0.10.2

Added

• reduce_by_index is now a good bit faster on operators whose
  arguments are two 32-bit values.

• The type checker warns on size annotations for function parameters
  and return types that will not be visible from the outside,
  because they refer to names nested inside tuples or records. For
  example, the function

  let f (n: i32, m: i32): [n][m]i32 = ...
  

  will cause such a warning. It should instead be written

  let f (n: i32) (m: i32): [n][m]i32 = ...
  

• A new library function
  futhark_context_config_select_device_interactively() has been
  added.

Fixed

• Fix reading and writing of binary files for C-compiled executables
  on Windows.

• Fixed a couple of overly strict internal sanity checks related to
  in-place updates (#735, #736).

• Fixed a couple of convoluted defunctorisation bugs (#739).

0.10.1

Added

• Using definitions from the intrinsic module outside the prelude
  now results in a warning.

• reduce_by_index with vectorised operators (e.g. map2 (+)) is
  orders of magnitude faster than before.

• Executables generated with the pyopencl backend now support the
  options --default-tile-size, --default-group-size,
  --default-num-groups, --default-threshold, and --size.

• Executables generated with c and opencl now print a help text
  if run with invalid options. The py and pyopencl backends
  already did this.

• Generated executables now support a --tuning flag for passing
  many tuned sizes in a file.

• Executables generated with the cuda backend now take an
  --nvrtc-option option.

• Executables generated with the opencl backend now take a
  --build-option option.

Removed

• The old futhark-* executables have been removed.

Changed

• If an array is passed for a function parameter of a polymorphic
  type, all arrays passed for parameters of that type must have the
  same shape. For example, given a function

  let pair 't (x: t) (y: t) = (x, y)
  

  The application pair [1] [2,3] will now fail at run-time.

• futhark test now numbers un-named data sets from 1 rather than
  0. This only affects the text output and the generated JSON
  files, and fits the tuple element ordering in Futhark.

• String literals are now of type []u8 and contain UTF-8 encoded
  bytes.

Fixed

• An significant problematic interaction between empty arrays and
  inner size declarations has been closed (#714). This follows a
  range of lesser empty-array fixes from 0.9.1.

• futhark datacmp now prints to stdout, not stderr.

• Fixed a major potential out-of-bounds access when sequentialising
  reduce_by_index (in most cases the bug was hidden by subsequent
  C compiler optimisations).

• The result of an anonymous function is now also forbidden from
  aliasing a global variable, just as with named functions.

• Parallel scans now work correctly when using a CPU OpenCL
  implementation.

• reduce_by_index was broken on newer NVIDIA GPUs when using fancy
  operators. This has been fixed.

0.9.1

Added

• futhark cuda: a new CUDA backend by Jakob Stokholm Bertelsen.

• New command for comparing data files: futhark datacmp.

• An :mtype command for futhark repl that shows the type of a
  module expression.

• futhark run takes a -w option for disabling warnings.

Changed

• Major command reorganisation: all Futhark programs have been
  combined into a single all-powerful futhark program. Instead of
  e.g. futhark-foo, use futhark foo. Wrappers will be kept
  around under the old names for a little while. futharki has
  been split into two commands: futhark repl and futhark run.
  Also, py has become python and cs has become csharp, but
  pyopencl and csopencl have remained as they were.

• The result of a function is now forbidden from aliasing a global
  variable. Surprisingly little code is affected by this.

• A global definition may not be ascribed a unique type. This never
  had any effect in the first place, but now the compiler will
  explicitly complain.

• Source spans are now printed in a slightly different format, with
  ending the line number omitted when it is the same as the start
  line number.

Fixed

• futharki now reports source locations of trace expressions
  properly.

• The type checker now properly complains if you try to define a
  type abbreviation that has unused size parameters.

0.8.1

Added

• Now warns when /futlib/... files are redundantly imported.

• futharki now prints warnings for files that are ":load"ed.

• The compiler now warns when entry points are declared with types
  that will become unnamed and opaque, and thus impossible to
  provide from the outside.

• Type variables invented by the type checker will now have a
  unicode subscript to distinguish them from type parameters
  originating in the source code.

• futhark-test and futhark-bench now support generating random
  test data.

• The library backends now generate proper names for arrays of
  opaque values.

• The parser now permits empty programs.

• Most transpositions are now a good bit faster, especially on
  NVIDIA GPUs.

Removed

• The <- symbol can no longer be used for in-place updates and
  record updates (deprecated in 0.7.3).

Changed

• Entry points that accept a single tuple-typed parameter are no
  longer silently rewritten to accept multiple parameters.

Fixed

• The :type command in futharki can now handle polymorphic
  expressions (#669).

• Fixed serious bug related to chaining record updates.

• Fixed type inference of record fields (#677).

• futharki no longer goes in an infinite loop if a for loop
  contains a negative upper bound.

• Overloaded number types can no longer carry aliases (#682).

0.7.4

Added

• Support type parameters for operator specs defined with val.

Fixed

• Fixed nasty defunctionalisation bug (#661).

• cabal/stack sdist works now.

0.7.3

Added

• Significant performance changes: there is now a constant extra
  compilation overhead (less than 200ms on most machines). However,
  the rest of the compiler is 30-40% faster (or more in some cases).

• A warning when ambiguously typed expressions are assigned a
  default (i32 or f64).

• In-place updates and record updates are now written with =
  instead of <-. The latter is deprecated and will be removed in
  the next major version (#650).

Fixed

• Polymorphic value bindings now work properly with module type
  ascription.

• The type checker no longer requires types used inside local
  functions to be unambiguous at the point where the local function
  is defined. They must still be unambiguous by the time the
  top-level function ends. This is similar to what other ML
  languages do.

• futhark-bench now writes "μs" instead of "us".

• Type inference for infix operators now works properly.

0.7.2

Added

• futhark-pkg now supports GitLab.

• futhark-tests --notty option now has a --no-terminal alias.
  --notty is deprecated, but still works.

• futhark-test now supports multiple entry points per test block.

• Functional record updates: r with f <- x.

Fixed

• Fix the -C option for futhark-test.

• Fixed incorrect type of reduce_by_index.

• Segmented reduce_by_index now uses much less memory.

0.7.1

Added

• C# backend by Mikkel Storgaard Knudsen (futhark-cs/futhark-csopencl).

• futhark-test and futhark-bench now take a --runner option.

• futharki now uses a new interpreter that directly interprets the
  source language, rather than operating on the desugared core
  language. In practice, this means that the interactive mode is
  better, but that interpretation is also much slower.

• A trace function that is semantically id, but makes futharki
  print out the value.

• A break function that is semantically id, but makes futharki
  stop and provide the opportunity to inspect variables in scope.

• A new SOAC, reduce_by_index, for expressing generalised
  reductions (sometimes called histograms). Designed and
  implemented by Sune Hellfritzsch.

Removed

• Most of futlib has been removed. Use external packages instead:

  • futlib/colour => https://github.com/athas/matte

  • futlib/complex => https://github.com/diku-dk/complex

  • futlib/date => https://github.com/diku-dk/date

  • futlib/fft => https://github.com/diku-dk/fft

  • futlib/linalg => https://github.com/diku-dk/fft

  • futlib/merge_sort, futlib/radix_sort => https://github.com/diku-dk/sorts

  • futlib/random => https://github.com/diku-dk/cpprandom

  • futlib/segmented => https://github.com/diku-dk/segmented

  • futlib/sobol => https://github.com/diku-dk/sobol

  • futlib/vector => https://github.com/athas/vector

  No replacement: futlib/mss, futlib/lss.

• zip6/zip7/zip8 and their unzip variants have been removed.
  If you build gigantic tuples, you're on your own.

• The >>> operator has been removed. Use an unsigned integer type
  if you want zero-extended right shifts.

Changed

• The largest/smallest values for numeric modules have been
  renamed highest/lowest.

Fixed

• Many small things.

0.6.3

Added

• Added a package manager: futhark-pkg. See also the
  documentation.

• Added log2 and log10 functions to f32 and f64.

• Module type refinement (with) now permits refining parametric
  types.

• Better error message when invalid values are passed to generated
  Python entry points.

• futhark-doc now ignores files whose doc comment is the word
  "ignore".

• copy now works on values of any type, not just arrays.

• Better type inference for array indexing.

Fixed

• Floating-point numbers are now correctly rounded to nearest even
  integer, even in exotic cases (#377).

• Fixed a nasty bug in the type checking of calls to consuming
  functions (#596).

0.6.2

Added

• Bounds checking errors now show the erroneous index and the size
  of the indexed array. Some other size-related errors also show
  more information, but it will be a while before they are all
  converted (and say something useful - it's not entirely
  straightforward).

• Opaque types now have significantly more readable names,
  especially if you add manual size annotations to the entry point
  definitions.

• Backticked infix operators can now be used in operator sections.

Fixed

• f64.e is no longer pi.

• Generated C library code will no longer abort() on application
  errors (#584).

• Fix file imports on Windows.

• futhark-c and futhark-opencl now generates thread-safe code (#586).

• Significantly better behaviour in OOM situations.

• Fixed an unsound interaction between in-place updates and
  parametric polymorphism (#589).