ShardedLevels · finch-tensor/Finch.jl@f607751

Triggered via pull request January 7, 2025 18:57

willow-ahrens

synchronize #654

wma/sharded_levels

Status Failure

Total duration 44m 27s

Artifacts –

CI.yml

on: pull_request

Matrix: test

Annotations

7 errors and 2 notices

Documentation: docs/src/docs/internals/virtualization.md#L286

doctest failure in src/docs/internals/virtualization.md:286-311 ```jldoctest example2; setup = :(using Finch) julia> prgm_inst = Finch.@finch_program_instance for i = _ s[] += A[i] end; julia> println(prgm_inst) loop_instance(index_instance(i), Finch.FinchNotation.Dimensionless(), assign_instance(access_instance(tag_instance(variable_instance(:s), Scalar{0, Int64}(0)), updater_instance()), tag_instance(variable_instance(:+), literal_instance(+)), access_instance(tag_instance(variable_instance(:A), Tensor(SparseList{Int64}(Element{0, Int64, Int64}([2, 3]), 5, [1, 3], [2, 5]))), reader_instance(), tag_instance(variable_instance(:i), index_instance(i))))) julia> prgm_inst Finch program instance: for i = Dimensionless() tag(s, Scalar{0, Int64})[] <<tag(+, +)>>= tag(A, Tensor(SparseList(Element(0))))[tag(i, i)] end julia> prgm = Finch.@finch_program for i = _ s[] += A[i] end; julia> println(prgm) loop(index(i), virtual(Finch.FinchNotation.Dimensionless()), assign(access(literal(Scalar{0, Int64}(0)), literal(Finch.FinchNotation.Updater())), literal(+), access(literal(Tensor(SparseList{Int64}(Element{0, Int64, Int64}([2, 3]), 5, [1, 3], [2, 5]))), literal(Finch.FinchNotation.Reader()), index(i)))) julia> prgm Finch program: for i = virtual(Finch.FinchNotation.Dimensionless) Scalar{0, Int64}(0)[] <<+>>= Tensor(SparseList{Int64}(Element{0, Int64, Int64}([2, 3]), 5, [1, 3], [2, 5]))[i] end ``` Subexpression: println(prgm) Evaluated output: loop(index(i), virtual(Finch.FinchNotation.Dimensionless()), assign(access(literal(Scalar{0, Int64}(0)), updater()), literal(+), access(literal(Tensor(SparseList{Int64}(Element{0, Int64, Int64}([2, 3]), 5, [1, 3], [2, 5]))), reader(), index(i)))) Expected output: loop(index(i), virtual(Finch.FinchNotation.Dimensionless()), assign(access(literal(Scalar{0, Int64}(0)), literal(Finch.FinchNotation.Updater())), literal(+), access(literal(Tensor(SparseList{Int64}(Element{0, Int64, Int64}([2, 3]), 5, [1, 3], [2, 5]))), literal(Finch.FinchNotation.Reader()), index(i)))) diff = Warning: Diff output requires color. loop(index(i), virtual(Finch.FinchNotation.Dimensionless()), assign(access(literal(Scalar{0, Int64}(0)), literal(Finch.FinchNotation.Updater())), updater()), literal(+), access(literal(Tensor(SparseList{Int64}(Element{0, Int64, Int64}([2, 3]), 5, [1, 3], [2, 5]))), literal(Finch.FinchNotation.Reader()), reader(), index(i))))

Documentation: docs/src/docs/internals/virtualization.md#L319

doctest failure in src/docs/internals/virtualization.md:319-331 ```jldoctest example2; setup = :(using Finch, AbstractTrees, SyntaxInterface, RewriteTools) julia> using Finch.FinchNotation; julia> PostOrderDFS(prgm) PostOrderDFS{FinchNode}(loop(index(i), virtual(Dimensionless()), assign(access(literal(Scalar{0, Int64}(0)), literal(Updater())), literal(+), access(literal(Tensor(SparseList{Int64}(Element{0, Int64, Int64}([2, 3]), 5, [1, 3], [2, 5]))), literal(Reader()), index(i))))) julia> (@capture prgm loop(~idx, ~ext, ~val)) true julia> idx Finch program: i ``` Subexpression: PostOrderDFS(prgm) Evaluated output: PostOrderDFS{FinchNode}(loop(index(i), virtual(Dimensionless()), assign(access(literal(Scalar{0, Int64}(0)), updater()), literal(+), access(literal(Tensor(SparseList{Int64}(Element{0, Int64, Int64}([2, 3]), 5, [1, 3], [2, 5]))), reader(), index(i))))) Expected output: PostOrderDFS{FinchNode}(loop(index(i), virtual(Dimensionless()), assign(access(literal(Scalar{0, Int64}(0)), literal(Updater())), literal(+), access(literal(Tensor(SparseList{Int64}(Element{0, Int64, Int64}([2, 3]), 5, [1, 3], [2, 5]))), literal(Reader()), index(i))))) diff = Warning: Diff output requires color. PostOrderDFS{FinchNode}(loop(index(i), virtual(Dimensionless()), assign(access(literal(Scalar{0, Int64}(0)), literal(Updater())), updater()), literal(+), access(literal(Tensor(SparseList{Int64}(Element{0, Int64, Int64}([2, 3]), 5, [1, 3], [2, 5]))), literal(Reader()), reader(), index(i)))))

Documentation