will the R500 be pushed to the limits with 1st generation games?
- Thread starter Amil330
- Start date
Amil330 said:
It probably depends what you mean. You could probably "push it to its limits" on present day stuff just by using ultra high quality textures, rendering at 1080i, turning on triple buffering, trilinear filtering, high AA and high Aniso.
If you mean will devs be able to make it realize its potential, so far that generally seems to take a couple of generations of games.
Nite_Hawk
No one can quanitfy what the "limits" actually are.
Of one thing i'm sure though, even in next gen games there will be The Slowdown Ghost. Whether it's because of some hardware limitation or due to the programmers having a hard time it doesn't matter, since the programmers will probably tell us about the hardware limitations, while the hardware manufacturers will tell us about how the programmers aren't trying hard enough.
Of one thing i'm sure though, even in next gen games there will be The Slowdown Ghost. Whether it's because of some hardware limitation or due to the programmers having a hard time it doesn't matter, since the programmers will probably tell us about the hardware limitations, while the hardware manufacturers will tell us about how the programmers aren't trying hard enough.
It always takes 2-3 generation of games minimum before you start seeing developers really get the most out of hardware. It takes at least that long for them to really learn their way around the system.
Launch games are developed almost entirely on systems that offer only partial performance or functionality of the final hardware. They will push a system to it's limits, but usually because of poorly optimized coding due to lack of proper hardware to learn on and work with.
Launch games are developed almost entirely on systems that offer only partial performance or functionality of the final hardware. They will push a system to it's limits, but usually because of poorly optimized coding due to lack of proper hardware to learn on and work with.
New GPU architecture + 4 months of dev/debug time (chips went out in June/July) = No*
* If by "pushed" meaning utlizing/exposing the complete featureset of the chip and maximizing its potential.
Unless a game is CPU/Memory bandwidth limited (or something else) it will be GPU limited. "Limited" and "Potential" and "Maximized" are all different things. Being GPU limited means nothing. You could be limited doing something VERY simple.
Put another way, on the GC Digimon World 4 pushes Flipper as much as RE4 or Zelda does:
DW4
RE4
Zelda
So the worse looking 360 games will be "pushing" the GPU as much as its best looking games like PGR3 and GOW. Obviously some games make better use of the HW.
That said, a lot of 1st gen games wont even utilize some of the unique Xeno features--there is just not enough time. Devs have been working on 9800/X800 class hardware with SM 2.0, no 3Dc (when on the 9800pro), fixed traditional pipeline, no hardware tesselation, no MEMEXPORT, etc... The systems also did not have the Cache Lock to feed info directly from the CPU. And realistically, with July, Aug, Sept, Oct to work on their games MOST of that will be debugging. Most games are content/feature complete already. 4 months to get to a new platform and debug the game is NOT a lot of time!! And rrealistically it is more like 2-3 months due to production requirements.
And this is why most 2nd/3rd gen games are far superior to 1st gen software. The devs cut their teeth for a couple months on launch titles, start from scratch on their next project... and their LAST games on the console, after really getting into the system tend to shine because they have got to really build 1 game top to finish on the system and see where it shines and where it does not and then design a game with those in mind. Your system do water well but stink at lighting? Maybe a boat game is better than a dark creep game then.
And example of how far a game can come: Virtua Fighters on the Sega Saturn. VF1 had no textures at all. VF2 was waaaay better looking and ran better
More detail, TEXTURES, and even fingers! The devs were rushed--like all launches--and were fighting new hardware. Both games pushed the hardware, but one obviously used it better!
Games like GOW and PGR3 are always the exception at launch. Most launches have 2 or 3 good titles... rarely a killer app. Last gen only the Xbox had a killer app at launch (Halo) and the gen before that only the N64 (Mario 64... one could argue the SS, but still).
* If by "pushed" meaning utlizing/exposing the complete featureset of the chip and maximizing its potential.
Unless a game is CPU/Memory bandwidth limited (or something else) it will be GPU limited. "Limited" and "Potential" and "Maximized" are all different things. Being GPU limited means nothing. You could be limited doing something VERY simple.
Put another way, on the GC Digimon World 4 pushes Flipper as much as RE4 or Zelda does:
DW4
RE4
Zelda
So the worse looking 360 games will be "pushing" the GPU as much as its best looking games like PGR3 and GOW. Obviously some games make better use of the HW.
That said, a lot of 1st gen games wont even utilize some of the unique Xeno features--there is just not enough time. Devs have been working on 9800/X800 class hardware with SM 2.0, no 3Dc (when on the 9800pro), fixed traditional pipeline, no hardware tesselation, no MEMEXPORT, etc... The systems also did not have the Cache Lock to feed info directly from the CPU. And realistically, with July, Aug, Sept, Oct to work on their games MOST of that will be debugging. Most games are content/feature complete already. 4 months to get to a new platform and debug the game is NOT a lot of time!! And rrealistically it is more like 2-3 months due to production requirements.
And this is why most 2nd/3rd gen games are far superior to 1st gen software. The devs cut their teeth for a couple months on launch titles, start from scratch on their next project... and their LAST games on the console, after really getting into the system tend to shine because they have got to really build 1 game top to finish on the system and see where it shines and where it does not and then design a game with those in mind. Your system do water well but stink at lighting? Maybe a boat game is better than a dark creep game then.
And example of how far a game can come: Virtua Fighters on the Sega Saturn. VF1 had no textures at all. VF2 was waaaay better looking and ran better
More detail, TEXTURES, and even fingers! The devs were rushed--like all launches--and were fighting new hardware. Both games pushed the hardware, but one obviously used it better!Games like GOW and PGR3 are always the exception at launch. Most launches have 2 or 3 good titles... rarely a killer app. Last gen only the Xbox had a killer app at launch (Halo) and the gen before that only the N64 (Mario 64... one could argue the SS, but still).
Umm, can't a game be limited through software and not through the graphics chip being maxed out? Like say poor coding that causes hiccups in the code, or just turning on vsync? Certainly a Namco Museum isn't pushing a system to its limits in any way, and if vsync were turned off you'd probably get thousands of fps.
YeuEmMaiMai
Regular
it will most likely take them 6-9 months after launch to fully realise the hardware since they will have had time to learn about bugs, performance issues, etc. It will take them another year to be able to program games that have little or no slow down.
