v0.24.2

GPUCompiler v0.24.2

Diff since v0.24.0

Merged pull requests:

• Don't ICE on non-void return types. (#512) (@maleadt)
• Improve error message on unsupported type usage. (#513) (@maleadt)
• Verify IR after our own validation. (#514) (@maleadt)

v0.24.1

GPUCompiler v0.24.1

Diff since v0.24.0

Merged pull requests:

• Don't ICE on non-void return types. (#512) (@maleadt)

v0.24.0

GPUCompiler v0.24.0

Diff since v0.23.0

Closed issues:

• compile(:asm) returns LLVM IR without requiring a context (#420)

Merged pull requests:

• Use NewPM wherever possible (#482) (@pchintalapudi)
• Switch to ReTestItems.jl (#505) (@maleadt)
• Ensure global variables have a safe name. (#507) (@maleadt)
• Update manifest (#509) (@github-actions[bot])
• Don't disallow Bottom returns. (#510) (@maleadt)

v0.23.0

GPUCompiler v0.23.0

Diff since v0.22.0

Closed issues:

• How to set fast math for CUDA (#491)

Merged pull requests:

• remove method_table from ci_cache_populate args (#493) (@vchuravy)
• Bump minimum Julia requirement to Julia 1.8 (#495) (@maleadt)
• Update manifest (#496) (@github-actions[bot])
• 1.11 enablement (#497) (@maleadt)
• enable dependabot for GitHub actions (#498) (@ranocha)
• Bump peter-evans/create-pull-request from 3 to 5 (#499) (@dependabot[bot])
• Bump actions/cache from 1 to 3 (#500) (@dependabot[bot])
• Bump codecov/codecov-action from 1 to 3 (#501) (@dependabot[bot])
• Bump actions/checkout from 2 to 3 (#502) (@dependabot[bot])
• Update manifest (#503) (@github-actions[bot])

v0.22.0

GPUCompiler v0.22.0

Diff since v0.21.4

Closed issues:

• Linking bitcode libraries fail on 1.11 (#489)

Merged pull requests:

• Update manifest (#487) (@github-actions[bot])
• adjust concrete_eval_eligible overload to v1.10 (#488) (@aviatesk)
• Add dwarf_version function so targets can choose (#490) (@gbaraldi)
• Add fastmath flag to PTXCompilerTarget (#492) (@Zentrik)

v0.21.4

GPUCompiler v0.21.4

Diff since v0.21.3

Merged pull requests:

• Correct inlinealloc (#484) (@wsmoses)
• Update manifest (#485) (@github-actions[bot])

v0.20.4

GPUCompiler v0.20.4

Diff since v0.20.3

Closed issues:

• GPUCompiler.compile replaces dots with underscores in extern functions (#464)
• UndefVarError: compile_cache not defined (#470)
• Adapt to specsig changes (#472)

Merged pull requests:

• Don't rename function declarations, only definitions. (#465) (@maleadt)
• PTX: Lower unreachable control flow to avoid bad CFG reconstruction (#467) (@maleadt)
• Update manifest (#468) (@github-actions[bot])
• Define compile_cache outside init() (#471) (@simonbyrne)
• Adapt to upstream changes. (#473) (@maleadt)
• During kernel state rewrite, handle functions passed as arguments. (#474) (@maleadt)
• Adapt to LLVM.jl 6. (#476) (@maleadt)
• Update manifest (#478) (@github-actions[bot])
• Version the compilecache. (#481) (@maleadt)
• Correct inlinealloc (#484) (@wsmoses)
• Update manifest (#485) (@github-actions[bot])

v0.21.3

GPUCompiler v0.21.3

Diff since v0.21.2

Merged pull requests:

• Update manifest (#478) (@github-actions[bot])
• Version the compilecache. (#481) (@maleadt)

v0.21.2

GPUCompiler v0.21.2

Diff since v0.21.1

v0.21.1

GPUCompiler v0.21.1

Diff since v0.21.0

Closed issues:

• UndefVarError: compile_cache not defined (#470)
• Adapt to specsig changes (#472)

Merged pull requests:

• Define compile_cache outside init() (#471) (@simonbyrne)
• Adapt to upstream changes. (#473) (@maleadt)
• During kernel state rewrite, handle functions passed as arguments. (#474) (@maleadt)
• Adapt to LLVM.jl 6. (#476) (@maleadt)