OrdinaryMagician/vkdoom_m
OrdinaryMagician/vkdoom_m
1
Fork 0
Code Activity

- enable core profile by default on GL 3.x.

Browse source 
After doing some profiling it was very obvious that this has better performance than client arrays. Persistent buffers are still better, though, especially for handling dynamic lights.
This commit is contained in:
Christoph Oelckers 2016-08-29 11:33:20 +02:00
commit 8f535997f8
5 changed files with 14 additions and 66 deletions
Download patch file Download diff file Expand all files Collapse all files

4
src/gl/data/gl_vertexbuffer.cpp
View file

@ -148,6 +148,7 @@ FFlatVertexBuffer::FFlatVertexBuffer(int width, int height)
glBindBuffer(GL_ARRAY_BUFFER, vbo_id);
glBufferStorage(GL_ARRAY_BUFFER, bytesize, NULL, GL_MAP_WRITE_BIT | GL_MAP_PERSISTENT_BIT | GL_MAP_COHERENT_BIT);
map = (FFlatVertex*)glMapBufferRange(GL_ARRAY_BUFFER, 0, bytesize, GL_MAP_WRITE_BIT | GL_MAP_PERSISTENT_BIT | GL_MAP_COHERENT_BIT);
DPrintf(DMSG_NOTIFY, "Using persistent buffer\n");
break;
}
@ -157,13 +158,14 @@ FFlatVertexBuffer::FFlatVertexBuffer(int width, int height)
glBindBuffer(GL_ARRAY_BUFFER, vbo_id);
glBufferData(GL_ARRAY_BUFFER, bytesize, NULL, GL_STREAM_DRAW);
map = nullptr;
DPrintf(DMSG_NOTIFY, "Using deferred buffer\n");
break;
}
case BM_CLIENTARRAY:
{
// The fallback path uses immediate mode rendering and does not set up an actual vertex buffer
map = new FFlatVertex[BUFFER_SIZE];
DPrintf(DMSG_NOTIFY, "Using client array buffer\n");
break;
}
}