Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
HBase >> mail # dev >> Comparing the performance of 0.98.4 RC0 and 0.98.0 using YCSB - 23% perf regression in workload E


Copy link to this message
-
Re: Comparing the performance of 0.98.4 RC0 and 0.98.0 using YCSB - 23% perf regression in workload E
Workload E tests on a new testbed instance with careful attention to
configuration do not produce the same results. We have instead:

*​Workload E*  - 0.98.4RC0

[OVERALL] RunTime(ms)1270229 [OVERALL]Throughput(ops/sec)
7944[INSERT] Operations
499175 [INSERT]AverageLatency(us)
18[INSERT]MinLatency(us)
5[INSERT] MaxLatency(us)
571160 [INSERT]95thPercentileLatency(ms)
0[INSERT] 99thPercentileLatency(ms)
0 [SCAN]Operations
9500825[SCAN]AverageLatency(us)
​​
 ​​
21089[SCAN] MinLatency(us)
772 [SCAN]MaxLatency(us)
3300020[SCAN]95thPercentileLatency(ms)
107[SCAN] 99thPercentileLatency(ms)
152

I ran workload E a few times to insure the results were consistent. They
vary a bit due to natural variance but not by 23%.
On Thu, Jul 17, 2014 at 5:08 PM, Andrew Purtell <[EMAIL PROTECTED]> wrote:
Best regards,

   - Andy

Problems worthy of attack prove their worth by hitting back. - Piet Hein
(via Tom White)