Issue 9460015: Do not count invocations but usage of a function, i.e., increment a function's counter at IC calls …

Issue 9460015: Do not count invocations but usage of a function, i.e., increment a function's counter at IC calls … (Closed)

Created:
8 years, 10 months ago by srdjan

Modified:
8 years, 10 months ago

Reviewers:
regis

CC:
reviews_dartlang.org

Base URL:
http://dart.googlecode.com/svn/branches/bleeding_edge/dart/

Visibility:
Public.

More Reviews

Description

Do not count invocations but usage of a function, i.e., increment a function's counter at IC calls and at return operations in unoptimized code. (TODO: increment count at static calls as well). There is no checking at method entry any more. The function counter reporting is not measuring the invocation count but much more how much time we spend in a method. Removed counter at backward branches. Renamed flags to: --optimization_counter_threshold (default 2000) --report_usage_count Committed: https://code.google.com/p/dart/source/detail?r=4583

Patch Set 1 #

Patch Set 2 : #

Patch Set 3 : #

Patch Set 4 : #

Patch Set 5 : #

Patch Set 6 : #

Patch Set 7 : #

Total comments: 8

Patch Set 8 : #

Created: 8 years, 10 months ago

Download [raw] [tar.bz2]

	Unified diffs	Side-by-side diffs	Delta from patch set	Stats (+117 lines, -195 lines)			Patch
M	runtime/vm/code_generator.cc	View	1 2 3 4 5 6 7	4 chunks	+24 lines, -4 lines	0 comments	Download
	runtime/vm/code_generator_arm.h	View	1 2 3 4 5 6 7	1 chunk	+3 lines, -0 lines	0 comments	Download
M	runtime/vm/code_generator_ia32.h	View	1 2 3 4 5 6 7	1 chunk	+3 lines, -2 lines	0 comments	Download
M	runtime/vm/code_generator_ia32.cc	View	1 2 3 4 5 6 7	7 chunks	+27 lines, -53 lines	0 comments	Download
M	runtime/vm/code_generator_x64.h	View	1 2 3 4 5 6 7	1 chunk	+3 lines, -2 lines	0 comments	Download
M	runtime/vm/code_generator_x64.cc	View	1 2 3 4 5 6 7	6 chunks	+5 lines, -54 lines	0 comments	Download
M	runtime/vm/instructions_ia32_test.cc	View	1 2 3 4 5 6 7	2 chunks	+3 lines, -3 lines	0 comments	Download
M	runtime/vm/instructions_x64_test.cc	View	1 2 3 4 5 6 7	2 chunks	+3 lines, -3 lines	0 comments	Download
M	runtime/vm/isolate.cc	View	1 2 3 4 5 6 7	5 chunks	+10 lines, -10 lines	0 comments	Download
M	runtime/vm/object.h	View	1 2 3 4 5 6 7	2 chunks	+10 lines, -6 lines	0 comments	Download
M	runtime/vm/object.cc	View	1 2 3 4 5 6 7	1 chunk	+1 line, -1 line	0 comments	Download
M	runtime/vm/opt_code_generator_ia32.h	View	1 2 3 4 5 6 7	1 chunk	+0 lines, -1 line	0 comments	Download
M	runtime/vm/raw_object.h	View	1 2 3 4 5 6 7	1 chunk	+1 line, -1 line	0 comments	Download
M	runtime/vm/raw_object_snapshot.cc	View	1 2 3 4 5 6 7	2 chunks	+2 lines, -2 lines	0 comments	Download
M	runtime/vm/stub_code.h	View	1 2 3 4 5 6 7	1 chunk	+0 lines, -1 line	0 comments	Download
M	runtime/vm/stub_code_arm.cc	View	1 2 3 4 5 6 7	1 chunk	+0 lines, -5 lines	0 comments	Download
M	runtime/vm/stub_code_ia32.cc	View	1 2 3 4 5 6 7	3 chunks	+20 lines, -22 lines	0 comments	Download
M	runtime/vm/stub_code_x64.cc	View	1 2 3 4 5 6 7	2 chunks	+1 line, -23 lines	0 comments	Download
M	tests/corelib/src/CollectionToStringTest.dart	View	1 2 3 4 5 6 7	1 chunk	+0 lines, -1 line	0 comments	Download
M	tests/language/src/MathVMTest.dart	View	1 2 3 4 5 6 7	1 chunk	+1 line, -1 line	0 comments	Download

Messages

Total messages: 3 (0 generated)

Expand Messages | Collapse Messages

regis

LGTM https://chromiumcodereview.appspot.com/9460015/diff/1076/runtime/vm/code_generator.cc File runtime/vm/code_generator.cc (right): https://chromiumcodereview.appspot.com/9460015/diff/1076/runtime/vm/code_generator.cc#newcode31 runtime/vm/code_generator.cc:31: DECLARE_FLAG(int, optimization_counter_threshold); It should not be necessary to ...

8 years, 10 months ago (2012-02-24 23:56:25 UTC) #2

srdjan

8 years, 10 months ago (2012-02-25 00:22:10 UTC) #3

Thanks!

https://chromiumcodereview.appspot.com/9460015/diff/1076/runtime/vm/code_gene...
File runtime/vm/code_generator.cc (right):

https://chromiumcodereview.appspot.com/9460015/diff/1076/runtime/vm/code_gene...
runtime/vm/code_generator.cc:31: DECLARE_FLAG(int,
optimization_counter_threshold);
On 2012/02/24 23:56:25, regis wrote:
> It should not be necessary to declare a flag that was defined above.

Done.

https://chromiumcodereview.appspot.com/9460015/diff/1076/runtime/vm/code_gene...
File runtime/vm/code_generator_x64.cc (right):

https://chromiumcodereview.appspot.com/9460015/diff/1076/runtime/vm/code_gene...
runtime/vm/code_generator_x64.cc:277: // - type checks are enabled.
On 2012/02/24 23:56:25, regis wrote:
> In the 32-bit version you added a TODO here:
> // TODO(srdjan): Still needed?

Done.

https://chromiumcodereview.appspot.com/9460015/diff/1076/runtime/vm/isolate.cc
File runtime/vm/isolate.cc (right):

https://chromiumcodereview.appspot.com/9460015/diff/1076/runtime/vm/isolate.c...
runtime/vm/isolate.cc:255: static int MostCalledFunctionFirst(const Function*
const* a,
On 2012/02/24 23:56:25, regis wrote:
> Rename to MostUsedFunctionFirst?

Done.

https://chromiumcodereview.appspot.com/9460015/diff/1076/runtime/vm/stub_code...
File runtime/vm/stub_code_x64.cc (right):

https://chromiumcodereview.appspot.com/9460015/diff/1076/runtime/vm/stub_code...
runtime/vm/stub_code_x64.cc:1526: intptr_t num_args) {
On 2012/02/24 23:56:25, regis wrote:
> How about adding a TODO to check the usage counter and optimize the function
as
> in the 32-bit version?

Done.

Expand Messages | Collapse Messages