shader cache polishing #748
Conversation
|
@devshgraphicsprogramming did i miss anything? |
|
Did you test the hash cache works nicely on example 23 (arithmetic test) and 62 CAD ? |
Signed-off-by: Ali Cheraghi <alichraghi@proton.me>
Signed-off-by: Ali Cheraghi <alichraghi@proton.me>
Signed-off-by: Ali Cheraghi <alichraghi@proton.me>
Signed-off-by: Ali Cheraghi <alichraghi@proton.me>
Signed-off-by: Ali Cheraghi <alichraghi@proton.me>
| std::vector<unsigned char> compressedSpirv = {}; | ||
| compressedSpirv.resize(propsSize + destLen); |
There was a problem hiding this comment.
why are you making a temp vector just to memcpy it later into a spirv buffer?
just make the buffer now and write into that
There was a problem hiding this comment.
the final size after compression is unknown and ICPUBuffer can't be resized (unless you ask me to add a resize method?)
There was a problem hiding this comment.
ok didn't realize.
In that case, there's a thing called CCPUBuffer with a custom allocator that can also adopt memory
you can malloc instead of having a vector, then realloc and after realloc is complete you can create that CCustomAllocatorCPUBuffer with core::adopt_memory tag constructor overload
You just need to provide an "allocator" which will call free() on the adopted pointer
There was a problem hiding this comment.
i'll address this in the upcoming Lib Shader Stage PR
Signed-off-by: Ali Cheraghi <alichraghi@proton.me>
|
@Erfan-Ahmadi you can merge, this will probably help you a lot with the cache size |
Description
Testing
TODO list: