Details
-
Bug
-
Status: Resolved
-
Critical
-
Resolution: Fixed
-
6.0.0
-
None
Description
HPCC-15489 introduced a change which caused a large performance degradation when many channels per slave are used.
Time / contention is now spent in:
roxiemem::CChunkedHeap::inlineDoAllocate
CLookupJoinActivityBase<CLookupManyHT>::nextRow()
Attachments
Issue Links
- relates to
-
HPCC-15489 channelsPerSlave>1 configuration with child datasets hit allocator caching issue
-
- Resolved
-