[arrow-rs] branch master updated: Add sort boolean benchmark (#457)

[alamb]

This is an automated email from the ASF dual-hosted git repository.

alamb pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/arrow-rs.git


The following commit(s) were added to refs/heads/master by this push:
     new 2d8e299  Add sort boolean benchmark (#457)
2d8e299 is described below

commit 2d8e299e1c02f4b1c76a03b56e386767f165fd01
Author: Andrew Lamb <and...@nerdnetworks.org>
AuthorDate: Sat Jun 19 07:11:16 2021 -0400

    Add sort boolean benchmark (#457)
    
    * Add sort boolean benchmark
    
    * fmt
---
 arrow/benches/sort_kernel.rs | 57 +++++++++++++++++++++++++++++---------------
 1 file changed, 38 insertions(+), 19 deletions(-)

diff --git a/arrow/benches/sort_kernel.rs b/arrow/benches/sort_kernel.rs
index 8467b50..f9f5f24 100644
--- a/arrow/benches/sort_kernel.rs
+++ b/arrow/benches/sort_kernel.rs
@@ -27,12 +27,19 @@ use arrow::compute::kernels::sort::{lexsort, SortColumn};
 use arrow::util::bench_util::*;
 use arrow::{array::*, datatypes::Float32Type};
 
-fn create_array(size: usize, with_nulls: bool) -> ArrayRef {
+fn create_f32_array(size: usize, with_nulls: bool) -> ArrayRef {
     let null_density = if with_nulls { 0.5 } else { 0.0 };
     let array = create_primitive_array::<Float32Type>(size, null_density);
     Arc::new(array)
 }
 
+fn create_bool_array(size: usize, with_nulls: bool) -> ArrayRef {
+    let null_density = if with_nulls { 0.5 } else { 0.0 };
+    let true_density = 0.5;
+    let array = create_boolean_array(size, null_density, true_density);
+    Arc::new(array)
+}
+
 fn bench_sort(array_a: &ArrayRef, array_b: &ArrayRef, limit: Option<usize>) {
     let columns = vec![
         SortColumn {
@@ -49,58 +56,70 @@ fn bench_sort(array_a: &ArrayRef, array_b: &ArrayRef, 
limit: Option<usize>) {
 }
 
 fn add_benchmark(c: &mut Criterion) {
-    let arr_a = create_array(2u64.pow(10) as usize, false);
-    let arr_b = create_array(2u64.pow(10) as usize, false);
+    let arr_a = create_f32_array(2u64.pow(10) as usize, false);
+    let arr_b = create_f32_array(2u64.pow(10) as usize, false);
 
     c.bench_function("sort 2^10", |b| b.iter(|| bench_sort(&arr_a, &arr_b, 
None)));
 
-    let arr_a = create_array(2u64.pow(12) as usize, false);
-    let arr_b = create_array(2u64.pow(12) as usize, false);
+    let arr_a = create_f32_array(2u64.pow(12) as usize, false);
+    let arr_b = create_f32_array(2u64.pow(12) as usize, false);
 
     c.bench_function("sort 2^12", |b| b.iter(|| bench_sort(&arr_a, &arr_b, 
None)));
 
-    let arr_a = create_array(2u64.pow(10) as usize, true);
-    let arr_b = create_array(2u64.pow(10) as usize, true);
+    let arr_a = create_f32_array(2u64.pow(10) as usize, true);
+    let arr_b = create_f32_array(2u64.pow(10) as usize, true);
 
     c.bench_function("sort nulls 2^10", |b| {
         b.iter(|| bench_sort(&arr_a, &arr_b, None))
     });
 
-    let arr_a = create_array(2u64.pow(12) as usize, true);
-    let arr_b = create_array(2u64.pow(12) as usize, true);
+    let arr_a = create_f32_array(2u64.pow(12) as usize, true);
+    let arr_b = create_f32_array(2u64.pow(12) as usize, true);
 
     c.bench_function("sort nulls 2^12", |b| {
         b.iter(|| bench_sort(&arr_a, &arr_b, None))
     });
 
+    let arr_a = create_bool_array(2u64.pow(12) as usize, false);
+    let arr_b = create_bool_array(2u64.pow(12) as usize, false);
+    c.bench_function("bool sort 2^12", |b| {
+        b.iter(|| bench_sort(&arr_a, &arr_b, None))
+    });
+
+    let arr_a = create_bool_array(2u64.pow(12) as usize, true);
+    let arr_b = create_bool_array(2u64.pow(12) as usize, true);
+    c.bench_function("bool sort nulls 2^12", |b| {
+        b.iter(|| bench_sort(&arr_a, &arr_b, None))
+    });
+
     // with limit
     {
-        let arr_a = create_array(2u64.pow(12) as usize, false);
-        let arr_b = create_array(2u64.pow(12) as usize, false);
+        let arr_a = create_f32_array(2u64.pow(12) as usize, false);
+        let arr_b = create_f32_array(2u64.pow(12) as usize, false);
         c.bench_function("sort 2^12 limit 10", |b| {
             b.iter(|| bench_sort(&arr_a, &arr_b, Some(10)))
         });
 
-        let arr_a = create_array(2u64.pow(12) as usize, false);
-        let arr_b = create_array(2u64.pow(12) as usize, false);
+        let arr_a = create_f32_array(2u64.pow(12) as usize, false);
+        let arr_b = create_f32_array(2u64.pow(12) as usize, false);
         c.bench_function("sort 2^12 limit 100", |b| {
             b.iter(|| bench_sort(&arr_a, &arr_b, Some(100)))
         });
 
-        let arr_a = create_array(2u64.pow(12) as usize, false);
-        let arr_b = create_array(2u64.pow(12) as usize, false);
+        let arr_a = create_f32_array(2u64.pow(12) as usize, false);
+        let arr_b = create_f32_array(2u64.pow(12) as usize, false);
         c.bench_function("sort 2^12 limit 1000", |b| {
             b.iter(|| bench_sort(&arr_a, &arr_b, Some(1000)))
         });
 
-        let arr_a = create_array(2u64.pow(12) as usize, false);
-        let arr_b = create_array(2u64.pow(12) as usize, false);
+        let arr_a = create_f32_array(2u64.pow(12) as usize, false);
+        let arr_b = create_f32_array(2u64.pow(12) as usize, false);
         c.bench_function("sort 2^12 limit 2^12", |b| {
             b.iter(|| bench_sort(&arr_a, &arr_b, Some(2u64.pow(12) as usize)))
         });
 
-        let arr_a = create_array(2u64.pow(12) as usize, true);
-        let arr_b = create_array(2u64.pow(12) as usize, true);
+        let arr_a = create_f32_array(2u64.pow(12) as usize, true);
+        let arr_b = create_f32_array(2u64.pow(12) as usize, true);
 
         c.bench_function("sort nulls 2^12 limit 10", |b| {
             b.iter(|| bench_sort(&arr_a, &arr_b, Some(10)))