Confirmed users
753
edits
Platform/GFX/DeferredGPUs: Difference between revisions
→Draw-calls are expensive
| Line 96: | Line 96: | ||
== Draw-calls are expensive == | == Draw-calls are expensive == | ||
They are always expensive regardless of the type of GPU, but talking with ARM people it sounded like there was something about deferred that made it worse. The ARM document explains: '' "The processing | They are always expensive regardless of the type of GPU, but talking with ARM people it sounded like there was something about deferred that made it worse. The ARM document explains: ''"The processing required by draw calls include allocating memory, copying data, and processing data. The overhead is the same whether you draw a single triangle or thousands of triangles in a draw call."'' | ||
required by draw calls include allocating memory, copying data, and processing data. The overhead is the same whether you draw a single triangle or thousands of triangles in a draw call." '' | |||
The consensus seems to be that it is worth using texture atlas to minimize draw calls, even if (in the case of a browser) that means dynamically updating the atlas with glTexSubImage2D. Well, at least on drivers where glTexSubImage2D is not broken. | The consensus seems to be that it is worth using texture atlas to minimize draw calls, even if (in the case of a browser) that means dynamically updating the atlas with glTexSubImage2D. Well, at least on drivers where glTexSubImage2D is not broken. | ||