| name | mojo-simd-optimize |
| description | Apply SIMD optimizations to Mojo code for parallel computation. Use when optimizing performance-critical tensor and array operations. |
| mcp_fallback | none |
| category | mojo |
SIMD Optimization Skill
Parallelize tensor and array operations using SIMD.
When to Use
- Optimizing tensor operations
- Vectorizing element-wise computations
- Performance-critical loops (>1000 elements)
- Benchmark results show optimization potential
Quick Reference
from sys.info import simdwidthof
comptime width = simdwidthof[DType.float32]()
# SIMD vector add
for i in range(0, size, width):
result.store(i, a.load[width](i) + b.load[width](i))
Workflow
- Identify bottleneck - Profile code to find hot loops
- Get SIMD width - Use
simdwidthof[dtype]()
- Vectorize loop - Process
width elements per iteration
- Handle remainder - Process leftover elements
- Benchmark - Verify performance improvement (4x-8x expected)
Mojo-Specific Notes
- SIMD width varies by CPU and dtype (usually 8-16 for float32)
- Always handle remainder elements with scalar loop
- Prefer
alias for compile-time SIMD width constants
- Test on target hardware - SIMD width is platform-specific
Error Handling
| Error |
Cause |
Solution |
Out of bounds |
Remainder not handled |
Add scalar remainder loop |
No speedup |
Wrong SIMD width |
Use simdwidthof[dtype]() |
Compilation fails |
Type mismatch |
Check load/store types match |
Segfault |
Misaligned access |
Ensure stride is correct |
References
.claude/shared/mojo-guidelines.md - SIMD patterns section- Mojo manual: SIMD documentation