diff --git a/.github/workflows/lint.yml b/.github/workflows/lint.yml
index 0335049..d64a429 100644
--- a/.github/workflows/lint.yml
+++ b/.github/workflows/lint.yml
@@ -9,7 +9,7 @@ jobs:
   lint:
     strategy:
       matrix:
-        go-version: [1.21.x]
+        go-version: [1.18.x]
         platform: [ubuntu-latest]
 
     runs-on: ${{ matrix.platform }}
diff --git a/.github/workflows/test.yml b/.github/workflows/test.yml
index 07321c5..59f3b8f 100644
--- a/.github/workflows/test.yml
+++ b/.github/workflows/test.yml
@@ -9,7 +9,7 @@ jobs:
   test:
     strategy:
       matrix:
-        go-version: [ 1.21.x ]
+        go-version: [ 1.18.x ]
         platform: [ ubuntu-latest ]
 
     runs-on: ${{ matrix.platform }}
diff --git a/internal/hashtable/map.go b/internal/hashtable/map.go
index 2ee48a2..eb073a6 100644
--- a/internal/hashtable/map.go
+++ b/internal/hashtable/map.go
@@ -69,7 +69,7 @@ type Map[K comparable, V any] struct {
 	// used to wake up resize waiters (concurrent modifications)
 	resizeCond sync.Cond
 	// resize in progress flag; updated atomically
-	resizing atomic.Int64
+	resizing int64
 }
 
 type table[K comparable] struct {
@@ -400,7 +400,7 @@ func (m *Map[K, V]) resize(known *table[K], hint resizeHint) {
 		}
 	}
 	// slow path.
-	if !m.resizing.CompareAndSwap(0, 1) {
+	if !atomic.CompareAndSwapInt64(&m.resizing, 0, 1) {
 		// someone else started resize. Wait for it to finish.
 		m.waitForResize()
 		return
@@ -420,7 +420,7 @@ func (m *Map[K, V]) resize(known *table[K], hint resizeHint) {
 		} else {
 			// no need to shrink, wake up all waiters and give up.
 			m.resizeMutex.Lock()
-			m.resizing.Store(0)
+			atomic.StoreInt64(&m.resizing, 0)
 			m.resizeCond.Broadcast()
 			m.resizeMutex.Unlock()
 			return
@@ -440,7 +440,7 @@ func (m *Map[K, V]) resize(known *table[K], hint resizeHint) {
 	// publish the new table and wake up all waiters.
 	atomic.StorePointer(&m.table, unsafe.Pointer(nt))
 	m.resizeMutex.Lock()
-	m.resizing.Store(0)
+	atomic.StoreInt64(&m.resizing, 0)
 	m.resizeCond.Broadcast()
 	m.resizeMutex.Unlock()
 }
@@ -473,7 +473,7 @@ func (m *Map[K, V]) newerTableExists(table *table[K]) bool {
 }
 
 func (m *Map[K, V]) resizeInProgress() bool {
-	return m.resizing.Load() == 1
+	return atomic.LoadInt64(&m.resizing) == 1
 }
 
 func (m *Map[K, V]) waitForResize() {
diff --git a/internal/lossy/buffer.go b/internal/lossy/buffer.go
index 6bb35a6..179a759 100644
--- a/internal/lossy/buffer.go
+++ b/internal/lossy/buffer.go
@@ -48,10 +48,10 @@ type PolicyBuffers[K comparable, V any] struct {
 //
 // This implementation is striped to further increase concurrency.
 type Buffer[K comparable, V any] struct {
-	head                 atomic.Uint64
-	headPadding          [xruntime.CacheLineSize - unsafe.Sizeof(atomic.Uint64{})]byte
-	tail                 atomic.Uint64
-	tailPadding          [xruntime.CacheLineSize - unsafe.Sizeof(atomic.Uint64{})]byte
+	head                 uint64
+	headPadding          [xruntime.CacheLineSize - 8]byte
+	tail                 uint64
+	tailPadding          [xruntime.CacheLineSize - 8]byte
 	nodeManager          *node.Manager[K, V]
 	returned             unsafe.Pointer
 	returnedPadding      [xruntime.CacheLineSize - 2*8]byte
@@ -77,14 +77,14 @@ func New[K comparable, V any](nodeManager *node.Manager[K, V]) *Buffer[K, V] {
 //
 // item may be lost due to contention.
 func (b *Buffer[K, V]) Add(n node.Node[K, V]) *PolicyBuffers[K, V] {
-	head := b.head.Load()
-	tail := b.tail.Load()
+	head := atomic.LoadUint64(&b.head)
+	tail := atomic.LoadUint64(&b.tail)
 	size := tail - head
 	if size >= capacity {
 		// full buffer
 		return nil
 	}
-	if b.tail.CompareAndSwap(tail, tail+1) {
+	if atomic.CompareAndSwapUint64(&b.tail, tail, tail+1) {
 		// success
 		index := int(tail & mask)
 		atomic.StorePointer(&b.buffer[index], n.AsPointer())
@@ -108,7 +108,7 @@ func (b *Buffer[K, V]) Add(n node.Node[K, V]) *PolicyBuffers[K, V] {
 				head++
 			}
 
-			b.head.Store(head)
+			atomic.StoreUint64(&b.head, head)
 			return pb
 		}
 	}
@@ -136,6 +136,6 @@ func (b *Buffer[K, V]) Clear() {
 		atomic.StorePointer(&b.buffer[i], nil)
 	}
 	b.Free()
-	b.tail.Store(0)
-	b.head.Store(0)
+	atomic.StoreUint64(&b.tail, 0)
+	atomic.StoreUint64(&b.head, 0)
 }
diff --git a/internal/queue/mpsc.go b/internal/queue/mpsc.go
index 3f36316..a537aea 100644
--- a/internal/queue/mpsc.go
+++ b/internal/queue/mpsc.go
@@ -12,7 +12,6 @@ package queue
 import (
 	"runtime"
 	"sync/atomic"
-	"unsafe"
 
 	"github.com/maypok86/otter/internal/xruntime"
 )
@@ -36,18 +35,18 @@ func zeroValue[T any]() T {
 type MPSC[T any] struct {
 	capacity     uint64
 	sleep        chan struct{}
-	head         atomic.Uint64
-	headPadding  [xruntime.CacheLineSize - unsafe.Sizeof(atomic.Uint64{})]byte
+	head         uint64
+	headPadding  [xruntime.CacheLineSize - 8]byte
 	tail         uint64
 	tailPadding  [xruntime.CacheLineSize - 8]byte
-	isSleep      atomic.Uint64
-	sleepPadding [xruntime.CacheLineSize - unsafe.Sizeof(atomic.Uint64{})]byte
+	isSleep      uint64
+	sleepPadding [xruntime.CacheLineSize - 8]byte
 	slots        []slot[T]
 }
 
 type slot[T any] struct {
-	// atomic.Uint64 is used here to get proper 8 byte alignment on 32-bit archs.
-	turn atomic.Uint64
+	// uint64 is used here to get proper 8 byte alignment on 32-bit archs.
+	turn uint64
 	item T
 }
 
@@ -63,13 +62,13 @@ func NewMPSC[T any](capacity int) *MPSC[T] {
 // Insert inserts the given item into the queue.
 // Blocks, if the queue is full.
 func (q *MPSC[T]) Insert(item T) {
-	head := q.head.Add(1) - 1
+	head := atomic.AddUint64(&q.head, 1) - 1
 	q.wakeUpConsumer()
 
 	slot := &q.slots[q.idx(head)]
 	turn := q.turn(head) * 2
 	retries := 0
-	for slot.turn.Load() != turn {
+	for atomic.LoadUint64(&slot.turn) != turn {
 		if retries == maxRetries {
 			q.wakeUpConsumer()
 			retries = 0
@@ -80,7 +79,7 @@ func (q *MPSC[T]) Insert(item T) {
 	}
 
 	slot.item = item
-	slot.turn.Store(turn + 1)
+	atomic.StoreUint64(&slot.turn, turn+1)
 }
 
 // Remove retrieves and removes the item from the head of the queue.
@@ -90,7 +89,7 @@ func (q *MPSC[T]) Remove() T {
 	slot := &q.slots[q.idx(tail)]
 	turn := 2*q.turn(tail) + 1
 	retries := 0
-	for slot.turn.Load() != turn {
+	for atomic.LoadUint64(&slot.turn) != turn {
 		if retries == maxRetries {
 			q.sleepConsumer()
 			retries = 0
@@ -101,7 +100,7 @@ func (q *MPSC[T]) Remove() T {
 	}
 	item := slot.item
 	slot.item = zeroValue[T]()
-	slot.turn.Store(turn + 1)
+	atomic.StoreUint64(&slot.turn, turn+1)
 	q.tail++
 	return item
 }
@@ -119,7 +118,7 @@ func (q *MPSC[T]) Capacity() int {
 }
 
 func (q *MPSC[T]) wakeUpConsumer() {
-	if q.isSleep.Load() == 1 && q.isSleep.CompareAndSwap(1, 0) {
+	if atomic.LoadUint64(&q.isSleep) == 1 && atomic.CompareAndSwapUint64(&q.isSleep, 1, 0) {
 		// if the consumer is asleep, we'll wake him up.
 		q.sleep <- struct{}{}
 	}
@@ -127,12 +126,12 @@ func (q *MPSC[T]) wakeUpConsumer() {
 
 func (q *MPSC[T]) sleepConsumer() {
 	// if the queue's been empty for too long, we fall asleep.
-	q.isSleep.Store(1)
+	atomic.StoreUint64(&q.isSleep, 1)
 	<-q.sleep
 }
 
 func (q *MPSC[T]) isEmpty() bool {
-	return q.tail == q.head.Load()
+	return q.tail == atomic.LoadUint64(&q.head)
 }
 
 func (q *MPSC[T]) idx(i uint64) uint64 {