v0.27.3

GPUCompiler v0.27.3

Diff since v0.27.2

Merged pull requests:

• Fix Report exception issue on nested codegen (#619) (@wsmoses)

v0.27.2

GPUCompiler v0.27.2

Diff since v0.27.1

Merged pull requests:

• Metal: Replace unreachable control flow with exit block branches (on macOS <15) (#618) (@maleadt)

v0.27.1

GPUCompiler v0.27.1

Diff since v0.27.0

Merged pull requests:

• Add CI jobs for AMDGPU (#612) (@pxl-th)
• Fix use of Base.rename for 1.12 (#613) (@nhz2)
• fix mangling of booleans and zeros (#614) (@simonbyrne)
• Bump LLVM compat. (#617) (@maleadt)

Closed issues:

• Mangler does not generate C++ valid names with boolean-valued parameters (#609)
• Julia -> C function (Create pthead) -> Julia CUDA kernel issue (#615)
• Deprecations thrown on current release? (#616)

v0.27.0

GPUCompiler v0.27.0

Diff since v0.26.7

Merged pull requests:

• Bump minimal Julia requirement to v1.10. (#598) (@maleadt)
• Set codegen kwargs based on toplevel setting (NFC) (#600) (@maleadt)
• Switch to PrecompileTools.jl. (#601) (@maleadt)
• [Metal] Fix llvm.minimum (#602) (@tgymnich)
• Be more strict about runtime function misuse. (#603) (@maleadt)
• Default to using NewPM. (#604) (@maleadt)
• Remove legacy pass manager pipelines. (#605) (@maleadt)
• [CI] Move to julia-actions/cache (#606) (@giordano)
• Metal: Fix IR validation. (#608) (@maleadt)

Closed issues:

• GPUCompiler emit_exception has wrong number o args (#596)
• Error when precompiling with Julia 1.11.0-rc2 (#610)

v0.26.7

GPUCompiler v0.26.7

Diff since v0.26.6

Merged pull requests:

• Add disk cache infrastructure for Julia 1.11 (#557) (@vchuravy)
• Fix issue with __init__ when built in sysimg (#589) (@sjkelly)
• Work around buggy Base.pkgversion again. (#594) (@maleadt)
• Fix NVVMReflect pass usage. (#595) (@maleadt)

Closed issues:

• pkgversion can return nothing. (#593)

v0.26.6

GPUCompiler v0.26.6

Diff since v0.26.5

Merged pull requests:

• Adapt to LLVM.jl 8 (#591) (@maleadt)

v0.26.5

GPUCompiler v0.26.5

Diff since v0.26.4

Merged pull requests:

• Update New PassManager Pipeline (#559) (@vchuravy)
• Add missing backticks (#566) (@jonas-schulze)
• Add a testsuite for inference caching (#567) (@vchuravy)
• Fixes for 1.12 (#568) (@maleadt)
• Don't opt into typeinf tracking before v1.11 (#569) (@vchuravy)
• [Metal] Add check for i128 usage (#571) (@tgymnich)
• Add an interface extension point to prepare before anything happens. (#572) (@maleadt)
• Make the optimization level configurable. (#577) (@maleadt)
• Bump julia-actions/setup-julia from 1 to 2 (#578) (@dependabot[bot])
• Adapt to LLVM 17. (#583) (@maleadt)
• Fixes for LLVM 17 (#584) (@maleadt)

Closed issues:

• PTX: Don't lower unreachable on LLVM 17 (#475)
• Opaque pointers support (#511)
• Readiness for LLVM 17 (#573)

v0.26.4

GPUCompiler v0.26.4

Diff since v0.26.3

Merged pull requests:

• Metal: Work around function attribute changes in LLVM 16. (#564) (@maleadt)

v0.26.3

GPUCompiler v0.26.3

Diff since v0.26.2

Merged pull requests:

• Adapt to upcoming Julia changes (#562) (@maleadt)
• Include the compiler config in the runtime cache lookup. (#563) (@maleadt)

Closed issues:

• Cache ignores compiler target properties (#560)

v0.26.2

GPUCompiler v0.26.2

Diff since v0.26.1

Merged pull requests:

• Support internal cache (#545) (@vchuravy)
• Use the new jl_method_lookup_by_tt API on 1.11. (#550) (@maleadt)
• Add CI for 1.11. (#551) (@maleadt)
• Specialize MI if necessary (#553) (@vchuravy)
• Fix construction of byval attribute. (#555) (@maleadt)

Closed issues:

• methodinstance(foo, types) returns MethodError permanently if method is not defined (#530)
• SPIRV_LLVM_Translator needs build for LLVM 16/17 (#543)
• Metal calling convention in LLVM used by ARM in LLVM 16 (#544)