480 Commits

Author	SHA1	Message	Date
firestar99	62662c92c7	explicit layout: WIP7, build equivalent test, showing bad shader optimizations	2025-12-10 12:11:43 +01:00
firestar99	449bbf485c	explicit layout: WIP6, fix offset calculation in test	2025-12-10 12:11:07 +01:00
firestar99	91486163ef	explicit layout: WIP5, move compiletest to separate folder	2025-12-10 12:10:41 +01:00
firestar99	9bd442ed9c	explicit layout: WIP4, dedup slice copy fns	2025-12-10 11:51:21 +01:00
firestar99	e4941b9727	explicit layout: WIP3, compiletest vector load	2025-12-10 11:51:21 +01:00
firestar99	9ae3114a88	explicit layout: WIP2, compiles	2025-12-10 11:51:21 +01:00
firestar99	8ded5cabf8	explicit layout: WIP1, array to slice cast fails	2025-12-10 11:51:21 +01:00
firestar99	ae25e1a47b	image texel write: change texel write arg from any vector to SampleResult, allows scalars ... * breakage: no longer allows any vector, type must match image format * allows texel writes into single-component scalar textures	2025-12-08 07:33:28 +00:00
firestar99	4a75a05549	image texel write: compiletests with explicit formats, fails for scalar formats	2025-12-08 07:33:28 +00:00
firestar99	ecb41c22dc	dylib: remove crate-type declarations from examples and difftests	2025-12-05 23:29:56 +00:00
firestar99	fe941fec1e	compiletest: ignore arch/shared on vulkan1.0/spv1.2 and lower	2025-12-05 10:35:47 +00:00
firestar99	18f9e0266d	replace tabs with spaces	2025-12-05 10:35:47 +00:00
firestar99	77708a3d3e	ScalarComposite: rename ScalarOrVectorComposite to ScalarComposite, bulk rename	2025-11-28 15:16:26 +00:00
firestar99	7c73d49942	ScalarComposite: derive enums and improve enum docs	2025-11-28 15:16:26 +00:00
firestar99	e9b5a74000	ScalarComposite: adjust subgroup_all_equal to accept composites	2025-11-28 15:16:26 +00:00
firestar99	589d51ff19	ScalarComposite: add #[derive(ScalarOrVectorComposite)] for structs	2025-11-28 15:16:26 +00:00
firestar99	403dd84ea4	ScalarComposite: add ScalarOrVectorComposite for subgroup intrinsics	2025-11-28 15:16:26 +00:00
firestar99	a49f4233fd	shared: only retain entry point vars with preserve-bindings	2025-11-27 09:43:14 +00:00
firestar99	02806ee961	shared: fix unused shared memory failing validation	2025-11-27 09:43:14 +00:00
firestar99	b01acc6840	shared: compiletest for unused shared memory, failing	2025-11-27 09:43:14 +00:00
firestar99	a496aeabd8	Matrix4x3: verify abi layout in difftest	2025-11-27 09:39:55 +00:00
Firestar99	703488181f	difftest: force wgpu to use Vulkan, disable dx12 ... naga fails to transpile our spirv to hlsl	2025-11-27 00:12:30 +00:00
firestar99	b3eda4df98	fix compiletest for spv1.4 and up	2025-11-12 23:35:58 +00:00
Christian Legnitto	2232fcb7b5	Do not ICE on indirect fn pointers ... Fixes https://github.com/Rust-GPU/rust-gpu/issues/452.	2025-10-28 08:49:53 +00:00
firestar99	67ccf2f6aa	format args decompiler: move support functions in spirv-std to mod debug_printf ... Currently, every time `lib.rs` changes, there's a change in the debug printf compiletest stderr, which is causing a lot of churn.	2025-10-22 13:42:03 +00:00
firestar99	8c3fa7b7e8	subgroup: add subgroup_broadcast_const variant for pre-spv1.5 broadcasts	2025-10-22 13:31:48 +00:00
firestar99	f438c8c817	scalar or vector: rename VectorOrScalar to ScalarOrVector and give it its own mod	2025-10-22 12:30:23 +00:00
firestar99	08a245cb23	scalar or vector: move all scalar traits and impl macros to mod scalar	2025-10-22 12:30:23 +00:00
firestar99	0ada17d0ec	scalar or vector: reexport all Scalar traits in lib.rs, make mods private ... apart from `mod float`, which contains functionality	2025-10-22 12:30:23 +00:00
firestar99	a28bd3a333	SpecConstant: add arrayed spec constants	2025-10-22 08:37:42 +00:00
firestar99	300da9c21d	compiletest: pass -Zui-testing to rustc	2025-10-22 03:55:58 +00:00
firestar99	50019a74a8	compiletest: normalize subgroup tests	2025-10-22 03:55:58 +00:00
firestar99	e7971d3695	compiletest: update readme normalize examples	2025-10-15 09:42:01 +00:00
firestar99	adf6ee9523	abi layout: add some compiletests	2025-10-15 09:42:01 +00:00
firestar99	7e02135fb8	abi layout: limit vectors to at most 4 components, as spec states	2025-10-15 09:42:01 +00:00
firestar99	9218e64b59	abi layout: minor code cleanups	2025-10-15 09:42:01 +00:00
firestar99	95837d7999	abi layout difftest: cuda and scalar-math feature testing	2025-10-15 09:42:01 +00:00
firestar99	6ed642b5bb	abi layout difftest: cuda and scalar-math feature forwarding	2025-10-15 09:42:01 +00:00
firestar99	45147a21d5	abi layout difftest: add all remaining glam types	2025-10-15 09:42:01 +00:00
firestar99	a5442acd31	abi layout compiletest: bless complex_image_sample_inst	2025-10-15 09:42:01 +00:00
firestar99	9b2860639f	abi layout: change Subgroup from transparent struct to typedef	2025-10-15 09:42:01 +00:00
firestar99	f30056a1c1	abi layout compiletest: fix invalid-matrix-type	2025-10-15 09:42:01 +00:00
firestar99	e7f4f3b81b	abi layout: glam::BVec support	2025-10-15 09:42:01 +00:00
firestar99	cb8df101b4	update glam: switch from #[repr(SIMD)] to #[rust_gpu::vector::v1]	2025-10-15 09:42:01 +00:00
firestar99	c83ef32ab7	abi layout difftest: add member offset checking	2025-10-15 09:42:01 +00:00
firestar99	5037751f8a	abi layout difftest: add difftest for layout size and alignment	2025-10-15 09:42:01 +00:00
firestar99	a26e0df73a	compiletest: shared memory reductions, using the same type in buffers and as shared memory	2025-10-11 10:58:29 +00:00
Eduard-Mihai Burtescu	8c8ff2fe1f	Erase explicit layout decorations (Offset/ArrayStride) when disallowed by Vulkan.	2025-10-11 10:58:29 +00:00
firestar99	7a72a2fa40	byte_addr_buffer: compiletests with mixed scalar and scalar pairs, store is failing	2025-10-11 09:50:40 +00:00
Christian Legnitto	56cc132145	Fix divide by zero ... We assumed every array/vector/matrix element has a non-zero stride, but `[()]` inside `BitSlice` produced a zero stride which is divide by zero. Fixes https://github.com/Rust-GPU/rust-gpu/issues/424.	2025-10-09 08:49:19 +00:00