add find all function

Author: thomasjungblut

recordio/simd/magic_number_search.go

@@ -19,3 +19,32 @@ func FindMagicNumber(data []byte, off int) int {
 	}
 	return -1
 }
+
+// FindAllMagicNumbers finds all occurrences of the magic number pattern in the data,
+// starting from the given offset. Returns a slice of all offsets where the pattern was found.
+func FindAllMagicNumbers(data []byte, off int) []int {
+	if len(data) < 3 {
+		return nil
+	}
+	if off >= len(data) || off < 0 {
+		return nil
+	}
+
+	var results []int
+	pos := off
+
+	for {
+		next := FindMagicNumber(data, pos)
+		if next < 0 {
+			break
+		}
+		results = append(results, next)
+		// Start searching from the next position after this match
+		pos = next + 1
+		if pos >= len(data)-2 {
+			break
+		}
+	}
+
+	return results
+}

recordio/simd/magic_number_search_bench_test.go

@@ -42,3 +42,42 @@ func BenchmarkFindMagicNumber(b *testing.B) {
 		})
 	}
 }
+
+func BenchmarkFindAllMagicNumbers(b *testing.B) {
+	// Create 1GB buffer with magic numbers every 50KB
+	const dataSize = 1024 * 1024 * 1024 // 1GB
+	const markerInterval = 50 * 1024    // 50KB
+
+	data := make([]byte, dataSize)
+	pattern := []byte{145, 141, 76}
+
+	// Place magic numbers every 50KB
+	for i := 0; i < len(data)-3; i += markerInterval {
+		data[i] = pattern[0]
+		data[i+1] = pattern[1]
+		data[i+2] = pattern[2]
+	}
+
+	// Place one at the end to ensure we find something
+	if len(data) >= 3 {
+		data[len(data)-3] = pattern[0]
+		data[len(data)-2] = pattern[1]
+		data[len(data)-1] = pattern[2]
+	}
+
+	// Calculate expected number of matches
+	// Markers placed at: 0, 50KB, 100KB, ..., n*50KB where n*50KB < dataSize-3
+	// Plus one at the end (dataSize-3)
+	expectedMatches := (dataSize-3)/markerInterval + 2 // floor division + loop markers + end marker
+
+	b.ResetTimer()
+	b.ReportAllocs()
+	b.SetBytes(dataSize) // Report bandwidth in GB/s
+
+	for i := 0; i < b.N; i++ {
+		results := FindAllMagicNumbers(data, 0)
+		if len(results) != expectedMatches {
+			b.Fatalf("expected %d matches, got %d", expectedMatches, len(results))
+		}
+	}
+}

recordio/simd/magic_number_search_cgo.go

@@ -91,3 +91,32 @@ func FindMagicNumber(data []byte, off int) int {
 
 	return cgo_find_magic_numbers_scalar(data, off)
 }
+
+// FindAllMagicNumbers finds all occurrences of the magic number pattern in the data,
+// starting from the given offset. Returns a slice of all offsets where the pattern was found.
+func FindAllMagicNumbers(data []byte, off int) []int {
+	if len(data) < 3 {
+		return nil
+	}
+	if off >= len(data) || off < 0 {
+		return nil
+	}
+
+	var results []int
+	pos := off
+
+	for {
+		next := FindMagicNumber(data, pos)
+		if next < 0 {
+			break
+		}
+		results = append(results, next)
+		// Start searching from the next position after this match
+		pos = next + 1
+		if pos >= len(data)-2 {
+			break
+		}
+	}
+
+	return results
+}