Q: what's faster in VRAM Same tpage?

[gwald]

1: A large single 120x 24 GsSprite, which is mostly blank (transparent)
2: Or 2 smaller 24x24 GsSprites

I know in openGL, #2 is faster, but I know LibGS has render call overheads.
Anyone know?

[Administrator]

Well if you have one large image, the PSX doesn't have to reference so many positions in the framebuffer, so in theory, that would be the quickest. Usually less instructions means the system will execute quicker. Don't think you'll know for sure though unless you compare with a DTL-H2700.

[Orion_]

try to bench it.
display 100 sprites with method 1 then using method 2, and measure it using the value returned by Vsync.

GsSortSprite_WithMethod1();
cpu_time = VSync(1);
DrawSync(0);
gpu_time = VSync(0);
GsSwapDispBuff();
...

[gwald]

Cool, thanks both for the replies