Merge pull request #62 from lemire/master

Adding bulk iteration.
bits-and-blooms · Apr 26, 2018 · 8ce1146 · 8ce1146
2 parents 1a37ad9 + 7f1803e
commit 8ce1146
Show file tree

Hide file tree

Showing 3 changed files with 98 additions and 2 deletions.
diff --git a/Makefile b/Makefile
@@ -168,7 +168,7 @@ docs:
 	@echo '<html><head><meta http-equiv="refresh" content="0;./127.0.0.1:6060/pkg/'${CVSPATH}'/'${PROJECT}'/index.html"/></head><a href="./127.0.0.1:6060/pkg/'${CVSPATH}'/'${PROJECT}'/index.html">'${PKGNAME}' Documentation ...</a></html>' > target/docs/index.html
 
 # Alias to run all quality-assurance checks
-qa: fmtcheck test vet lint coverage cyclo ineffassign misspell structcheck varcheck errcheck gosimple astscan
+qa: fmtcheck test vet lint coverage cyclo ineffassign misspell structcheck varcheck errcheck gosimple 
 
 # --- INSTALL ---
 

diff --git a/bitset.go b/bitset.go
@@ -230,6 +230,56 @@ func (b *BitSet) NextSet(i uint) (uint, bool) {
 	return 0, false
 }
 
+// NextSetMany returns many next bit sets from the specified index,
+// including possibly the current index and up to cap(buffer).
+// If the returned slice has len zero, then no more set bits were found
+//
+//    buffer := make([]uint, 256)
+//    j := uint(0)
+//    j, buffer = bitmap.NextSetMany(j, buffer)
+//    for ; len(buffer) > 0; j, buffer = bitmap.NextSetMany(j,buffer) {
+//     for k := range buffer {
+//      do something with buffer[k]
+//     }
+//     j += 1
+//    }
+//
+func (b *BitSet) NextSetMany(i uint, buffer []uint) (uint, []uint) {
+	myanswer := buffer[:0]
+
+	x := int(i >> log2WordSize)
+	if x >= len(b.set) {
+		return 0, myanswer
+	}
+	w := b.set[x]
+	w = w >> (i & (wordSize - 1))
+	base := uint(x << 6)
+	capacity := cap(buffer)
+	for len(myanswer) < capacity {
+		for w != 0 {
+			t := w & ((^w) + 1)
+			r := trailingZeroes64(w)
+			myanswer = append(myanswer, r+base)
+			if len(myanswer) == capacity {
+				goto End
+			}
+			w = w ^ t
+		}
+		x += 1
+		if x == len(b.set) {
+			break
+		}
+		base += 64
+		w = b.set[x]
+	}
+End:
+	if len(myanswer) > 0 {
+		return myanswer[len(myanswer)-1], myanswer
+	} else {
+		return 0, myanswer
+	}
+}
+
 // NextClear returns the next clear bit from the specified index,
 // including possibly the current index
 // along with an error code (true = valid, false = no bit found i.e. all bits are set)

diff --git a/bitset_benchmark_test.go b/bitset_benchmark_test.go
@@ -126,11 +126,57 @@ func BenchmarkLemireIterate(b *testing.B) {
 	b.ResetTimer()
 	sum := uint(0)
 	for i := 0; i < b.N; i++ {
-		for i, e := bitmap.NextSet(0); e; i, e = bitmap.NextSet(i + 1) {
+		for j, e := bitmap.NextSet(0); e; j, e = bitmap.NextSet(j + 1) {
 			sum++
 		}
 	}
 	if sum == 0 { // added just to fool ineffassign
 		return
 	}
 }
+
+// go test -bench=LemireIterateb
+// see http://lemire.me/blog/2016/09/22/swift-versus-java-the-bitset-performance-test/
+func BenchmarkLemireIterateb(b *testing.B) {
+	bitmap := New(100000000)
+	for v := uint(0); v <= 100000000; v += 100 {
+		bitmap.Set(v)
+	}
+	b.ResetTimer()
+	sum := uint(0)
+	for i := 0; i < b.N; i++ {
+		for j, e := bitmap.NextSet(0); e; j, e = bitmap.NextSet(j + 1) {
+			sum += j
+		}
+	}
+
+	if sum == 0 { // added just to fool ineffassign
+		return
+	}
+}
+
+// go test -bench=BenchmarkLemireIterateManyb
+// see http://lemire.me/blog/2016/09/22/swift-versus-java-the-bitset-performance-test/
+func BenchmarkLemireIterateManyb(b *testing.B) {
+	bitmap := New(100000000)
+	for v := uint(0); v <= 100000000; v += 100 {
+		bitmap.Set(v)
+	}
+	b.ResetTimer()
+	sum := uint(0)
+	buffer := make([]uint, 256)
+	for i := 0; i < b.N; i++ {
+		j := uint(0)
+		j, buffer = bitmap.NextSetMany(j, buffer)
+		for ; len(buffer) > 0; j, buffer = bitmap.NextSetMany(j, buffer) {
+			for k := range buffer {
+				sum += buffer[k]
+			}
+			j += 1
+		}
+	}
+
+	if sum == 0 { // added just to fool ineffassign
+		return
+	}
+}