Implement shader storage buffer operations using new Load/Store instructions (#4993)

* Implement storage buffer operations using new Load/Store instruction

* Extend GenerateMultiTargetStorageOp to also match access with constant offset, and log and comments

* Remove now unused code

* Catch more complex cases of global memory usage

* Shader cache version bump

* Extend global access elimination to work with more shared memory cases

* Change alignment requirement from 16 bytes to 8 bytes, handle cases where we need more than 16 storage buffers

* Tweak preferencing to catch more cases

* Enable CB0 elimination even when host storage buffer alignment is > 16 (for Intel)

* Fix storage buffer bindings

* Simplify some code

* Shader cache version bump

* Fix typo

* Extend global memory elimination to handle shared memory with multiple possible offsets and local memory
This commit is contained in:
gdkchan 2023-06-03 20:12:18 -03:00 committed by GitHub
parent 81c9052847
commit 21c9ac6240
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
42 changed files with 1468 additions and 1259 deletions

View file

@ -79,10 +79,8 @@ namespace Ryujinx.Graphics.Shader.IntermediateRepresentation
ImageAtomic,
IsNan,
Load,
LoadGlobal,
LoadLocal,
LoadShared,
LoadStorage,
Lod,
LogarithmB2,
LogicalAnd,
@ -117,16 +115,10 @@ namespace Ryujinx.Graphics.Shader.IntermediateRepresentation
Sine,
SquareRoot,
Store,
StoreGlobal,
StoreGlobal16,
StoreGlobal8,
StoreLocal,
StoreShared,
StoreShared16,
StoreShared8,
StoreStorage,
StoreStorage16,
StoreStorage8,
Subtract,
SwizzleAdd,
TextureSample,

View file

@ -11,7 +11,12 @@ namespace Ryujinx.Graphics.Shader.IntermediateRepresentation
StorageBuffer,
LocalMemory,
SharedMemory,
GlobalMemory
GlobalMemory,
// TODO: Remove those and store type as a field on the Operation class itself.
GlobalMemoryS8,
GlobalMemoryS16,
GlobalMemoryU8,
GlobalMemoryU16
}
static class StorageKindExtensions