[interpreter] Add relative numbers to dispatch report

[interpreter] Add relative numbers to dispatch report

This adds an additional column with percentages to the output of
bytecode_dispatch_report.py --top-dispatches-for-bytecode.
The percentages always represent the relative number of dispatches to
the target bytecode to all dispatches from the source bytecode.
The additional flag --sort-sources-relative/-r allows sorting the
"Top sources of dispatches to" the given bytecode by this column to more
easily find bytecodes that significantly often dispatch to the target.

BUG=v8:4899
LOG=N

Committed: https://crrev.com/a3f598fc85b274320dd4556b36c6c36087e80bcd
Cr-Commit-Position: refs/heads/master@{#37873}

[klaasb, 2016-07-18 18:50:02]

Description was changed from

==========
[interpreter] Add relative numbers to dispatch report

This adds an additional column with percentages to the output of
bytecode_dispatch_report.py --top-dispatches-for-bytecode.
The percentages always represent the relative number of dispatches to
the target bytecode to all dispatches from the source bytecode.
The additional flag --sort-sources-relative/-r allows sorting the
"Top sources of dispatches to" the given bytecode by this column to more
easily find bytecodes that significantly often dispatch to the target.

BUG=v8:4899
LOG=N
==========

to

==========
[interpreter] Add relative numbers to dispatch report

This adds an additional column with percentages to the output of
bytecode_dispatch_report.py --top-dispatches-for-bytecode.
The percentages always represent the relative number of dispatches to
the target bytecode to all dispatches from the source bytecode.
The additional flag --sort-sources-relative/-r allows sorting the
"Top sources of dispatches to" the given bytecode by this column to more
easily find bytecodes that significantly often dispatch to the target.

BUG=v8:4899
LOG=N
==========

[klaasb, 2016-07-18 18:50:03]

klaasb@google.com changed reviewers:
+ rmcilroy@chromium.org

[klaasb, 2016-07-18 18:56:49]

Hi,

when trying to find out which bytecodes actually dispatch often to Star, I added
a small feature to the dispatch report script.

[rmcilroy, 2016-07-19 09:23:08]

Nice, looks like a useful feature. One suggestion, let me know what you think

https://codereview.chromium.org/2159683003/diff/1/tools/ignition/bytecode_dis...
File tools/ignition/bytecode_dispatches_report.py (right):

https://codereview.chromium.org/2159683003/diff/1/tools/ignition/bytecode_dis...
tools/ignition/bytecode_dispatches_report.py:222: "speficied bytecode")
fix typo

https://codereview.chromium.org/2159683003/diff/1/tools/ignition/bytecode_dis...
tools/ignition/bytecode_dispatches_report.py:255: destinations[destination] =
(count, count / total)
I think it would be better to do this as a processing stage in
print_top_dispatch_sources_and_destinations (with the result being a new table,
rather than modifying the existing one), since ratio is only used there. WDYT?

[klaasb, 2016-07-19 10:24:27]

https://codereview.chromium.org/2159683003/diff/1/tools/ignition/bytecode_dis...
File tools/ignition/bytecode_dispatches_report.py (right):

https://codereview.chromium.org/2159683003/diff/1/tools/ignition/bytecode_dis...
tools/ignition/bytecode_dispatches_report.py:222: "speficied bytecode")
On 2016/07/19 09:23:08, rmcilroy wrote:
> fix typo

Done.

https://codereview.chromium.org/2159683003/diff/1/tools/ignition/bytecode_dis...
tools/ignition/bytecode_dispatches_report.py:255: destinations[destination] =
(count, count / total)
On 2016/07/19 09:23:08, rmcilroy wrote:
> I think it would be better to do this as a processing stage in
> print_top_dispatch_sources_and_destinations (with the result being a new
table,
> rather than modifying the existing one), since ratio is only used there. WDYT?

Done.

[rmcilroy, 2016-07-19 11:33:15]

lgtm

https://codereview.chromium.org/2159683003/diff/20001/tools/ignition/bytecode...
File tools/ignition/bytecode_dispatches_report.py (right):

https://codereview.chromium.org/2159683003/diff/20001/tools/ignition/bytecode...
tools/ignition/bytecode_dispatches_report.py:95: def
find_top_dispatch_sources(dispatches_table, bytecode, top_count,
update name (find_top_dispatch_sources_and_destinations)

https://codereview.chromium.org/2159683003/diff/20001/tools/ignition/bytecode...
tools/ignition/bytecode_dispatches_report.py:106: for source, table_row in
iteritems(relative_table):
It seems a bit strange to have an inner function part-way through the function
which captures something initialized earlier in the function. 

Could you either pass relative_table as an argument to this, or just make this
build a table inline rather than being an inner generator function.

[klaasb, 2016-07-19 13:52:28]

https://codereview.chromium.org/2159683003/diff/20001/tools/ignition/bytecode...
File tools/ignition/bytecode_dispatches_report.py (right):

https://codereview.chromium.org/2159683003/diff/20001/tools/ignition/bytecode...
tools/ignition/bytecode_dispatches_report.py:95: def
find_top_dispatch_sources(dispatches_table, bytecode, top_count,
On 2016/07/19 11:33:15, rmcilroy wrote:
> update name (find_top_dispatch_sources_and_destinations)

Done.

https://codereview.chromium.org/2159683003/diff/20001/tools/ignition/bytecode...
tools/ignition/bytecode_dispatches_report.py:106: for source, table_row in
iteritems(relative_table):
On 2016/07/19 11:33:15, rmcilroy wrote:
> It seems a bit strange to have an inner function part-way through the function
> which captures something initialized earlier in the function. 
> 
> Could you either pass relative_table as an argument to this, or just make this
> build a table inline rather than being an inner generator function.

Done.

[klaasb, 2016-07-19 13:53:18]

The CQ bit was checked by klaasb@google.com

[klaasb, 2016-07-19 13:53:18]

The patchset sent to the CQ was uploaded after l-g-t-m from
rmcilroy@chromium.org
Link to the patchset: https://codereview.chromium.org/2159683003/#ps40001
(title: "don't generate whole table and don't use generator")

[commit-bot: I haz the power, 2016-07-19 13:53:23]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[klaasb, 2016-07-19 13:53:35]

The CQ bit was unchecked by klaasb@google.com

[klaasb, 2016-07-19 13:53:48]

The CQ bit was checked by klaasb@google.com

[commit-bot: I haz the power, 2016-07-19 13:53:52]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-19 14:17:23]

Description was changed from

==========
[interpreter] Add relative numbers to dispatch report

This adds an additional column with percentages to the output of
bytecode_dispatch_report.py --top-dispatches-for-bytecode.
The percentages always represent the relative number of dispatches to
the target bytecode to all dispatches from the source bytecode.
The additional flag --sort-sources-relative/-r allows sorting the
"Top sources of dispatches to" the given bytecode by this column to more
easily find bytecodes that significantly often dispatch to the target.

BUG=v8:4899
LOG=N
==========

to

==========
[interpreter] Add relative numbers to dispatch report

This adds an additional column with percentages to the output of
bytecode_dispatch_report.py --top-dispatches-for-bytecode.
The percentages always represent the relative number of dispatches to
the target bytecode to all dispatches from the source bytecode.
The additional flag --sort-sources-relative/-r allows sorting the
"Top sources of dispatches to" the given bytecode by this column to more
easily find bytecodes that significantly often dispatch to the target.

BUG=v8:4899
LOG=N
==========

[commit-bot: I haz the power, 2016-07-19 14:17:24]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-07-19 14:17:29]

CQ bit was unchecked.

[commit-bot: I haz the power, 2016-07-19 14:17:41]

Description was changed from

==========
[interpreter] Add relative numbers to dispatch report

This adds an additional column with percentages to the output of
bytecode_dispatch_report.py --top-dispatches-for-bytecode.
The percentages always represent the relative number of dispatches to
the target bytecode to all dispatches from the source bytecode.
The additional flag --sort-sources-relative/-r allows sorting the
"Top sources of dispatches to" the given bytecode by this column to more
easily find bytecodes that significantly often dispatch to the target.

BUG=v8:4899
LOG=N
==========

to

==========
[interpreter] Add relative numbers to dispatch report

This adds an additional column with percentages to the output of
bytecode_dispatch_report.py --top-dispatches-for-bytecode.
The percentages always represent the relative number of dispatches to
the target bytecode to all dispatches from the source bytecode.
The additional flag --sort-sources-relative/-r allows sorting the
"Top sources of dispatches to" the given bytecode by this column to more
easily find bytecodes that significantly often dispatch to the target.

BUG=v8:4899
LOG=N

Committed: https://crrev.com/a3f598fc85b274320dd4556b36c6c36087e80bcd
Cr-Commit-Position: refs/heads/master@{#37873}
==========

[commit-bot: I haz the power, 2016-07-19 14:17:42]

Patchset 3 (id:??) landed as
https://crrev.com/a3f598fc85b274320dd4556b36c6c36087e80bcd
Cr-Commit-Position: refs/heads/master@{#37873}