0.21.3

Added

• Parse errors now list possible expected tokens.

• Lexer errors now mention the file.

Fixed

• Overloaded number literals cannot be sum types (#1557).

• Defective GPU code generation for vectorised non-commutative
  operatators (#1559).

• Excessive memory usage for some programs (#1325).

0.21.2

Added

• New functions: reduce_by_index_2d, reduce_by_index_3d.

• Manifests now contain compiler version information.

Fixed

• Allocation insertion pass bug (#1546).

• An exotic bug involving TLS and dynamically loading code generated
  by the multicore backend.

• Unconstrained ambiguous types now default to () (#1552). This
  should essentially never have any observable impact, except that
  more programs will type check.

• Double buffering compiler crash (#1553).

0.21.1

Added

• Top-level value definitions can (and should) now be declared with
  with def, although let still works.

• New tool: futhark defs, for printing locations of top-level
  definitions.

Changed

• def is now a reserved word.

Fixed

• Contrived intra-group code versions with no actual parallelism would
  be given a group size of zero (#1524).

0.20.8

Added

• futhark repl now allows Ctrl-c to interrupt execution.

Fixed

• Alias tracking of sum types.

• Proper checking that a function declared to return a unique-typed
  value actually does so.

• Faulty uniqueness checking and inference for lambdas (#1535).

• Monomorphisation would duplicate functions under rare circumstances
  (#1537).

• Interpreter didn't check that the arguments passed to unflatten
  made sense (#1539).

• futhark literate now supports a $loaddata builtin function for
  passing datasets to Futhark programs.

0.20.7

Added

• Better exploitation of parallelism in fused nested segmented
  reductions.

• Prelude function not for negating booleans.

Fixed

• Some incorrect removal of copies (#1505).

• Handling of parametric modules with top-level existentials (#1510).

• Module substitution fixes (#1512, #1518).

• Invalid in-place lowering (#1523).

• Incorrect code generation for some intra-group parallel code versions.

• Flattening crash in the presence of irregular parallelism (#1525).

• Incorrect substitution of type abbreviations with hidden sizes (#1531).

• Proper handling of NaN in min/max functions for
  f16/f32/f64 in interpreter (#1528).

0.20.6

Added

• Much better code generation for segmented scans with vectorisable
  operators.

Fixed

• Fixes to extremely exotic GPU scans involving array operators.

• Missing alias tracking led to invalid rewrites, causing a compiler
  crash (#1499).

• Top-level bindings with existential sizes were mishandled (#1500, #1501).

• A variety of memory leaks in the multicore backend, mostly (or
  perhaps exclusively) centered around context freeing or failing
  programs - this should not have affected many people.

• Various fixes to f16 handling in the GPU backends.

0.20.5

Added

• Existential sizes can now be explicitly quantified in type
  expressions (#1308).

• Significantly expanded error index.

• Attributes can now be numeric.

• Patterns can now have attributes. None have any effect at the
  moment.

• futhark autotune and futhark bench now take a --spec-file
  option for loading a test specification from another file.

Fixed

• auto output reference datasets are now recreated when the program
  is newer than the data files.

• Exotic hoisting bug (#1490).

0.20.4

Added

• Tuning parameters now (officially) exposed in the C API.

• futhark autotune is now 2-3x faster on many programs, as it now
  keeps the process running.

• Negative numeric literals are now allowed in case patterns.

Fixed

• futhark_context_config_set_profiling was missing for the c backend.

• Correct handling of nested entry points (#1478).

• Incorrect type information recorded when doing in-place lowering (#1481).

0.20.3

Added

• Executables produced by C backends now take a --no-print-result option.

• The C backends now generate a manifest when compiling with
  --library. This can be used by FFI generators (#1465).

• The beginnings of a Rust-style error index.

• scan on newer CUDA devices is now much faster.

Fixed

• Unique opaque types are named properly in entry points.

• The CUDA backend in library mode no longer exit()s the process if
  NVRTC initialisation fails.

0.20.2

Fixed

• Simplification bug (#1455).

• In-place-lowering bug (#1457).

• Another in-place-lowering bug (#1460).

• Don't try to tile inside loops with parameters with variant sizes (#1462).

• Don't consider it an ICE when the user passes invalid command line
  options (#1464).