v0.2a

Bug fixes.

Known issue:

• Multi-precision code with if branch fails to generate mask for A64FX.
• Multi-precision mode generates slow code. Thus, for example, you should generate (F|S|U)64-only code instead of (F|S|U)64/(F|S|U)32 multi-precision code.

v0.2

Bug fix for A64FX (ARM SVE) mode. Test implementation of CUDA kernel.

v0.1c

New grammer for --class-file option. bug fix for if statement is included.

v0.1b

bug fix

v0.1a

bug fix for c and fortran interface. add sample/c/Nbody.

v0.1

Add c interface and fortran interface

v0.0c

bug fix

v0.0b

Modify if else state specification in doc/kernel_generator_spec.pdf.

v0.0a

Various bugs are modified. No new feature.

v0.0

First release