Hey all,

I'm going to buy a Macbook Pro next month, and I was wondering if the X1600 is a good enough card to use for shaders. I'm upgrading from my 12" Powerbook because its Go5200 is just not powerful enough for shader experiments (my current game uses a dead simple fragment shader and loses 30-40 FPS), and also to get shiny new hardware that can run Windows on the side

Anyone been using the X1600 for GLSL? I assume it should be much better than the R9700M in my friend's Powerbook 15", which would satisfy my needs for awhile.

Quote:I'm going to buy a Macbook Pro next month, and I was wondering if the X1600 is a good enough card to use for shaders.

Yes. The design itself is fairly flexible (being SM3 compliant) as well.

Quote:I assume it should be much better than the R9700M in my friend's Powerbook 15", which would satisfy my needs for awhile.

Yes the R5xx series (which the X1600 is a member of) is much better wrt programmable vertex / fragment pipeline performance than the R3xx you are using now is... As per usual the only thing you loose out on (in general versus a higher-end desktop GPU) tends to be fill rate / memory bandwidth.

That's good (great, actually. I had heard ATI lags in Shader Model 3 support -- how well do the OS X drivers support it?). My desktop PC has an NV6600GT; how does that stack up (if at all)?

I'm still buying the machine, but I'm trying to figure out where the machine sits in the "graphics power continuum" for lack of a better term.

Thanks so far.

ATI's drivers on Mac OS X are *much* better than NVidia's. The X1600 fully supports dynamic branching and looping within GLSL on Mac OS X. I don't think the GeForce 6 & 7 family yet do. The X1k family of cards don't support vertex texturing at all (but the GeForce 6 & 7 family's capability isn't exposed through GLSL on Mac OS X either). Render-to-vertex-buffer works.

OneSadCookie Wrote:The X1600 fully supports dynamic branching and looping within GLSL on Mac OS X. I don't think the GeForce 6 & 7 family yet do.

Dynamic branching is supported on Geforce 6 and 7. (Note that as of 10.4.3, NV_fragment_program2 is exported, which adds the necessary IF, ELSE, REP etc support.)

I know it's supported by ARB_fragment_program via NV_fragment_program2, but I didn't think it worked in GLSL... and I thought the same was true of vertex texturing. Am I wrong?

OneSadCookie Wrote:I know it's supported by ARB_fragment_program via NV_fragment_program2, but I didn't think it worked in GLSL... and I thought the same was true of vertex texturing. Am I wrong?

GLSL leverages the NV_ extensions.
Vertex texturing is not supported yet (MAX_VERTEX_TEXTURE_IMAGE_UNITS_ARB = 0)

Okay. I take it the GMA 950 in the Macbook is not much better than my Go5200, if at all?

It may be better for pixel shaders, but the lack of hardware T&L and vertex shaders will seriously hinder it.

There's no hardware transform & lighting? Cripes.

MBP for sure then, even though it's a couple extra inches of aluminum to carry around all day.

Thanks, guys.