SHOW:
|
|
- or go back to the newest paste.
1 | 1. Make sure all GPU stored / synthesized assets (FBOs, shaders, textures) can be read-back and reconstructed, test by destroying and rebuilding the GL context. | |
2 | 2. Break out all GL/EGL code to go via per-GPU populated vtables. | |
3 | 3. Add config options to reference GL library etc. per GPU, and priority/swap list for GPUs. | |
4 | 3. Test GPU- swapping by performing 1, switch backing GPU in 2. | |
5 | 4. Add affinity-tracking to all GPU backed store (bitmap of synched GPU) | |
6 | 5. Increase daily cocktail budget. | |
7 | 6. On new client-shm or compositor side buffers, murder vram by PBO+mirroring or lose some sanity by deferring the transfer to trigger ‘on-demand’ when the same | |
8 | object backing store is required on multiple GPUs. | |
9 | 7. Shader-compilation time, do that for all GPUs. | |
10 | 8. Test-case that performs the same scene graph on both GPUs, readback and pixel-pixel comparison. | |
11 | 9. Increase daily cocktail budget. | |
12 | 10. Add protocol paths to hint/bias GPU. | |
13 | 11. Repeat the same process on client side. | |
14 | 12. Enjoy the hollow fruits of having performed the equivalent of a 500m long 2km underground caving squeeze in a space filled with spiders. That nibble at your face. |