Add exemplar reservoir parallel benchmarks (#7441)

This also fixes a bug introduced in
https://github.com/open-telemetry/opentelemetry-go/pull/7423, where we
were only locking around storage and not around other shared fields
(e.g. count). Fixing the bug is required for benchmarks to run properly,
but wasn't caught by concurrent safe tests because the SDK does not
currently call exemplar methods concurrently.

```
goos: linux
goarch: amd64
pkg: go.opentelemetry.io/otel/sdk/metric/exemplar
cpu: Intel(R) Xeon(R) CPU @ 2.20GHz
BenchmarkFixedSizeReservoirOffer-24    	  498955	       248.4 ns/op	       0 B/op	       0 allocs/op
BenchmarkHistogramReservoirOffer-24    	  478068	       250.1 ns/op	       0 B/op	       0 allocs/op
```

sdk/metric/exemplar/benchmark_test.go

@@ -0,0 +1,50 @@
+// Copyright The OpenTelemetry Authors
+// SPDX-License-Identifier: Apache-2.0
+
+package exemplar // import "go.opentelemetry.io/otel/sdk/metric/exemplar"
+
+import (
+	"runtime"
+	"testing"
+	"time"
+)
+
+func BenchmarkFixedSizeReservoirOffer(b *testing.B) {
+	ts := time.Now()
+	val := NewValue[int64](25)
+	ctx := b.Context()
+	reservoir := NewFixedSizeReservoir(runtime.NumCPU())
+	b.RunParallel(func(pb *testing.PB) {
+		i := 0
+		for pb.Next() {
+			reservoir.Offer(ctx, ts, val, nil)
+			// Periodically trigger a reset, because the algorithm for fixed-size
+			// reservoirs records exemplars very infrequently after a large
+			// number of collect calls.
+			if i%100 == 99 {
+				reservoir.mu.Lock()
+				reservoir.reset()
+				reservoir.mu.Unlock()
+			}
+			i++
+		}
+	})
+}
+
+func BenchmarkHistogramReservoirOffer(b *testing.B) {
+	ts := time.Now()
+	ctx := b.Context()
+	buckets := []float64{0, 5, 10, 25, 50, 75, 100, 250, 500, 750, 1000, 2500, 5000, 7500, 10000}
+	values := make([]Value, len(buckets))
+	for i, bucket := range buckets {
+		values[i] = NewValue[float64](bucket + 1)
+	}
+	res := NewHistogramReservoir(buckets)
+	b.RunParallel(func(pb *testing.PB) {
+		i := 0
+		for pb.Next() {
+			res.Offer(ctx, ts, values[i%len(values)], nil)
+			i++
+		}
+	})
+}

sdk/metric/exemplar/fixed_size_reservoir.go

@@ -125,6 +125,8 @@ func (r *FixedSizeReservoir) Offer(ctx context.Context, t time.Time, n Value, a
 	// https://github.com/MrAlias/reservoir-sampling for a performance
 	// comparison of reservoir sampling algorithms.
 
+	r.mu.Lock()
+	defer r.mu.Unlock()
 	if int(r.count) < cap(r.measurements) {
 		r.store(int(r.count), newMeasurement(ctx, t, n, a))
 	} else if r.count == r.next {

sdk/metric/exemplar/histogram_reservoir.go

@@ -68,5 +68,11 @@ func (r *HistogramReservoir) Offer(ctx context.Context, t time.Time, v Value, a
 	default:
 		panic("unknown value type")
 	}
-	r.store(sort.SearchFloat64s(r.bounds, n), newMeasurement(ctx, t, v, a))
+
+	idx := sort.SearchFloat64s(r.bounds, n)
+	m := newMeasurement(ctx, t, v, a)
+
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	r.store(idx, m)
 }

sdk/metric/exemplar/storage.go

@@ -27,8 +27,6 @@ func newStorage(n int) *storage {
 }
 
 func (r *storage) store(idx int, m measurement) {
-	r.mu.Lock()
-	defer r.mu.Unlock()
 	r.measurements[idx] = m
 }