Nintendo DS 3d performance

yipchunyu

Newcomer
1272193-3lELEkra1O.jpg


If the linked image is real. does nintendo DS capable to display 3d graphic. The performance? above N64?
 
...

Vertex Transform Capability : Max 4 million vertices/s
Polygon Rendering Capability : Max 120K polygons/s
Pixel Rendering Capability : Max 30 million pixels/s

Must have a built in T&L engine, as it would be impossible to achieve 4 million vertices/s without one on ARM9. Polygon rasterization capability is comparable to N64, but T&L has been improved.

Is Nintendo recycling the N64 GPU macro here???
 
The "not so obvious details" -

LCD Size
256x192 RPG dots x 2

2D Graphic Engine(A,B) - (edit: whatever that means)
BG(edit: background?): max 4
OBJ(edit: objects?): max 128

3D Graphic Engine
Vertex Transform?(edit: dunno what's the english for this): 4M/sec
Polygon output: max 120K/sec
Pixel output: max 30M/sec


The 1st 11 pages of the document may contain more interesting information.
 
...

I don't know what to make of it. The GPU is comparable to what N64/PSX had, yet the T&L performance is ~4X the PSX level. 4 MB RAM is fairly restrictive, but not as bad as PSP since it feeds off cart instead of a disc.
 
Re: ...

Deadmeat said:
Vertex Transform Capability : Max 4 million vertices/s
Polygon Rendering Capability : Max 120K polygons/s
Pixel Rendering Capability : Max 30 million pixels/s

Must have a built in T&L engine, as it would be impossible to achieve 4 million vertices/s without one on ARM9. Polygon rasterization capability is comparable to N64, but T&L has been improved.

Is Nintendo recycling the N64 GPU macro here???

Deadmeat,

VFP9-S Benefits
The vector processing capability of the ARM VFP9-S offers increased performance for floating point arithmetic used in automotive powertrain and body control applications, imaging applications such as scaling, transforms and font generation in printing, 3D transforms, FFT and filtering in graphics. The next generation of consumer products such as Internet appliances, set-top boxes, and home gateways, can directly benefit from the ARM VFP9.

At 67 MHz this would be close to 87.1 MFLOPS which is not bad.

Still, at 4 MVertices/s that might call for a faster Hardwired T&L unit.
 
I don't know what to make of it. The GPU is comparable to what N64/PSX had, yet the T&L performance is ~4X the PSX level. 4 MB RAM is fairly restrictive, but not as bad as PSP since it feeds off cart instead of a disc.

Who knows this thing could be tile base, and 120k/s is poly on screen :?:

I want picture :D
 
Re: ...

Deadmeat said:
I don't know what to make of it. The GPU is comparable to what N64/PSX had, yet the T&L performance is ~4X the PSX level. 4 MB RAM is fairly restrictive, but not as bad as PSP since it feeds off cart instead of a disc.

Your cartridge point is not a bad one: I think that PSP will have some nice surprises ( more RAM IMHO ).
 
Nintendo Insider learned from a source today new, exciting developments on the Nintendo DS hardware. As confirmed earlier today, the Nintendo DS was originally called the "Nitro." Apparently, Nintendo is also referring to the Nintendo DS as "Nitro" in the "Nitro Programming Manual" and other development documents.

Our source states that the best thing about the DS is that it will cost much less than the PSP, it has "very different hardware" [than the PSP] and it "takes a totally different approach in [developing] 3D games." While it is easy to develop for -- if you develop a new title for the system, there will be no problems -- it is "very hard/impossible to convert an existing nextgen title to the DS." The hardware is not like on the PC or GameCube, but is instead really different.

Our source states that when the specs are released, you will think that the system is very low tech, but this is not true; "it really has some decent power." Our source was willing to give us this regarding the DS's power: "The best way to describe it is a 2xGBA plus 1xN64 [built] into a portable hardware, but without [those] strict ROM limits of the old consoles.." Nintendo can and will still change the hardware up to the system's release, and the games will and "MUST" be different than on other consoles or portables.

http://nintendoinsider.com/site/EpZlAZpFuueSWVPXga.php
 
Why would there be a section on a "2D Engine" if this thing has a N64-quality GPU? In that case 2D + 3D would both be handled similiarly (as they are on Playstation where "sprite" is just a special case of "triangle"). Then such things as "number of backgrounds" would be arbitrary - in fact that metric only makes sense if "Nitro" was to have an old-school scanline sprite engine instaed of a 3D GPU. That's very fishy.
 
...

Still, at 4 MVertices/s that might call for a faster Hardwired T&L unit.
Maybe not.

Since a 4x4 dot 4x1 would require 28 FLOPS, only 96 MFLOPS is needed to claim 4 million vertices transform figure.

Remembering DC, Sega could have claimed 22 million polys/s for DC but they never did due to their honesty. Well, honest people lose in this business....
 
Re: ...

Deadmeat said:
Still, at 4 MVertices/s that might call for a faster Hardwired T&L unit.
Maybe not.

Since a 4x4 dot 4x1 would require 28 FLOPS, only 96 MFLOPS is needed to claim 4 million vertices transform figure.

Darn, it is good to talk without flaming any company: kudos Deadmeat.
 
Re: ...

Deadmeat said:
Still, at 4 MVertices/s that might call for a faster Hardwired T&L unit.
Maybe not.

Since a 4x4 dot 4x1 would require 28 FLOPS, only 96 MFLOPS is needed to claim 4 million vertices transform figure.

Well, ARM claims 1.3 MFLOPS/MHz... maybe they have some special instruction that would save some FP ops for 4x4 * 4x1 operations ?

This figure is in N64 range then: the N64 had a ~100 MFLOPS FPU.
 
Why would there be a section on a "2D Engine" if this thing has a N64-quality GPU? In that case 2D + 3D would both be handled similiarly (as they are on Playstation where "sprite" is just a special case of "triangle"). Then such things as "number of backgrounds" would be arbitrary - in fact that metric only makes sense if "Nitro" was to have an old-school scanline sprite engine instaed of a 3D GPU. That's very fishy.

Its probably the GBA bit. Who knows maybe they'll allow GBA compatibility.
 
V3, it is the GBA bit.

The 4 BGs and 128 OBJs ( Sprites ) are quite explicit to peopel who have coded for the GBA.

No, they would not vary even if the CPU is clocked at 2x the frequency as they are both Hardware limited: the OAM ( Object Attribute Memory ) for example only supports 128 entries.
 
...

This figure is in N64 range then: the N64 had a ~100 MFLOPS FPU.
A N64 level 3D performance, a bit more because of improved FPU.(N64 GPU was geometry starved) This is pretty good for GBA screen resolutin, anything over that is an overkill.
 
Re: ...

Deadmeat said:
This figure is in N64 range then: the N64 had a ~100 MFLOPS FPU.
A N64 level 3D performance, a bit more because of improved FPU.(N64 GPU was geometry starved) This is pretty good for GBA screen resolutin, anything over that is an overkill.

The FPU appears few MFLOPS short of the N64 FPU, but if the memory subsystem is better those are MFLOPS you would have never gotten on the N64 anyways.

You are mentioning the RSP which was used for T&L and more by some developers.

Yet, considering the screen resolution, the fill-rate, etc... I do not think this is something we should cry home about: after all both N64 and the DS seem limited in Rasterization of said polygons ( 120 K polygons/s ).
 
...

The FPU appears few MFLOPS short of the N64 FPU
I expect the sustained number of ARM9-VFPU to be be higher than R4300i because

1. It is a newer design.
2. It has dedicated vector instructions, while R4300i relied on scalar float loops.

You are forgetting about the RSP which was used for T&L and more by some developers.
RSP never supported 4-byte floats, it was strictly for a 8-slot vector of 2-byte integers.
 
Back
Top