PATCH (not to be comitted): Support for printing instruction statistics

PATCH (not to be committed): Support for printing instruction statistics

This change adds a --print-instruction-stats flag to the VM which will print out the
distribution of RX code size based on type of instruction.


The statistics will be sorted and printed like this:
 ...
 1.66 %    43957 bytes      1864 count    23.58 avg bytes/entry    -    StoreIndexed
 1.69 %    44639 bytes       288 count   155.00 avg bytes/entry    -    <copy parameters>
 1.99 %    52616 bytes     15085 count     3.49 avg bytes/entry    -    TargetEntry
 2.45 %    64872 bytes     26897 count     2.41 avg bytes/entry    -    PushArgument
 2.52 %    66550 bytes      2918 count    22.81 avg bytes/entry    -    CheckedSmiOp
 2.62 %    69338 bytes      3780 count    18.34 avg bytes/entry    -    <frame-entry>
 2.81 %    74344 bytes      6886 count    10.80 avg bytes/entry    -    Goto
 3.37 %    89178 bytes       320 count   278.68 avg bytes/entry    -    <stub-code>
 3.38 %    89333 bytes      6701 count    13.33 avg bytes/entry    -    Branch
 3.44 %    90961 bytes      1327 count    68.55 avg bytes/entry    -    <try-sync-spilling-code>
 4.14 %   109423 bytes      4939 count    22.15 avg bytes/entry    -    Return
 4.33 %   114471 bytes      2918 count    39.23 avg bytes/entry    -    <slow-path:checked-smi>
 4.39 %   115981 bytes     36639 count     3.17 avg bytes/entry    -    ParallelMove
 4.63 %   122436 bytes      3222 count    38.00 avg bytes/entry    -    <checked-entry-prologue>
 4.75 %   125730 bytes      2319 count    54.22 avg bytes/entry    -    <slow-path:checked-smi-compare>
 6.70 %   177308 bytes      3863 count    45.90 avg bytes/entry    -    <slow-path:stack-overflow>
 8.99 %   237752 bytes      8696 count    27.34 avg bytes/entry    -    InstanceCall
 9.47 %   250515 bytes      9934 count    25.22 avg bytes/entry    -    StoreInstanceField
--------------------
 0.03 %      680 bytes unaccounted
 2.77 %    73237 bytes alignment
 5.80 %   153312 bytes instruction object header
91.41 %  2417571 bytes instructions
--------------------
100.00 %  2644800 bytes in total
--------------------
       0 return-constant functions
       0 return-constant-with-load-field functions
      45 wasteful functions (body < 2 * frame overhead)
--------------------
--------------------
Emitted 4791 functions
--------------------

[kustermann, 2016-12-15 15:44:35]

Description was changed from

==========
PATCH (not to be comitted): Support for printing instruction statistics

This change adds a --print-instruction-stats flag to the VM which will print out
the
distribution of RX code size based on type of instruction.
==========

to

==========
PATCH (not to be committed): Support for printing instruction statistics

This change adds a --print-instruction-stats flag to the VM which will print out
the
distribution of RX code size based on type of instruction.


The statistics will be sorted and printed like this:
 ...
 1.93 %   139096 bytes       905 count   153.70 avg bytes/entry    -   
<stub-code>
 1.94 %   139842 bytes     15538 count     9.00 avg bytes/entry    -    Return
 1.97 %   142581 bytes      4924 count    28.96 avg bytes/entry    -   
PolymorphicInstanceCall
 2.22 %   160481 bytes      6174 count    25.99 avg bytes/entry    -   
StoreIndexed
 2.60 %   187603 bytes     21487 count     8.73 avg bytes/entry    -    Goto
 2.95 %   212938 bytes      5595 count    38.06 avg bytes/entry    -   
<slow-path:checked-smi>
 3.05 %   220148 bytes     12063 count    18.25 avg bytes/entry    -   
<frame-entry>
 3.42 %   246597 bytes    101243 count     2.44 avg bytes/entry    -   
PushArgument
 3.45 %   248765 bytes     20491 count    12.14 avg bytes/entry    -    Branch
 4.64 %   335067 bytes    124178 count     2.70 avg bytes/entry    -   
ParallelMove
 6.52 %   471064 bytes     12351 count    38.14 avg bytes/entry    -   
<slow-path:stack-overflow>
 9.59 %   692718 bytes     25399 count    27.27 avg bytes/entry    -   
StoreInstanceField
15.42 %  1113056 bytes     41254 count    26.98 avg bytes/entry    -   
InstanceCall
--------------------
 9.64 %   695864 bytes unaccounted
 3.39 %   244752 bytes alignment
 6.90 %   498080 bytes instruction object header
80.07 %  5780984 bytes instructions
--------------------
100.00 %  7219680 bytes in total
--------------------
    3617 return-constant functions
    3829 return-constant-with-load-field functions
    1248 wasteful functions (body < 2 * frame overhead)
--------------------
--------------------
Emitted 15565 functions
--------------------
==========

[kustermann, 2016-12-15 22:48:13]

Description was changed from

==========
PATCH (not to be committed): Support for printing instruction statistics

This change adds a --print-instruction-stats flag to the VM which will print out
the
distribution of RX code size based on type of instruction.


The statistics will be sorted and printed like this:
 ...
 1.93 %   139096 bytes       905 count   153.70 avg bytes/entry    -   
<stub-code>
 1.94 %   139842 bytes     15538 count     9.00 avg bytes/entry    -    Return
 1.97 %   142581 bytes      4924 count    28.96 avg bytes/entry    -   
PolymorphicInstanceCall
 2.22 %   160481 bytes      6174 count    25.99 avg bytes/entry    -   
StoreIndexed
 2.60 %   187603 bytes     21487 count     8.73 avg bytes/entry    -    Goto
 2.95 %   212938 bytes      5595 count    38.06 avg bytes/entry    -   
<slow-path:checked-smi>
 3.05 %   220148 bytes     12063 count    18.25 avg bytes/entry    -   
<frame-entry>
 3.42 %   246597 bytes    101243 count     2.44 avg bytes/entry    -   
PushArgument
 3.45 %   248765 bytes     20491 count    12.14 avg bytes/entry    -    Branch
 4.64 %   335067 bytes    124178 count     2.70 avg bytes/entry    -   
ParallelMove
 6.52 %   471064 bytes     12351 count    38.14 avg bytes/entry    -   
<slow-path:stack-overflow>
 9.59 %   692718 bytes     25399 count    27.27 avg bytes/entry    -   
StoreInstanceField
15.42 %  1113056 bytes     41254 count    26.98 avg bytes/entry    -   
InstanceCall
--------------------
 9.64 %   695864 bytes unaccounted
 3.39 %   244752 bytes alignment
 6.90 %   498080 bytes instruction object header
80.07 %  5780984 bytes instructions
--------------------
100.00 %  7219680 bytes in total
--------------------
    3617 return-constant functions
    3829 return-constant-with-load-field functions
    1248 wasteful functions (body < 2 * frame overhead)
--------------------
--------------------
Emitted 15565 functions
--------------------
==========

to

==========
PATCH (not to be committed): Support for printing instruction statistics

This change adds a --print-instruction-stats flag to the VM which will print out
the
distribution of RX code size based on type of instruction.


The statistics will be sorted and printed like this:
 ...
 1.66 %    43957 bytes      1864 count    23.58 avg bytes/entry    -   
StoreIndexed
 1.69 %    44639 bytes       288 count   155.00 avg bytes/entry    -    <copy
parameters>
 1.99 %    52616 bytes     15085 count     3.49 avg bytes/entry    -   
TargetEntry
 2.45 %    64872 bytes     26897 count     2.41 avg bytes/entry    -   
PushArgument
 2.52 %    66550 bytes      2918 count    22.81 avg bytes/entry    -   
CheckedSmiOp
 2.62 %    69338 bytes      3780 count    18.34 avg bytes/entry    -   
<frame-entry>
 2.81 %    74344 bytes      6886 count    10.80 avg bytes/entry    -    Goto
 3.37 %    89178 bytes       320 count   278.68 avg bytes/entry    -   
<stub-code>
 3.38 %    89333 bytes      6701 count    13.33 avg bytes/entry    -    Branch
 3.44 %    90961 bytes      1327 count    68.55 avg bytes/entry    -   
<try-sync-spilling-code>
 4.14 %   109423 bytes      4939 count    22.15 avg bytes/entry    -    Return
 4.33 %   114471 bytes      2918 count    39.23 avg bytes/entry    -   
<slow-path:checked-smi>
 4.39 %   115981 bytes     36639 count     3.17 avg bytes/entry    -   
ParallelMove
 4.63 %   122436 bytes      3222 count    38.00 avg bytes/entry    -   
<checked-entry-prologue>
 4.75 %   125730 bytes      2319 count    54.22 avg bytes/entry    -   
<slow-path:checked-smi-compare>
 6.70 %   177308 bytes      3863 count    45.90 avg bytes/entry    -   
<slow-path:stack-overflow>
 8.99 %   237752 bytes      8696 count    27.34 avg bytes/entry    -   
InstanceCall
 9.47 %   250515 bytes      9934 count    25.22 avg bytes/entry    -   
StoreInstanceField
--------------------
 0.03 %      680 bytes unaccounted
 2.77 %    73237 bytes alignment
 5.80 %   153312 bytes instruction object header
91.41 %  2417571 bytes instructions
--------------------
100.00 %  2644800 bytes in total
--------------------
       0 return-constant functions
       0 return-constant-with-load-field functions
      45 wasteful functions (body < 2 * frame overhead)
--------------------
--------------------
Emitted 4791 functions
--------------------
==========