Re: [HACKERS] WAL CPU overhead/optimization (was Master-slave visibility order)

2013-09-03 Thread Andres Freund
On 2013-09-03 23:25:01 +0200, Hannu Krosing wrote: > On 09/03/2013 05:27 PM, Robert Haas wrote: > > On Thu, Aug 29, 2013 at 8:22 PM, Ants Aasma wrote: > >> I might give it a shot later this cycle as I have familiarized my self > >> with the problem domain anyway. I understand the appeal of staying

Re: [HACKERS] WAL CPU overhead/optimization (was Master-slave visibility order)

2013-09-03 Thread Hannu Krosing
On 09/03/2013 05:27 PM, Robert Haas wrote: > On Thu, Aug 29, 2013 at 8:22 PM, Ants Aasma wrote: >> I might give it a shot later this cycle as I have familiarized my self >> with the problem domain anyway. I understand the appeal of staying >> with what we have, but this would cap the speedup at 4x

Re: [HACKERS] WAL CPU overhead/optimization (was Master-slave visibility order)

2013-09-03 Thread Robert Haas
On Thu, Aug 29, 2013 at 8:22 PM, Ants Aasma wrote: > I might give it a shot later this cycle as I have familiarized my self > with the problem domain anyway. I understand the appeal of staying > with what we have, but this would cap the speedup at 4x and has large > caveats with the extra lookup t

Re: [HACKERS] WAL CPU overhead/optimization (was Master-slave visibility order)

2013-08-30 Thread k...@rice.edu
On Fri, Aug 30, 2013 at 03:22:37AM +0300, Ants Aasma wrote: > On Fri, Aug 30, 2013 at 3:02 AM, Andres Freund wrote: > > I am not sure "hot cache large buffer performance" is really the > > interesting case. Most of the XLogInsert()s are pretty small in the > > common workloads. I vaguely recall tr

Re: [HACKERS] WAL CPU overhead/optimization (was Master-slave visibility order)

2013-08-29 Thread Ants Aasma
On Fri, Aug 30, 2013 at 3:02 AM, Andres Freund wrote: > I am not sure "hot cache large buffer performance" is really the > interesting case. Most of the XLogInsert()s are pretty small in the > common workloads. I vaguely recall trying 8 and getting worse > performance on many workloads, but that m

Re: [HACKERS] WAL CPU overhead/optimization (was Master-slave visibility order)

2013-08-29 Thread Andres Freund
On 2013-08-30 02:53:54 +0300, Ants Aasma wrote: > On Fri, Aug 30, 2013 at 1:30 AM, Andres Freund wrote: > > On 2013-08-30 01:10:40 +0300, Ants Aasma wrote: > >> On Fri, Aug 30, 2013 at 12:33 AM, Andres Freund > >> wrote: > >> > FWIW, WAL is still the major bottleneck for INSERT heavy workloads.

Re: [HACKERS] WAL CPU overhead/optimization (was Master-slave visibility order)

2013-08-29 Thread Ants Aasma
On Fri, Aug 30, 2013 at 1:30 AM, Andres Freund wrote: > On 2013-08-30 01:10:40 +0300, Ants Aasma wrote: >> On Fri, Aug 30, 2013 at 12:33 AM, Andres Freund >> wrote: >> > FWIW, WAL is still the major bottleneck for INSERT heavy workloads. The >> > per CPU overhead actually minimally increased (at

[HACKERS] WAL CPU overhead/optimization (was Master-slave visibility order)

2013-08-29 Thread Andres Freund
On 2013-08-30 01:10:40 +0300, Ants Aasma wrote: > On Fri, Aug 30, 2013 at 12:33 AM, Andres Freund > wrote: > > FWIW, WAL is still the major bottleneck for INSERT heavy workloads. The > > per CPU overhead actually minimally increased (at least in my tests), it > > just scales noticeably better tha