rust - https://github.com/rust-lang/rust

Age	Commit message (Collapse)	Author	Lines
2025-07-22	Rename `tests/codegen` into `tests/codegen-llvm`	Guillaume Gomez	-1730/+0

2025-07-20	So many test updates x_x	Scott McMurray	-473/+102

2025-07-03	Block SIMD in transmute_immediate; delete `OperandValueKind`	Scott McMurray	-4/+2
	See conversation in <https://rust-lang.zulipchat.com/#narrow/channel/131828-t-compiler/topic/Is.20transmuting.20a.20.60T.60.20to.20.60Tx1.60.20.28one-element.20SIMD.20vector.29.20UB.3F/near/526262799>.
2025-06-04	x86 (32/64): go back to passing SIMD vectors by-ptr	Ralf Jung	-15/+5

2025-04-20	simd intrinsics with mask: accept unsigned integer masks	Ralf Jung	-0/+57

2025-02-27	remove most `simd_` intrinsic declaration in tests	Folkert de Vries	-193/+131
	instead, we can just import the intrinsics from core
2025-02-25	remove `simd_fpow` and `simd_fpowi`	Folkert de Vries	-174/+0

2025-02-18	x86-sse2 ABI: use SSE registers for floats and SIMD	Ralf Jung	-2/+15

2025-01-27	Fix SIMD codegen tests on LLVM 20	Nikita Popov	-14/+14
	The splat contents are printed differently on LLVM 20.
2025-01-26	Consistently use the most significant bit of vector masks	Jörn Horstmann	-13/+110
	This improves the codegen for vector `select`, `gather`, `scatter` and boolean reduction intrinsics and fixes rust-lang/portable-simd#316. The current behavior of most mask operations during llvm codegen is to truncate the mask vector to <N x i1>, telling llvm to use the least significat bit. The exception is the `simd_bitmask` intrinsics, which already used the most signifiant bit. Since sse/avx instructions are defined to use the most significant bit, truncating means that llvm has to insert a left shift to move the bit into the most significant position, before the mask can actually be used. Similarly on aarch64, mask operations like blend work bit by bit, repeating the least significant bit across the whole lane involves shifting it into the sign position and then comparing against zero. By shifting before truncating to <N x i1>, we tell llvm that we only consider the most significant bit, removing the need for additional shift instructions in the assembly.
2024-11-07	Update test for LLVM 20's new vector splat syntax	Hans Wennborg	-3/+3
	that was introduced in https://github.com/llvm/llvm-project/pull/112548
2024-09-09	Ban non-array SIMD	Scott McMurray	-309/+162

2024-04-24	Fix tests and bless	Gary Guo	-1/+0

2024-02-25	fix use of platform_intrinsics in tests	Ralf Jung	-45/+45

2024-02-22	[AUTO_GENERATED] Migrate compiletest to use `ui_test`-style `//@` directives	许杰友 Jieyou Xu (Joe)	-23/+23

2024-02-20	delete a test that no longer makes sense	Ralf Jung	-47/+0

2023-12-12	Fix alignment passed down to LLVM for simd_masked_load	Jakub Okoński	-2/+2

2023-12-09	Add simd_masked_{load,store} platform-intrinsics	Jakub Okoński	-0/+66
	This maps to the LLVM intrinsics: llvm.masked.load and llvm.masked.store
2023-07-29	cg_llvm: simplify llvm.masked.gather/scatter naming with opaque pointers	Erik Desjardins	-4/+4
	With opaque pointers, there's no longer a need to generate a chain of pointer types in the intrinsic name when arguments are pointers to pointers.
2023-07-27	CHECK only for opaque ptr	Josh Stone	-10/+10

2023-07-08	Always name the return place.	Camille GILLOT	-89/+88

2023-04-04	Allow `transmute`s to produce `OperandValue`s instead of always using `alloca`s	Scott McMurray	-4/+47
	LLVM can usually optimize these away, but especially for things like transmutes of newtypes it's silly to generate the `alloc`+`store`+`load` at all when it's actually a nop at LLVM level.
2023-02-14	Try to fix codegen tests for ??? LLVM 14 ???	Ben Kimock	-5/+5

2023-02-14	Fix codegen tests	Ben Kimock	-92/+92

2023-02-12	Enable CopyProp by default, tune the impl a bit	Ben Kimock	-3/+0

2023-01-11	Move /src/test to /tests	Albert Larsan	-0/+2272