]

Now that the normal flame baiting statements are being made....... about hardware that is almost extinct (in terms of new sales).

I'm not trying to flame Pete. He's usually the "voice of reason" in these debates, so I was surprised to see him using that POS TRAoD as evidence of anything other than bad programming, and two games that don't exist yet.

I've said it before and I'll say it again: DX9 PS2 is about as important for last spring's cards as hardware T/L was for the original GF. Maybe even less.

It may well be a factor some day, but by the time it is, last spring's cards aren't the ones you'll want to run it. Anyone who bought ATI for it's PS2 capability last year will have to buy again this year.