Concurrent Full Text CONTAINSTABLE queries deadlocking on each other by J. Granden

We're experiencing 10-20 deadlocks per hour bewteen multiple instances of this query:

"SELECT i.[key] FROM    CONTAINSTABLE(items, SearchTerms, @IncludeContainsCondition)"

The problematic resource is a Sch-M lock on a "COMPRESSED_FRAGMENT"

HW is 4 node numa with 80 procs.

(full deadlock graph will be attached shortly)
Resource list from deadlock graph:

<resource-list>
<metadatalock subresource="COMPRESSED_FRAGMENT" classid="object_id = 2079398527, fragment_id = 10858772" dbid="15" id="lock456efcd180" mode="Sch-S">
<owner-list>
    <owner id="process723196b0c8" mode="Sch-S" />
</owner-list>
<waiter-list>
    <waiter id="process6a91518cf8" mode="Sch-M" requestType="wait" />
</waiter-list>
</metadatalock>
<metadatalock subresource="COMPRESSED_FRAGMENT" classid="object_id = 2079398527, fragment_id = 9903010" dbid="15" id="lock539185f080" mode="Sch-M">
<owner-list>
    <owner id="process95a4cf8" mode="Sch-S" requestType="wait" />
</owner-list>
<waiter-list>
    <waiter id="process723196b0c8" mode="Sch-S" requestType="wait" />
</waiter-list>
</metadatalock>
<metadatalock subresource="COMPRESSED_FRAGMENT" classid="object_id = 2079398527, fragment_id = 9903010" dbid="15" id="lock539185f080" mode="Sch-M">
<owner-list>
    <owner id="process6a91518cf8" mode="Sch-M" />
</owner-list>
<waiter-list>
    <waiter id="process95a4cf8" mode="Sch-S" requestType="wait" />
</waiter-list>
</metadatalock>
</resource-list>