introduced new collector store impl based on slice
counter-intuitively the list impl was faster than the heap the theory was the heap did more comparisons and swapping so even though it benefited from no interface and some cache locality, it was still slower the idea was to just use a raw slice kept in order this avoids the need for interface, but can take same comparison approach as the list it seems to work out: go test -run=xxx -bench=. -benchmem -cpuprofile=cpu.out BenchmarkTop10of100000Scores-4 5000 299959 ns/op 2600 B/op 36 allocs/op BenchmarkTop100of100000Scores-4 2000 601104 ns/op 20720 B/op 216 allocs/op BenchmarkTop10of1000000Scores-4 500 3450196 ns/op 2616 B/op 36 allocs/op BenchmarkTop100of1000000Scores-4 500 3874276 ns/op 20856 B/op 216 allocs/op PASS ok github.com/blevesearch/bleve/search/collectors 7.440s
This commit is contained in:
parent
47c239ca7b
commit
c9310b906d
63
search/collectors/slice.go
Normal file
63
search/collectors/slice.go
Normal file
|
@ -0,0 +1,63 @@
|
||||||
|
// Copyright (c) 2014 Couchbase, Inc.
|
||||||
|
// Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file
|
||||||
|
// except in compliance with the License. You may obtain a copy of the License at
|
||||||
|
// http://www.apache.org/licenses/LICENSE-2.0
|
||||||
|
// Unless required by applicable law or agreed to in writing, software distributed under the
|
||||||
|
// License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND,
|
||||||
|
// either express or implied. See the License for the specific language governing permissions
|
||||||
|
// and limitations under the License.
|
||||||
|
|
||||||
|
package collectors
|
||||||
|
|
||||||
|
import "github.com/blevesearch/bleve/search"
|
||||||
|
|
||||||
|
type collectStoreSlice struct {
|
||||||
|
slice search.DocumentMatchCollection
|
||||||
|
compare collectorCompare
|
||||||
|
}
|
||||||
|
|
||||||
|
func newStoreSlice(cap int, compare collectorCompare) *collectStoreSlice {
|
||||||
|
rv := &collectStoreSlice{
|
||||||
|
slice: make(search.DocumentMatchCollection, 0, cap),
|
||||||
|
compare: compare,
|
||||||
|
}
|
||||||
|
return rv
|
||||||
|
}
|
||||||
|
|
||||||
|
func (c *collectStoreSlice) Add(doc *search.DocumentMatch) {
|
||||||
|
// find where to insert, starting at end (lowest)
|
||||||
|
i := len(c.slice)
|
||||||
|
for ; i > 0; i-- {
|
||||||
|
cmp := c.compare(doc, c.slice[i-1])
|
||||||
|
if cmp >= 0 {
|
||||||
|
break
|
||||||
|
}
|
||||||
|
}
|
||||||
|
if i < 0 {
|
||||||
|
i = 0
|
||||||
|
}
|
||||||
|
// insert at i
|
||||||
|
c.slice = append(c.slice, nil)
|
||||||
|
copy(c.slice[i+1:], c.slice[i:])
|
||||||
|
c.slice[i] = doc
|
||||||
|
}
|
||||||
|
|
||||||
|
func (c *collectStoreSlice) RemoveLast() *search.DocumentMatch {
|
||||||
|
var rv *search.DocumentMatch
|
||||||
|
rv, c.slice = c.slice[len(c.slice)-1], c.slice[:len(c.slice)-1]
|
||||||
|
return rv
|
||||||
|
}
|
||||||
|
|
||||||
|
func (c *collectStoreSlice) Final(skip int, fixup collectorFixup) (search.DocumentMatchCollection, error) {
|
||||||
|
for i := skip; i < len(c.slice); i++ {
|
||||||
|
err := fixup(c.slice[i])
|
||||||
|
if err != nil {
|
||||||
|
return nil, err
|
||||||
|
}
|
||||||
|
}
|
||||||
|
return c.slice[skip:], nil
|
||||||
|
}
|
||||||
|
|
||||||
|
func (c *collectStoreSlice) Len() int {
|
||||||
|
return len(c.slice)
|
||||||
|
}
|
|
@ -32,7 +32,7 @@ type TopNCollector struct {
|
||||||
results search.DocumentMatchCollection
|
results search.DocumentMatchCollection
|
||||||
facetsBuilder *search.FacetsBuilder
|
facetsBuilder *search.FacetsBuilder
|
||||||
|
|
||||||
store *collectStoreList
|
store *collectStoreSlice
|
||||||
|
|
||||||
needDocIds bool
|
needDocIds bool
|
||||||
neededFields []string
|
neededFields []string
|
||||||
|
@ -52,7 +52,7 @@ func NewTopNCollector(size int, skip int, sort search.SortOrder) *TopNCollector
|
||||||
hc := &TopNCollector{size: size, skip: skip, sort: sort}
|
hc := &TopNCollector{size: size, skip: skip, sort: sort}
|
||||||
// pre-allocate space on the heap, we need size+skip results
|
// pre-allocate space on the heap, we need size+skip results
|
||||||
// +1 additional while figuring out which to evict
|
// +1 additional while figuring out which to evict
|
||||||
hc.store = newStoreList(size+skip+1, func(i, j *search.DocumentMatch) int {
|
hc.store = newStoreSlice(size+skip+1, func(i, j *search.DocumentMatch) int {
|
||||||
return hc.sort.Compare(hc.cachedScoring, hc.cachedDesc, i, j)
|
return hc.sort.Compare(hc.cachedScoring, hc.cachedDesc, i, j)
|
||||||
})
|
})
|
||||||
|
|
||||||
|
|
Loading…
Reference in New Issue
Block a user