jhorstmann commented on code in PR #4701:
URL: https://github.com/apache/arrow-rs/pull/4701#discussion_r1296252983
##########
arrow-ord/Cargo.toml:
##########
@@ -44,10 +44,3 @@ half = { version = "2.1", default-features = false, features
= ["num-traits"] }
[dev-dependencies]
rand = { version = "0.8", default-features = false, features = ["std",
"std_rng"] }
-
-[package.metadata.docs.rs]
-features = ["dyn_cmp_dict"]
-
-[features]
-dyn_cmp_dict = []
-simd = ["arrow-array/simd"]
Review Comment:
correct, though I feel like llvm has gotten a little better at this. on the
other hand, the simd kernels also did not generate "perfect" code due to
combining multiple masks into a single u64 (except for u8 comparisons which
have 64 lanes).
I think the tradeoff against improved compile time is ok, and longer term
this allows more code cleanups. The decimal types for example already had a
quite hacky support for simd. The last simd usages then would be the
aggregation kernels.
--
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]
