v0.1.9

FoldsCUDA v0.1.9

Diff since v0.1.8

Merged pull requests:

• Use TestFunctionRunner (#87) (@tkf)
• Use StaticArrays to compute histogram (#94) (@tkf)

v0.1.8

FoldsCUDA v0.1.8

Diff since v0.1.7

Merged pull requests:

• Use CUDA.return_type to infer device code (#91) (@tkf)
• Require CUDA.jl 3.8 (#92) (@tkf)

v0.1.7

FoldsCUDA v0.1.7

Diff since v0.1.6

Merged pull requests:

• Test with Julia 1.7 (#74) (@tkf)
• Coalesced non-commutative reduction using shfl (#78) (@tkf)
• Support Union values in shfl-based kernel (#80) (@tkf)
• Fix shfl-based reduction for large inputs (#81) (@tkf)
• Drop Julia 1.5 support (#83) (@tkf)
• Setup benchmark suite (#84) (@tkf)
• Fix for complete with side effects (#86) (@tkf)
• Support InitialValues 0.3 (#88) (@tkf)

v0.1.6

FoldsCUDA v0.1.6

Diff since v0.1.5

Merged pull requests:

• Test with TeeRF (#66) (@tkf)
• Support pairs (#67) (@tkf)
• Add a partition reduce example (#68) (@tkf)
• Don't directly depend on GPUArrays (#71) (@tkf)
• Free/fix docs/Manifest.toml (#72) (@tkf)
• Restrict to Random123 v1.3 for now (#73) (@tkf)
• Remove redundant synchronizations (#75) (@tkf)
• Don't call next on dummy; it's not pure (#76) (@tkf)
• Require UnionArrays 0.1.2 (#79) (@tkf)

v0.1.5

FoldsCUDA v0.1.5

Diff since v0.1.4

Merged pull requests:

• Support type-changing accumulator (#49) (@tkf)
• Turn tests to a package (#51) (@tkf)
• Update FLoops to the version that invokes init in loop (#52) (@tkf)
• Update CUDA.jl; stop using my fork (#53) (@tkf)
• Fix CPU tests (#54) (@tkf)
• Update UnionArrays (#55) (@tkf)
• Test with CUDA.jl 3.0 in Julia 1.6 (#57) (@tkf)

v0.1.4

FoldsCUDA v0.1.4

Diff since v0.1.3

Closed issues:

• Skip the non-GPU tests on Buildkite (#40)

Merged pull requests:

• Test with Folds.jl (#37) (@tkf)
• Set up Buildkite (#39) (@DilumAluthge)
• Demo in-place mutation with Referenceables.jl (#43) (@tkf)
• Demo simple in-place mutation (#44) (@tkf)
• Run CI with Julia 1.6 (#45) (@tkf)
• Disallow scalar in test (#46) (@tkf)
• Require Transducers 0.4.60 (#50) (@tkf)

v0.1.3

FoldsCUDA v0.1.3

Diff since v0.1.2

Merged pull requests:

• Use CUDAEx for CuArrays automatically (#36) (@tkf)

v0.1.2

FoldsCUDA v0.1.2

Diff since v0.1.1

Closed issues:

• An error from counter-based RNG example: .parent is of type Array{Tuple{Tuple{Int64}},1} which is not isbits. (#29)

Merged pull requests:

• Import Executor from Transducers.jl (#35) (@tkf)

v0.1.1

FoldsCUDA v0.1.1

Diff since v0.1.0

Merged pull requests:

• Demo with lazy mapping (and fix a bug) (#24) (@tkf)
• Add an example: estimating pi using Monte-Carlo method (#26) (@tkf)

v0.1.0

FoldsCUDA v0.1.0