Skip to content
Open
Changes from all commits
Commits
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
2 changes: 1 addition & 1 deletion content/english/hpc/simd/masking.md
Original file line number Diff line number Diff line change
Expand Up @@ -170,7 +170,7 @@ vptest ymm0, ymm0
je loop
```

This doesn't improve performance much because both both `vptest` and `vmovmskps` have a throughput of one and will bottleneck the computation regardless of anything else we do in the loop.
This doesn't improve performance much because both `vptest` and `vmovmskps` have a throughput of one and will bottleneck the computation regardless of anything else we do in the loop.

To work around this limitation, we can iterate in blocks of 16 elements and combine the results of independent comparisons of two 256-bit AVX2 registers using a bitwise `or`:

Expand Down