Incorrect OSPF Auto Cost Reference Bandwidth Calculations

Hi All,

We tend to configure the OSPF auto cost reference bandwidth to 100Gb, which means that 1Gb interfaces should have a cost of 100, on Juniper MX:

> show ospf interface ge-1/0/0.904 detail 
Interface State Area DR ID BDR ID Nbrs
ge-1/0/0.904 DR x.x.x.x x.x.x.x x.x.x.x 1
Type: LAN, Address: x.x.x.x, Mask: 255.255.255.254, MTU: 1500, Cost: 100

However, on Edgemax side the calculations seem to be incorrect (giving a cost of 10000, out by a factor of 100):

# show interfaces ethernet eth0 vif 904 ip ospf 
 dead-interval 3
 hello-interval 1

# show protocols ospf auto-cost reference-bandwidth 
 reference-bandwidth 100000

# run show ip ospf interface eth0.904
eth0.904 is up
  ifindex 34, MTU 1500 bytes, BW 0 Kbit <UP,BROADCAST,RUNNING,MULTICAST>
  Internet Address x.x.x.x/31, Area x.x.x.x [Stub]
  MTU mismatch detection:enabled
  Router ID x.x.x.x, Network Type BROADCAST, Cost: 10000

I can't see any reference to this issue elsewhere in the forum, the only references I can see seem to infer the correct and intended calulcation.

Is this the intended behaviour?