Uint fix

[rajuptvs]

Changed occurrences of Int to UInt to prevent deprecation warnings mentioned in #119
Please let me know if this is inline with the suggested changes.

Thanks!

[ehsanmok]

Thanks a lot! Could you pixi run format to fix the formatting please?

[rajuptvs]

Thanks a lot! Could you pixi run format to fix the formatting please?

yep, my bad! thank you.. fixed it

[rajuptvs]

Also while going through the puzzles myself..
found few occurences where there wasn't a special init as Int of variables

fn process_sliding_window(
    output: LayoutTensor[mut=True, dtype, vector_layout],
    a: LayoutTensor[mut=False, dtype, vector_layout],
):
    thread_id = thread_idx.x

    # Each thread processes a sliding window of 3 elements
    window_sum = Scalar[dtype](0.0)

    # Sum elements in sliding window: [i-1, i, i+1]
    for offset in range(ITER):
        idx = thread_id + offset - 1
        if 0 <= idx < SIZE:
            value = rebind[Scalar[dtype]](a[idx])
            window_sum += value

    output[thread_id] = window_sum

Output

[Thread 0x7fffa57fa640 (LWP 1345974) exited]
/home/raju/personal/mojo-gpu-puzzles/problems/p09/p09.mojo:41:27: warning: deprecated implicit conversion from 'Int' to 'UInt'
        idx = thread_id + offset - 1
                          ^~~~~~
/home/raju/personal/mojo-gpu-puzzles/problems/p09/p09.mojo:1:1: note: '@implicit' constructor 'UInt.__init__' declared here
from memory import UnsafePointer

This particular case I have converted locally by type casting like below, but not in this commit

fn process_sliding_window(
    output: LayoutTensor[mut=True, dtype, vector_layout],
    a: LayoutTensor[mut=False, dtype, vector_layout],
):
    thread_id = thread_idx.x

    # Each thread processes a sliding window of 3 elements
    window_sum = Scalar[dtype](0.0)

    # Sum elements in sliding window: [i-1, i, i+1]
    for offset in range(ITER):
        idx = thread_id + UInt(offset) - 1
        if 0 <= idx < SIZE:
            value = rebind[Scalar[dtype]](a[idx])
            window_sum += value

    output[thread_id] = window_sum

so was wondering are there any pointers as to when is the deprecation warning getting triggered as such.

[ehsanmok]

Right! I think completely resolving it would require some manual type conversions. In the case above, it will be

    for offset in range(ITER):
        idx = thread_id + UInt(offset) - 1
        if 0 <= Int(idx) < SIZE:

Because, idx is UInt but SIZE is Int.