v0.6.16

Finch v0.6.16

Diff since v0.6.15

Merged pull requests:

• better short circuits (#455) (@willow-ahrens)
• Update README.md (#458) (@willow-ahrens)
• Wma/shortcircuit (#459) (@willow-ahrens)
• more tweaks (#461) (@willow-ahrens)

v0.6.15

Finch v0.6.15

Diff since v0.6.14

Merged pull requests:

• High-Level Interface (#389) (@willow-ahrens)
• Add no-merge option for SparseRLE and DenseRLE, clean up docs and tests for levels (#449) (@willow-ahrens)
• Wma/fix countstored (#450) (@willow-ahrens)

Closed issues:

• Add a Tensor Utilities docs page (#228)
• Add const aliases for continuous paper terminology (#334)
• High-Level Interface (#377)
• virtual_size called too often (#448)
• DOK constructor issues (#451)
• Add dropdims (#453)

v0.6.14

Finch v0.6.14

Diff since v0.6.13

Merged pull requests:

• add a heavy compilation benchmark (#431) (@willow-ahrens)
• Improve Compiler Runtime by 25x (#432) (@willow-ahrens)
• Wma/blur fixes (#433) (@willow-ahrens)
• Merge Identical Runs in SparseRLE (#436) (@willow-ahrens)
• Fix SparseVBLLevel bug (#437) (@radha-patel)
• API: Add fill_value and fill_value_type config argument to Base.similar (#438) (@mtsokol)
• update docs (#439) (@willow-ahrens)
• Parameterize the test suite more (#440) (@willow-ahrens)
• okay, adds sparsebandlevels (#441) (@willow-ahrens)
• Small fix to fsparse_impl (#442) (@kylebd99)
• Fix SingleList (#445) (@nullplay)
• Wma/fix sparse band (#446) (@willow-ahrens)

Closed issues:

• Usage with index set construction (#318)
• Single block level (#330)
• Making it clear how dimensionalization works (#335)
• Understand and improve behavior on large kernels (#338)
• MethodError in 1-D Tensor indexing (#427)
• Multidimensional RLE (#435)
• Error for SparseBand in copyto! (#443)
• Wrong behavior of Spike (#444)
• Error for SingleList and SingleRLE in 2D Tensor creation (#447)

v0.6.13

Finch v0.6.13

Diff since v0.6.12

Merged pull requests:

• API: Add list indexing for 1D Tensors (#429) (@mtsokol)
• Update Project.toml (#430) (@willow-ahrens)

v0.6.12

Finch v0.6.12

Diff since v0.6.11

Merged pull requests:

• Add a return statement to finch (#394) (@willow-ahrens)
• add 1D hash, parameterize datastructure. (#411) (@willow-ahrens)
• API: Add eltype for SwizzleArray (#415) (@mtsokol)
• Wma/fix printing (#416) (@willow-ahrens)
• Wma/freezetrim (#418) (@willow-ahrens)
• Wma/continuousarray (#419) (@willow-ahrens)
• Wma/fix separate (#420) (@willow-ahrens)
• Wma/implicit yield (#422) (@willow-ahrens)
• Wma/protodocs (#423) (@willow-ahrens)
• Add a getindex for swizzles and also talk about code_warntype (#424) (@willow-ahrens)
• Update Project.toml (#425) (@willow-ahrens)

Closed issues:

• Add an optimized hash level for one dimension (#227)
• Merge continuous infrastructure to Main (#328)
• AsArray for SwizzleArrays (#397)
• CUDA.JL style atomics (#398)
• Seperation Level Clean up (#399)
• scansearch fails for scipy.sparse indices arrays (#406)
• Move trim! into freeze! (#412)
• make fiber printing less complicated (#413)
• Tensor <: AbstractArray (#414)
• Are Protocols Gone? (#421)

v0.6.11

Finch v0.6.11

Diff since v0.6.10

Merged pull requests:

• Trilinear Interpolation on ray sampling (#404) (@nullplay)
• API: Introduce OffByOneVector (#405) (@mtsokol)
• Remove type annotations for scansearch args (#407) (@mtsokol)
• Common Subexpression Elimination (#408) (@willow-ahrens)
• cut a release! (#410) (@willow-ahrens)

Closed issues:

• Common Subexpression Elimination on input IR (#134)
• RLE Tree level (#336)
• Off By One Array (#381)
• Copying in Tensor(lvl, arr) constructor (#403)

v0.6.10

Finch v0.6.10

Diff since v0.6.9

Merged pull requests:

• Tfc/atomic (#382) (@wraith1995)
• cleanup atomics (#401) (@willow-ahrens)
• Wma/swizzleswizzle (#402) (@willow-ahrens)

Closed issues:

• Atomic Array Wrapper (#246)
• Make dimensions mutable (#393)
• Clean up make_lock (#400)

v0.6.9

Finch v0.6.9

Diff since v0.6.8

Merged pull requests:

• continuous examples (#392) (@nullplay)
• Fix a swizzle issue (#395) (@willow-ahrens)
• bugfix release (#396) (@willow-ahrens)

Closed issues:

• Multiple let statements not parsing properly (#385)
• Question regarding 3d swizzle array (#387)

v0.6.8

Finch v0.6.8

Diff since v0.6.7

Merged pull requests:

• Wma/fix staging (#371) (@willow-ahrens)
• reorganize docs (#372) (@willow-ahrens)
• SingleRLE level (#373) (@nullplay)
• SingleList Level (#374) (@nullplay)
• Wma/leveldocs (#375) (@willow-ahrens)
• update status (#376) (@willow-ahrens)
• Wma/pretty instance (#379) (@willow-ahrens)
• add github link (#380) (@willow-ahrens)
• SparseRLE thaw (#384) (@nullplay)
• Fix Let Syntax Bug (#386) (@radha-patel)
• update deps (#388) (@willow-ahrens)

Closed issues:

• Docs for the levels (#226)
• Single run level (#329)
• Singleton level (#331)
• Python interface (#333)
• Function staging is broken for compiled code that contains closures. (#365)

v0.6.7

Finch v0.6.7

Diff since v0.6.6

Merged pull requests:

• Pointer Levels (#304) (@willow-ahrens)
• Better errors on unloaded extensions (#344) (@willow-ahrens)
• import changes to limits (#345) (@willow-ahrens)
• rename base to interface (#346) (@willow-ahrens)
• constructors to fix 321 (#348) (@willow-ahrens)
• Scale, Product Array Combinator (#349) (@nullplay)
• ensure the errors are defined (#350) (@willow-ahrens)
• Wma/fix 317 (#352) (@willow-ahrens)
• remove dead tests (#354) (@willow-ahrens)
• Dataflow Bug (#355) (@nullplay)
• Wma/sparselistthaw (#359) (@willow-ahrens)
• Update CI.yml (#360) (@willow-ahrens)
• make fsparse, etc. match definitions in Base (#361) (@willow-ahrens)
• make static hash more static (#363) (@willow-ahrens)
• Wma/rename fiber (#367) (@willow-ahrens)
• oops (#368) (@willow-ahrens)
• Update Project.toml (#369) (@willow-ahrens)

Closed issues:

• More Levels RFC (#81)
• RFC: Rename Fiber to TensorTree or just Tensor (#301)
• getroot() function (#317)
• UndefVarError: FinchProtocolError not defined (#319)
• Implement export functions for Finch Fibers (#321)
• Bugs in dataflow analysis (#339)
• Separated Memory Levels (#342)
• Better errors for missing fileio includes (#343)
• oh no! (#351)
• fsparse, etc. need to have identical interfaces to those in Base (#356)
• Consider block allocated list datastructure for parallelism (#357)
• All levels should support thaw (#358)