There is already an open issue about that on GoogleCloudPlatform's github issue tracker. The issue has been created in late Feb.

I suggest adding "+1" there by leaving a comment and monitoring the progress.

I can't confirm that isn't working using Container-Optimized OS (cos) 1.14.10-gke.27 (master version)

kind: Deployment
apiVersion: apps/v1
metadata:
  labels:
    app: test1
  name: test1
spec:
  replicas: 1
  selector:
    matchLabels:
      app: test1
  template:
    metadata:
      labels:
        app: test1
    spec:
      volumes:
      - name: demo
        emptyDir:
          medium: Memory
          sizeLimit: "1Gi"
      containers:
      - name: busy
        image: busybox
        command: ["/bin/sh"]
        args: ["-c", "sleep 3000"]
        volumeMounts:
        - name: demo
          mountPath: /dev/shm

It's 1,8Gi ~= 0.5 the underlying node memory size

Filesystem                Size      Used Available Use% Mounted on
overlay                  94.3G      2.9G     91.4G   3% /
tmpfs                     1.8G         0      1.8G   0% /dev/shm

Please note

the sizeLimit parameter set for emptyDir works differently (in this release at the moment) - eviction manager keeps monitoring the disk space used by pod emptyDir volume and it will evict pods when the usage exceeds the limit.

   medium       <string>
     What type of storage medium should back this directory. The default is ""
     which means to use the node's default medium. Must be an empty string
     (default) or Memory

   sizeLimit    <string>
     Total amount of local storage required for this EmptyDir volume. The size
     limit is also applicable for memory medium. 

     The maximum usage on memory medium EmptyDir would be the minimum value
     between the SizeLimit specifiedhere and the sum of memory limits of all
     containers in a pod. The default is nil which means that the limit is
     undefined

Hope this help.