Re: Compression of full-page-writes

On 08 October 2013 18:42 KONDO Mitsumasa wrote:
>(2013/10/08 20:13), Haribabu kommi wrote:
>> I will test with sync_commit=on mode and provide the test results.
>OK. Thanks!

Pgbench test results with synchronous_commit mode as on.

                        Thread-1                    Threads-2
                Head code        FPW compress    Head code        FPW compress
Pgbench-org 5min        138(0.24GB)        131(0.04GB)        160(0.28GB)        163(0.05GB)
Pgbench-1000 5min        140(0.29GB)        128(0.03GB)        160(0.33GB)        162(0.02GB)
Pgbench-org 15min        141(0.59GB)        136(0.12GB)        160(0.65GB)        162(0.14GB)
Pgbench-1000 15min    138(0.81GB)        134(0.11GB)        159(0.92GB)        162(0.18GB)

Pgbench-org - original pgbench
Pgbench-1000 - changed pgbench with a record size of 1000.
5 min - pgbench test carried out for 5 min.
15 min - pgbench test carried out for 15 min.


From the above readings it is observed that,
1. There a performance dip in one thread test, the amount of dip reduces with the test run time.
2. More than 75% WAL reduction in all scenarios.

Please find the attached sheet for more details regarding machine and test configuration


Regards,
Hari babu.