It seems NVidia also have low performance in Max Payne 2

[McArra]

Here is a review using Max Payne 2 to compare 5700U and 9600XT and the difference in this game is quite big. Damn, I thought NVidia really had improved with 5700U, but it seems new drivers are needed to improve performance for games, with optimization for each one.

 

[Corporate Thug]

killed in shadermark 2.0 as well

 

[Rage187]

my 5900nu runs it like a dream, w/ 8 af, and 4aa.


w00t w00t

 

[kylebisme]

my card runs some of the levels in max payne 2 like a dream...


... oh wait, some levels are dreams.

 

[Nebor]

Just because it doesn't run it as fast as the ATI cards doesn't mean it runs it slow. My card runs the game perfectly @ 1280x1024 4xAA 8xAF All High Details.

 

[kylebisme]

sure enough, but i think Brent at hardocp said it well when he stated:

The game was very playable at 1280x1024 4XAA/8XAF on both the GFFX 5950Ultra and the Radeon 9800XT; however the 9800XT gives you even more performance to play with if you wish to increase to 1600x1200 or perhaps 6XAA or 16XAF at 1280x1024.

http://www.hardocp.com/article.html?art=NTM3LDM=

 

[DannyBoy]

My 9100 plays it sweet on 1280x1024 w 4xaa and 4xaf

 

[Tab]

I have to wait another we until the 9600XTs are out Then it is Max Payne 2

 

[MercenaryForHire]

My Quadro4 played it just fine.

- M4H

 

[Pete]

Max Payne 2 numbers may need to be taken with a grain of salt. Apparently (I learned this at B3D), the devs kept their code very close to their vest, not showing it to anyone--so nV and ATi may not have had a chance to optimize for it. I've also read that the devs themselves said MP2 wasn't suited for benchmarking because of all the behind-the-scenes dynamic tweaking the game does, as it can't offer equal settings for an even comparison.

After the performance boosts nV gave Halo and TR:AoD and AM3 with new drivers, it's worth waiting to see what they can do with MP2. (I'm still waiting on the promised IQ comparisons to verify that their recent gains are legit, though.)

 

[reever]

Originally posted by: Pete
After the performance boosts nV gave Halo and TR:AoD and AM3 with new drivers, it's worth waiting to see what they can do with MP2. (I'm still waiting on the promised IQ comparisons to verify that their recent gains are legit, though.)

AM3 fine, but TRAOD? It gained fps, but every site that saw the fps gain was using the game WITHOUT any ps2.0 options on

 

[kylebisme]

Originally posted by: Pete
Max Payne 2 numbers may need to be taken with a grain of salt. Apparently (I learned this at B3D), the devs kept their code very close to their vest, not showing it to anyone--so nV and ATi may not have had a chance to optimize for it. I've also read that the devs themselves said MP2 wasn't suited for benchmarking because of all the behind-the-scenes dynamic tweaking the game does, as it can't offer equal settings for an even comparison.

After the performance boosts nV gave Halo and TR:AoD and AM3 with new drivers, it's worth waiting to see what they can do with MP2. (I'm still waiting on the promised IQ comparisons to verify that their recent gains are legit, though.)

i don't know it seems to me that image quality is what matters most and it needs to come at an acceptable framerate, regardless of what "behind-the-scenes dynamic tweaking" the game does. also, most people who wanted to play the game have already beat it by now so waiting to see what they can do comes off as too little to late imho.

 

[Pete]

Sorry, reever, I was playing fast and loose with the numbers. I thought nV gained on TR:AoD, but I don't exactly pay attn to that bench, so I was just assuming it saw a big boost.

Snowman, that's a good point that most people already bought and played MP2, so its current performance is important. That's a great point, in fact, and one I'll keep in mind. Future optimizations are nice on MP titles, but SP titles really need to perform well out of the gate.

 

[stardust]

Please pay attention to the Van's lights rite above Max payne's head. Notice that the ATI drivers were less rendered? (those lights are very close to the center of the screen so it shouldn't be affected by ATI's flower like AA filtering.

ATI drivers

Nvidia drivers

Very interesting indeed.

 

[Crapgame]

Wow interesting is an understatement, look at the liesence plate on the car to the left, ATI's screen the last digit and part of the plate is crunched away, on the next car over look at the grill detail, half of it is gone completely on the ATI screen. Call me an nVidiot if you will but nV is the clear winner as far as appearance goes in that comparison IMHO.

 

[InlineFive]

It's like choosing the lesser of two evils.

 

[McArra]

True. Why can that be, a bug? I hope so :disgust:

 

[Pete]

No offense, stardust, but that's not very interesting. I mean, that's the most minute IQ difference I've seen in the many, many comparative screenshots between the two architectures. A little higher AF on a couple of textures as small as those headlights and the pipes aren't going to affect benchmarks by more than a few percentage points, at worst; I'm guessing closer to 0% than 1%.

 

[stardust]

Originally posted by: Pete
No offense, stardust, but that's not very interesting. I mean, that's the most minute IQ difference I've seen in the many, many comparative screenshots between the two architectures. A little higher AF on a couple of textures as small as those headlights and the pipes aren't going to affect benchmarks by more than a few percentage points, at worst; I'm guessing closer to 0% than 1%.

OK pete, FYI only, FYI only, sorry

edit: everyone's getting so tense just b4 the end of the year

 

[Schadenfroh]

my SGI Reality Engine workstation runs it really good

 

[Pete]

Sorry, stardust, my response obviously came across as too harsh. I did say don't take offense, though. I'm just used to much larger IQ differences, and the ones you pointed out are pretty minute. My apologies if you thought I was too harsh or dismissive.

 

[jiffylube1024]

Originally posted by: McArra
Here is a review using Max Payne 2 to compare 5700U and 9600XT and the difference in this game is quite big. Damn, I thought NVidia really had improved with 5700U, but it seems new drivers are needed to improve performance for games, with optimization for each one.

2X AF, ugh. Why even bother. All it does is make the ATI go into biliniar. May as well leave it off! As for 2X AA, it's pretty useless, but not as bad as 2X AF!

Still, the 9600XT doess pull away from the 5700U in MP2, impressively so actually.

 

Originally posted by: TheSnowman

Originally posted by: Pete
Max Payne 2 numbers may need to be taken with a grain of salt. Apparently (I learned this at B3D), the devs kept their code very close to their vest, not showing it to anyone--so nV and ATi may not have had a chance to optimize for it. I've also read that the devs themselves said MP2 wasn't suited for benchmarking because of all the behind-the-scenes dynamic tweaking the game does, as it can't offer equal settings for an even comparison.

After the performance boosts nV gave Halo and TR:AoD and AM3 with new drivers, it's worth waiting to see what they can do with MP2. (I'm still waiting on the promised IQ comparisons to verify that their recent gains are legit, though.)

i don't know it seems to me that image quality is what matters most and it needs to come at an acceptable framerate, regardless of what "behind-the-scenes dynamic tweaking" the game does. also, most people who wanted to play the game have already beat it by now so waiting to see what they can do comes off as too little to late imho.

So if you play MP2, but an ATI card. If you play Neverwinter Nights, but an nvidia card. Simple. Buy the card that plays YOUR game the fastest and with best IQ for yourself.

 

Originally posted by: Pete
No offense, stardust, but that's not very interesting. I mean, that's the most minute IQ difference I've seen in the many, many comparative screenshots between the two architectures. A little higher AF on a couple of textures as small as those headlights and the pipes aren't going to affect benchmarks by more than a few percentage points, at worst; I'm guessing closer to 0% than 1%.

It is minute, but just from one screen shot to the other. But it's not so minute when the cards render billions of pixels and shaders per second. Every little bit counts when your dealing with collosal numbers of operations per millisecond.

 

[stardust]

Originally posted by: gorillaman
It is minute, but just from one screen shot to the other. But it's not so minute when the cards render billions of pixels and shaders per second. Every little bit counts when your dealing with collosal numbers of operations per millisecond.

That reminds me of Tomb Raider where there is a circle effect around the character in 3rd person games. In TR terms, it was a bug in the game (nv architecture compatibility problem) that forced the Geforce cards to over calculate the PS2 and create this blurry effect like a round halo above the main character's head. That may be one of the reasons why the FX5900 is so slow in that benchmark, but fixed in the FX5950 with det50s.


link