[Bug libstdc++/83938] Speed up inplace_merge() algorithm & fix inefficient logic

[john.chang at samba dot tv]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83938

--- Comment #3 from WeiChungChang  ---
Created attachment 43183
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=43183&action=edit
proposed solution

Proposed solution

[Bug libstdc++/83938] Speed up inplace_merge() algorithm & fix inefficient logic

[john.chang at samba dot tv]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83938

WeiChungChang  changed:

   What|Removed |Added

URL||https://drive.google.com/op
   ||en?id=1MTigPFnqy7tscE-eCPiw
   ||9Y8_1blDTL6O

--- Comment #2 from WeiChungChang  ---
Or access the report via this link

https://drive.google.com/open?id=1MTigPFnqy7tscE-eCPiw9Y8_1blDTL6O

[Bug libstdc++/83938] Speed up inplace_merge() algorithm & fix inefficient logic

[john.chang at samba dot tv]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83938

--- Comment #1 from WeiChungChang  ---
Created attachment 43182
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=43182&action=edit
report

report in detail

[Bug libstdc++/83938] New: Speed up inplace_merge() algorithm & fix inefficient logic

[john.chang at samba dot tv]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83938

Bug ID: 83938
   Summary: Speed up inplace_merge() algorithm & fix inefficient
logic
   Product: gcc
   Version: 7.2.0
Status: UNCONFIRMED
  Severity: normal
  Priority: P3
 Component: libstdc++
  Assignee: unassigned at gcc dot gnu.org
  Reporter: john.chang at samba dot tv
  Target Milestone: ---

In this report we review the inplace merge algorithm of current GCC libstdc++.

We survey a potential issue about downgrade of performance when merging two
sorted vectors given limited buffer.

The experiment result shows at most a 49.18% gain in execution time or reduce
50 % in memory usage.

Please estimate to enhance it.

Attached is the report in detail.

Thanks.