Skip to content

Optimize remainder validation#51

Draft
hkratz wants to merge 74 commits intomainfrom
remainder_optimization
Draft

Optimize remainder validation#51
hkratz wants to merge 74 commits intomainfrom
remainder_optimization

Conversation

@hkratz
Copy link
Contributor

@hkratz hkratz commented May 30, 2021 •
edited
Loading

  • AVX 2
  • SSE 4.2
  • aarch64 Neon

hkratz added 30 commits May 26, 2021 08:18
@hkratz
refactor: check_remainder()
f759a78
@hkratz
bit mor idiomatic
e562d48
@hkratz
WIP: remainder optimization
1edb3b2
@hkratz
aarch64 remainder processing: use lane loads
a7efda1
@hkratz
fix
b3c7636
@hkratz
SimdInput::new_partial() implementation
defd4d2
@hkratz
ascii special-casing
5df7b38
@hkratz
ascii-optimized remainder checking
7f0cf3b
@hkratz
test-bed for different partial load implementations
2cbf8a9
@hkratz
aarch64: select partial load implementation
42f1acd
@hkratz
cleanup
8e2706f
@hkratz
remove unnecessary pub
695c4f8
@hkratz
basic x86 partial load impl.
fa25b46
@hkratz
clippy
3d71822
@hkratz
add missing helper fns
d4d3ccd
@hkratz
clippy
d08c88b
@hkratz
add unit test for AVX 2 load_partial()
0c39149
@hkratz
Implement AVX 2 simd value Display and LowerHex traits for debugging
198e11c
@hkratz
add load_partial_direct() method
57b0790
@hkratz
only run avx2 masked load test if AVX 2 is available
38b58de
@hkratz
fix avx2 detection
b6b2388
@hkratz
prevent remainder loop unrolling in sse 4.2 which caused the methods …
44f8fd1 
…not to be inlined.
@hkratz
only implement Debug/LowerHex for tests
cb7ed19
@hkratz
Merge branch 'remainder_optimization' of github.com:rusticstuff/simdu…
11de67f 
…tf8 into remainder_optimization
@hkratz
expand benchmarks
10df336
@hkratz
x86: make delegation to std for small inputs in-code configurable
bab2fbf
@hkratz
comment
0bb570d
@hkratz
x86: delegate inputs < 9 bytes to std
cbae848
@hkratz
Dsiplay for SimdU8Value
5229492
@hkratz
SSE 4.2 load_partial()
2952db3
hkratz added 30 commits June 18, 2021 07:39
@hkratz
remove stray comment
1b2d9aa
@hkratz
loop unrolling prevention no longer needed
150103b
@hkratz
benchmark for small inputs of a random length in ranges
44cd43c
@hkratz
aarch64 partial load asm impl.
97794b3
@hkratz
Update Cargo.toml
138ea37 
update benchmark Cargo.toml
@hkratz
Merge branch 'remainder_optimization' of github.com:rusticstuff/simdu…
051dcc1 
…tf8 into remainder_optimization
@hkratz
fix/allow lint warnings
b6304c2
@hkratz
example is only for x86-64
7fea546
@hkratz
fix small std benchmark
b995b7f
@hkratz
Update lib.rs
1141636 
fmt
@hkratz
use classic lengths for throughput benchmark again now that we have a…
e1fd53c 
…n extra benchmark for small inputs.
@hkratz
add small_basic benchmark
b4ce94f
@hkratz
ARM64 CI
8babc9f
@hkratz
fix ci
4021faa
@hkratz
Update ci.yml
80b0bc8
@hkratz
fix ci
d29695d
@hkratz
more ARM64 ci
b2c0339
@hkratz
Merge branch 'main' into remainder_optimization
4abbcbe
@hkratz
fix ci
b1416f6
@hkratz
clippy
5753cad
@hkratz
add one more small benchmark
b86d00f
@hkratz
more consistent small basic benchmarks
8eb9ebf
@hkratz
rename fn
47588bc
@hkratz
make cargo test --all-features work on non-x86
119b622
@hkratz
remove aarch64 partial load assembly impl. non-inlined fn
bb58d71
@hkratz
asm fn -> method
45befdb
@hkratz
Merge branch 'main' into remainder_optimization
431e8cc
@hkratz
simplify
37f9198
@hkratz
Trigger GitHub actions
d3399cd
@hkratz
fixes for current nightly
82c5fee
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@hkratz