After the -fdefault-real-8 Option Is Enabled in Flang, You Are Not Advised to Use the Double Intrinsic Type

When -fdefault-real-8 is enabled, GFortran can upgrade the double intrinsic type to the corresponding quadruple-precision intrinsic type. However, Bisheng Compiler does not support the upgrade of the double intrinsic type. If the double intrinsic type is used, the precision may be incorrect.

Therefore, when -fdefault-real-8 is enabled, you are advised to use generic intrinsic types. In the following case, you can use acos instead of dacos.

1
2

intrinsic dacos
print *, dacos(0.65d0)

Change the preceding content to the following:

1

print *, acos(0.65d0)