max(tensor, axis = 1) doesn't work without TRITON_INTERPRET=1

[wa008]

Describe the bug

tensor.max(axis = 1) doesn't work and tl.max(tensor, axis = 1) works well without TRITON_INTERPRET=1

Actually, It's not a big deal...

key code:

    new_max_val = input_val.max(axis = 1) # it doesn't work
    # new_max_val = tl.max(input_val, axis = 1) # it works well

error information:

RuntimeError: Cannot call @triton.jit'd outside of the scope of a kernel

The above exception was the direct cause of the following exception:

CompilationError                          Traceback (most recent call last)
[/usr/local/lib/python3.10/dist-packages/triton/compiler/compiler.py](https://localhost:8080/#) in make_ir(self, options, codegen_fns, context)
    111 
    112     def make_ir(self, options, codegen_fns, context):
--> 113         return ast_to_ttir(self.fn, self, context=context, options=options, codegen_fns=codegen_fns)
    114 
    115     def parse_options(self):

CompilationError: at 23:26:
    target = tl.load(target_ptr + offsets_rows, mask = offsets_rows < M, other = 0.0)
    max_val = tl.full(target.shape, -float("inf"), dtype = tl.float32)
    sumexp = tl.zeros_like(max_val)

    allcurx = tl.zeros_like(max_val)
    for index in tl.range(0, N, BLOCK_SIZE_N):
        offsets_input = offsets_rows[:, None] * N + (offsets_cols + index)[None, :]
        mask_input = (offsets_rows[:, None] < M) & ((offsets_cols + index)[None, :] < N)
        input_val = tl.load(input_ptr + offsets_input, mask=mask_input, other = -float("inf"))

        if index == 0:
            new_max_val = input_val.max(axis = 1) # it doesn't work

full code: https://drive.google.com/file/d/1UBXqpBhELn8amEOU8h5zj1i6WvihZKJX/view?usp=drive_link

Environment details

Triton: 3.1.0
Environment: colab

[Jokeren]

Please attach a reproducer

[wa008]

Please attach a reproducer

I updated key code and full code on description, Thanks for your attention!

[Jokeren]

I cannot reproduce it on my end. Very possibly you're not using the latest version. Please consider build triton from source and retry.