Skip to content

Specify single-block limitation on puzzle 11 pattern #243

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
dunnoconnor merged 1 commit into from
Apr 21, 2026
Merged

Specify single-block limitation on puzzle 11 pattern #243

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 19 additions & 0 deletions book/src/puzzle_11/puzzle_11.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -165,6 +165,25 @@ The solution implements a sliding window sum using TileTensor with these key ste
window_sum = shared[i-2] + shared[i-1] + shared[i]
```

> **Single-block assumption:** This solution is correct because the puzzle is configured with
> `BLOCKS_PER_GRID = (1, 1)` and `SIZE == TPB = 8`, guaranteeing every thread belongs to the
> same block so `global_i == local_i`. Under this constraint, `local_i >= 2` whenever
> `global_i > 1`, so `shared[local_i - 2]` and `shared[local_i - 1]` are always valid.
>
> In a **multi-block** kernel the first two threads of each block beyond block 0 would have
> `local_i = 0` or `local_i = 1` while `global_i > 1`, causing out-of-bounds shared memory
> reads. The robust pattern for multi-block pooling guards with `local_i` and falls back to
> global reads for the halo elements:
>
> ```mojo
> if local_i >= 2:
> output[global_i] = shared[local_i-2] + shared[local_i-1] + shared[local_i]
> elif local_i == 1 and global_i >= 2:
> output[global_i] = a[global_i-2] + shared[0] + shared[1]
> elif local_i == 0 and global_i >= 2:
> output[global_i] = a[global_i-2] + a[global_i-1] + shared[0]
> ```

4. **Memory access pattern**
- One global read per thread into shared tensor
- Efficient neighbor access through shared memory
Expand Down
Loading