mirror of
https://github.com/matrix-construct/construct
synced 2024-06-18 01:48:27 +02:00
configure: Additional half-precision type checks.
This commit is contained in:
parent
57f9d3fdfb
commit
3a5bcbb0dd
|
@ -990,9 +990,12 @@ AC_CHECK_TYPES([__uint128_t])
|
|||
AC_CHECK_SIZEOF([__uint128_t])
|
||||
|
||||
dnl Gauge the support of experimental compact floating point formats
|
||||
AC_CHECK_SIZEOF([__fp16])
|
||||
AC_CHECK_SIZEOF([__bf16])
|
||||
AC_CHECK_TYPES([_Float16])
|
||||
AC_CHECK_SIZEOF([_Float16])
|
||||
AC_CHECK_TYPES([__fp16])
|
||||
AC_CHECK_SIZEOF([__fp16])
|
||||
AC_CHECK_TYPES([__bf16])
|
||||
AC_CHECK_SIZEOF([__bf16])
|
||||
|
||||
dnl -fsigned-char ensures 'char' is signed on all platforms
|
||||
RB_VAR_PREPEND([CFLAGS], ["-fsigned-char"])
|
||||
|
|
Loading…
Reference in a new issue