mbutrovich commented on PR #3368:
URL:
https://github.com/apache/datafusion-comet/pull/3368#issuecomment-3837925203
I asked Claude to summarize the net benefit of this PR based on how rustc
chooses targets when not given a target-cpu argument:
### Target CPU comparison
| Architecture | Before (default) | After (this PR) |
|--------------|------------------|-----------------|
| x86_64 Linux | `x86-64` | `x86-64-v3` |
| aarch64 Linux | `generic` | `neoverse-n1` |
### Key features gained
| Feature | x86-64 (before) | x86-64-v3 (after) | Benefit |
|---------|-----------------|-------------------|---------|
| AVX | ✗ | ✓ | Wider SIMD |
| AVX2 | ✗ | ✓ | 256-bit integer SIMD |
| FMA | ✗ | ✓ | Fused multiply-add |
| BMI1/BMI2 | ✗ | ✓ | Bit manipulation |
| Feature | generic (before) | neoverse-n1 (after) | Benefit |
|---------|------------------|---------------------|---------|
| LSE atomics | ✗ | ✓ | Single-instruction atomics, no LL/SC loops |
| AES/SHA2 | ✗ | ✓ | Hardware crypto |
| CRC | ✗ | ✓ | Hardware CRC32 |
| dotprod | ✗ | ✓ | Dot product acceleration |
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
