MirrorBackup/libstarlight
1
0
Fork 0
forked from Mirror/libstarlight
Code Pull Requests Activity

let's not waste precious VRAM on a depth buffer we don't use

Browse Source
This commit is contained in:
zetaPRIME 2017-03-18 16:13:12 -04:00
parent c17200e8db
commit ddb272fd66
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
libstarlight/source/starlight/gfx/RenderCore.cpp
View File

@ -51,14 +51,14 @@ namespace { // internals
return b;
}
inline int NextPow2(unsigned int x) {
inline unsigned int NextPow2(unsigned int x) {
--x;
x |= x >> 1;
x |= x >> 2;
x |= x >> 4;
x |= x >> 8;
x |= x >> 16;
return ++x >= 64 ? x : 64; // min size to keep gpu from locking
return std::min(std::max(++x, 64U), 1024U); // clamp size to keep gpu from locking
}
class CRawTexture : public CTexture {
@ -248,7 +248,7 @@ CRenderTarget::CRenderTarget(int width, int height, bool forceExact) {
auto w = forceExact ? width : NextPow2(width),
h = forceExact ? height : NextPow2(height);
txSize = Vector2(w, h);
tgt = C3D_RenderTargetCreate(w, h, GPU_RB_RGBA8, GPU_RB_DEPTH24_STENCIL8); // though actually, do we really need stenciling? could drop to 16 if we don't
tgt = C3D_RenderTargetCreate(w, h, GPU_RB_RGBA8, -1/*GPU_RB_DEPTH24_STENCIL8/**/); // I don't think we need a depth buffer >.>
Mtx_Ortho(&projection, 0.0f, w, 0.0f, h, 0.0f, 1.0f, true);
//Mtx_OrthoTilt(&projection, 0.0f, h, 0.0f, w, 0.0f, 1.0f, true);
}