Why did the Gamecube fail?

[Pemalite]

Leynos said:

curl-6 said:

Regarding PS2 and Gamecube, as I understand it the Cube's main graphical advantages were its high speed 1T-SRAM, and it's TEV unit which allows 8 textures to be combined in a single pass, thus giving it an advantage in things like bumpmapping and the like, correct?

I think Xbox only had 4. At least according to EGM in 2001.

The Xbox had 8 texture mapping units. It's Geforce 3/4 derived. It can match the Gamecube on this front.

Though the OG Xbox can handle larger textures due to it's higher system bandwidth and system memory capacity... The Cube had the texture cache and 1T-SRAM to make up some of that difference and can simply sample more textures and layer them appropriately in real time.
Star Wars definitely showcases this advantage for the Gamecube to a great extent.

curl-6 said:

Leynos said:

I think Xbox only had 4. At least according to EGM in 2001.

I remember reading that too. As I understand it (I could be wrong) Gamecube was faster at multitexturing due to being geared for it, while Xbox's programmable pixel and vertex shaders were less fast but more flexible.

Definitely. The Gamecube's TEV meant it was a texturing powerhouse, no doubt... And the games showcased that.
But for compute it was a no contest, this was simply one of nVidia's big key strengths at the time, no one would approach nVidia on this front (And we aren't talking raw specs here, but real-world capability) until ATI brought out the Radeon 9700 years later once the 7th generation was underway.

The Xbox being built on PC technology had a more 7th-gen hardware design which emphasized shaders and compute to pull off effects like deferred rendering in Shrek (Something that hit it big with Battlefield 3 during the 7th gen)...

And Ray Tracing in Conker: Live and Reloaded. (Global Illumination.)
In Conker if the normal looks towards the ground, the engine would shoot a single ray from the sky towards the ground and do collision detection with colour encoding calculations performed on the vertices.
Result was the ability for the engine to seamlessly do colour shifting between scenes.

Obviously Conkers implementation of Ray Tracing is rudimentary as it's only running on 6th gen hardware, but hard to ignore the benefits it brought forward.

Since then Ray Tracing has only become more intricate, demanding and impressive.
Conker and Shrek though are probably two of my favorite games that showcases the foundations of the main technologies we rely on today and which define current games and rendering approaches.

It was really forward looking hardware.

Last edited by Pemalite - on 13 October 2020