Re: 9.3 regression with dbt2

Applying your patch plus adding -fno-omit-frame-pointer, I got 54526.90
notpm.

The profile (part) below:
# Samples: 610K of event 'cycles'
# Event count (approx.): 6686532056450
#
# Overhead Command Shared Object
Symbol
# ........ .......... .............
..................................................
#
4.08% postgres postgres [.] hash_search_with_hash_value

|
--- hash_search_with_hash_value
|
--2.87%-- BufTableLookup
|
--2.86%-- ReadBuffer_common
ReadBufferExtended
|
|--0.86%-- index_fetch_heap
| |
| --0.62%-- index_getnext
| IndexNext
| ExecScan
| ExecProcNode
|
|--0.71%-- BitmapHeapNext
| ExecScan
| ExecProcNode
|
--0.57%-- _bt_relandgetbuf
|
--0.57%-- _bt_search

3.75% postgres postgres [.] heap_hot_search_buffer

|
--- heap_hot_search_buffer
|
|--1.92%-- BitmapHeapNext
| ExecScan
| ExecProcNode
| |
| --1.12%-- standard_ExecutorRun
| _SPI_execute_plan
| SPI_execute_plan
| |
| --0.70%-- payment
| ExecMakeFunctionResult
| ExecProject
| ExecResult
| ExecProcNode
| standard_ExecutorRun
| PortalRunSelect
| PortalRun
| PostgresMain
| ServerLoop
| PostmasterMain
| main
| __libc_start_main
|
--1.74%-- index_fetch_heap
|
--1.46%-- index_getnext
|
--1.45%-- IndexNext
ExecScan
ExecProcNode
|
--0.96%--
standard_ExecutorRun

_SPI_execute_plan

SPI_execute_plan
|
--0.50%--
new_order

ExecMakeFunctionResult

ExecProject

ExecResult

ExecProcNode

standard_ExecutorRun

PortalRunSelect

PortalRunFetch

PerformPortalFetch

standard_ProcessUtility

PortalRunUtility

FillPortalStore

PortalRun

PostgresMain

ServerLoop

PostmasterMain
main

__libc_start_main

3.15% postgres postgres [.] LWLockAcquire

|
--- LWLockAcquire
|
--1.65%-- ReadBuffer_common
ReadBufferExtended

2.74% postgres postgres [.] PinBuffer

|
--- PinBuffer
|
--2.72%-- ReadBuffer_common
ReadBufferExtended
|
|--0.71%-- index_fetch_heap
| |
| --0.51%-- index_getnext
| IndexNext
| ExecScan
| ExecProcNode
|
|--0.67%-- BitmapHeapNext
| ExecScan
| ExecProcNode
|
--0.60%-- heapgetpage
heapgettup_pagemode
heap_getnext
SeqNext
ExecScan
ExecProcNode

2.72% postgres postgres [.] _bt_compare

|
--- _bt_compare
|
--2.51%-- _bt_binsrch
|
|--1.18%-- _bt_search
| |
| --0.74%-- _bt_first
| |
| --0.72%-- btgettuple
|
FunctionCall2Coll
|
index_getnext_tid
| |
| --0.58%--
index_getnext
| |

|
--0.58%-- IndexNext
|
ExecScan
|
ExecProcNode
|
--0.97%-- _bt_first
|
--0.94%-- btgettuple
FunctionCall2Coll
index_getnext_tid
|
--0.66%-- index_getnext
|
--0.66%--
IndexNext

ExecScan

ExecProcNode
|

--0.51%-- standard_ExecutorRun

_SPI_execute_plan

SPI_execute_plan

2.47% postgres postgres [.] AllocSetAlloc

|
--- AllocSetAlloc
|
--0.54%-- MemoryContextAllocZeroAligned