[
https://issues.apache.org/jira/browse/HBASE-9969?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Matt Corgan updated HBASE-9969:
-------------------------------
Attachment: hbase-9969-pq-v1.patch
attaching hbase-9969-pq-v1.patch
* adds KeyValueScannerPriorityQueue, which is a stripped down copy of
PriorityQueue that we can play with
* adds KeyValueScannerHeap (almost identical to KeyValueHeap, but uses the
above)
* includes LoserTreeKeyValueHeap and LoserTree
* each of the 3 heaps implments BenchmarkableKeyValueHeap (not complete)
* enhances KeyValueHeapBenchmark to benchmark all 3 implementations
* always tests 1mm KVs, no matter how many scanners
* sorts the input KVs, though that doesn't seem to matter much
* does a few warmup runs
One problem is that this goes from 1 to 3 implementations behind the same
interface which may not get inlined as well. Absolute performance will
therefore be lower, but hopefully performance will be comparable across the 3
implementations.
> Improve KeyValueHeap using loser tree
> -------------------------------------
>
> Key: HBASE-9969
> URL: https://issues.apache.org/jira/browse/HBASE-9969
> Project: HBase
> Issue Type: Improvement
> Components: Performance, regionserver
> Reporter: Chao Shi
> Assignee: Chao Shi
> Fix For: 0.98.0, 0.96.1
>
> Attachments: 9969-0.94.txt, hbase-9969-pq-v1.patch,
> hbase-9969-v2.patch, hbase-9969-v3.patch, hbase-9969.patch, hbase-9969.patch,
> kvheap-benchmark.png, kvheap-benchmark.txt
>
>
> LoserTree is the better data structure than binary heap. It saves half of the
> comparisons on each next(), though the time complexity is on O(logN).
> Currently A scan or get will go through two KeyValueHeaps, one is merging KVs
> read from multiple HFiles in a single store, the other is merging results
> from multiple stores. This patch should improve the both cases whenever CPU
> is the bottleneck (e.g. scan with filter over cached blocks, HBASE-9811).
> All of the optimization work is done in KeyValueHeap and does not change its
> public interfaces. The new code looks more cleaner and simpler to understand.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
