[flang-commits] [PATCH] D139351: Lowering and runtime support for F08 transformational intrinsics: BESSEL_JN and BESSEL_YN

[Tarun Prabhu via Phabricator via flang-commits]

tarunprabhu updated this revision to Diff 481741.
tarunprabhu added a comment.

Significant changes to the patch:

- Use fir::IfOp so anchor points for the recursion are only calculated if necessary.

- Created runtime functions specialized for the `x == 0` case. This has cleaned up the signature of the runtime functions so only those arguments that are necessary are passed along.

- Use int32_t in the signature for the runtime functions instead of relying on the width of the system's `int` datatype.

- Added tests for the genBessel* lowering functions in addition to tests of the runtime functions themselves.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D139351/new/

https://reviews.llvm.org/D139351

Files:
  flang/include/flang/Optimizer/Builder/Runtime/Transformational.h
  flang/include/flang/Runtime/transformational.h
  flang/lib/Lower/IntrinsicCall.cpp
  flang/lib/Optimizer/Builder/Runtime/Transformational.cpp
  flang/runtime/transformational.cpp
  flang/test/Lower/Intrinsics/bessel_jn.f90
  flang/test/Lower/Intrinsics/bessel_yn.f90
  flang/unittests/Optimizer/Builder/Runtime/TransformationalTest.cpp
  flang/unittests/Runtime/Transformational.cpp

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D139351.481741.patch
Type: text/x-patch
Size: 67054 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/flang-commits/attachments/20221209/1018ed6b/attachment-0001.bin>