v0.3.2

MultiBroadcastFusion v0.3.2

Diff since v0.3.1

Merged pull requests:

• Revamp benchmarks (#36) (@charleskawczynski)
• Add buildkite pipeline for CUDA tests (#37) (@charleskawczynski)
• Update benchmarks (#38) (@charleskawczynski)
• Use verbose names for device types (#39) (@charleskawczynski)
• Split unit tests from benchmarks (#42) (@charleskawczynski)
• Bump patch version for new release (#44) (@charleskawczynski)

Closed issues:

• Add CUDA tests (#34)
• fused_copyto! for CUDA assumes linear indexing (#40)
• Split off unit tests from performance benchmarks (#41)

v0.3.1

MultiBroadcastFusion v0.3.1

Diff since v0.3.0

Merged pull requests:

• Add support for Julia 1.9 (#30) (@Sbozzolo)

Closed issues:

• Is julia 1.10 required? (#29)

v0.3.0

MultiBroadcastFusion v0.3.0

Diff since v0.2.0

Merged pull requests:

• Remove unused code (#18) (@charleskawczynski)
• Add support for Arrays/CuArrays with CUDA ext (#19) (@charleskawczynski)
• Fix cuda benchmark (#20) (@charleskawczynski)
• Generalize @make_fused (#21) (@charleskawczynski)
• Move tests into subfolders (#22) (@charleskawczynski)
• Improve names and docs (#23) (@charleskawczynski)
• Bump minor version (#25) (@charleskawczynski)

Closed issues:

• Add limited support for common types (#6)
• Improve names (#11)

v0.2.0

MultiBroadcastFusion v0.2.0

Diff since v0.1.1

Merged pull requests:

• Split macro into type and at-fused (#16) (@charleskawczynski)

Closed issues:

• Split the macro (#14)

v0.1.1

MultiBroadcastFusion v0.1.1

Merged pull requests:

• Fix test bugs, improve tests, restrict loops, func calls, and if-else (#1) (@charleskawczynski)
• Add broken test (#2) (@charleskawczynski)
• Fix code_lowered_single_expression bug (#3) (@charleskawczynski)
• Improve CPU performance implementation (#7) (@charleskawczynski)
• Make things more strict (#8) (@charleskawczynski)
• Update tagbot (#10) (@charleskawczynski)
• Add docs for custom macros (#12) (@charleskawczynski)
• Add support for fused pairs across barriers (#13) (@charleskawczynski)

Closed issues:

• Fix fused CPU kernel performance (#4)
• Add support for fused pairs across barriers (#5)
• TagBot trigger issue (#9)