kuba c13e09130c [ShaderCompiler] Resolve V_MOVREL* with selects instead of a spilled VGPR array ... A program with any V_MOVREL* instruction declared its whole VGPR file as a 512-entry private array indexed at run time, which Metal spills to thread memory, so every register access became a memory access. Keep one variable per register and resolve relative accesses with a select over the registers the program names. Octopath's instanced vertex shader (1536 x 8124): 112 ms -> 6.4 ms per draw in replay, output byte-identical.		2026-09-29 22:47:40 +02:00
..
SharpEmu.Libs.Tests	[ShaderCompiler] Resolve V_MOVREL* with selects instead of a spilled VGPR array	2026-09-29 22:47:40 +02:00
SharpEmu.ShaderCompiler.Metal.Tests	Amd fix opcode nid filecache wave32 (#903)	2026-09-24 15:00:42 +03:00
SharpEmu.ShaderCompiler.Tests	[SpirvTranslator] Give each used register its own private variable	2026-09-29 22:47:40 +02:00
SharpEmu.SourceGenerators.Tests	feat(hle): prefer loaded guest exports for selected registrations (#844)	2026-08-25 22:32:57 +02:00