Rasmus Wriedt Larsen
e2652591a5
Python: Change perf fix PoorMansFunctionResolution
...
Thanks @yoff, this leaves us with the following evaluation, which looks
very close to the one in the other fix (but with cleaner implementation)
-- both at 688k max tuples (although numbers are not exactly the same).
```
[2021-11-24 13:48:40] (14s) Tuple counts for PoorMansFunctionResolution::getSimpleMethodReferenceWithinClass#ff/2@e5f05asv after 74ms:
47493 ~3% {3} r1 = JOIN Class::Class::getAMethod_dispred#ff WITH py_Classes ON FIRST 1 OUTPUT Lhs.1, 0, Lhs.0
47335 ~0% {2} r2 = JOIN r1 WITH AstGenerated::Function_::getArg_dispred#fff ON FIRST 2 OUTPUT Rhs.2, Lhs.2
46683 ~0% {2} r3 = JOIN r2 WITH DataFlowPublic::ParameterNode::getParameter_dispred#fb_10#join_rhs ON FIRST 1 OUTPUT Rhs.1, Lhs.1
259968 ~4% {2} r4 = JOIN r3 WITH LocalSources::Cached::hasLocalSource#ff_10#join_rhs ON FIRST 1 OUTPUT Rhs.1, Lhs.1
161985 ~0% {3} r5 = JOIN r4 WITH Attributes::AttrRef::accesses_dispred#bff_102#join_rhs ON FIRST 1 OUTPUT Rhs.1 'result', Lhs.1, Rhs.2
161985 ~2% {3} r6 = JOIN r5 WITH Attributes::AttrRead#class#f ON FIRST 1 OUTPUT Lhs.2, Lhs.1, Lhs.0 'result'
688766 ~0% {3} r7 = JOIN r6 WITH Function::Function::getName_dispred#ff_10#join_rhs ON FIRST 1 OUTPUT Lhs.1, Rhs.1 'func', Lhs.2 'result'
20928 ~0% {2} r8 = JOIN r7 WITH Class::Class::getAMethod_dispred#ff ON FIRST 2 OUTPUT Lhs.1 'func', Lhs.2 'result'
return r8
```
2021-11-24 13:52:05 +01:00
Rasmus Wriedt Larsen
eaed870b31
Python: Fix performance problem in PoorMansFunctionResolution
...
Before these changes:
[2021-11-22 12:02:50] (8s) Tuple counts for PoorMansFunctionResolution::getSimpleMethodReferenceWithinClass#ff/2@cbddf257 after 8.6s:
387565 ~0% {3} r1 = JOIN Attributes::AttrRead#class#f WITH Attributes::AttrRef::accesses_dispred#bff ON FIRST 1 OUTPUT Rhs.2, Lhs.0 'result', Rhs.1
6548632 ~0% {3} r2 = JOIN r1 WITH Function::Function::getName_dispred#ff_10#join_rhs ON FIRST 1 OUTPUT Rhs.1 'func', Lhs.1 'result', Lhs.2
5640480 ~0% {4} r3 = JOIN r2 WITH Class::Class::getAMethod_dispred#ff_10#join_rhs ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'result', Lhs.2, Lhs.0 'func'
55660458 ~0% {5} r4 = JOIN r3 WITH Class::Class::getAMethod_dispred#ff ON FIRST 1 OUTPUT Rhs.1, 0, Lhs.1 'result', Lhs.2, Lhs.3 'func'
55621412 ~0% {4} r5 = JOIN r4 WITH AstGenerated::Function_::getArg_dispred#fff ON FIRST 2 OUTPUT Rhs.2, Lhs.2 'result', Lhs.3, Lhs.4 'func'
54467144 ~0% {4} r6 = JOIN r5 WITH DataFlowPublic::ParameterNode::getParameter_dispred#fb_10#join_rhs ON FIRST 1 OUTPUT Lhs.2, Rhs.1, Lhs.1 'result', Lhs.3 'func'
20928 ~0% {2} r7 = JOIN r6 WITH LocalSources::Cached::hasLocalSource#ff ON FIRST 2 OUTPUT Lhs.3 'func', Lhs.2 'result'
return r7
With these changes:
[2021-11-22 11:54:25] (415s) Tuple counts for PoorMansFunctionResolution::getSimpleMethodReferenceWithinClass_helper#fff/3@14db70a8 after 75ms:
388306 ~0% {2} r1 = JOIN Attributes::AttrRead#class#f WITH Attributes::AttrRef::getObject_dispred#bf ON FIRST 1 OUTPUT Rhs.1, Lhs.0 'read'
379420 ~4% {2} r2 = JOIN r1 WITH LocalSources::Cached::hasLocalSource#ff ON FIRST 1 OUTPUT Rhs.1, Lhs.1 'read'
175082 ~0% {2} r3 = JOIN r2 WITH DataFlowPublic::ParameterNode#class#fff ON FIRST 1 OUTPUT Rhs.2, Lhs.1 'read'
175082 ~2% {3} r4 = JOIN r3 WITH Essa::ParameterDefinition::getParameter_dispred#ff ON FIRST 1 OUTPUT 0, Rhs.1, Lhs.1 'read'
166798 ~0% {2} r5 = JOIN r4 WITH AstGenerated::Function_::getArg_dispred#fff_120#join_rhs ON FIRST 2 OUTPUT Rhs.2 'func', Lhs.2 'read'
162096 ~0% {3} r6 = JOIN r5 WITH Class::Class::getAMethod_dispred#ff_10#join_rhs ON FIRST 1 OUTPUT Lhs.0 'func', Rhs.1 'cls', Lhs.1 'read'
return r6
[2021-11-22 11:54:25] (415s) Tuple counts for PoorMansFunctionResolution::getSimpleMethodReferenceWithinClass_helper2#ffff/4@2b60f0s9 after 63ms:
162046 ~0% {3} r1 = SCAN PoorMansFunctionResolution::getSimpleMethodReferenceWithinClass_helper#fff OUTPUT In.2 'read', In.0 'func', In.1 'cls'
162046 ~0% {3} r2 = JOIN r1 WITH Attributes::AttrRead#class#f ON FIRST 1 OUTPUT Lhs.1 'func', Lhs.2 'cls', Lhs.0 'read'
162046 ~1% {3} r3 = JOIN r2 WITH py_Functions ON FIRST 1 OUTPUT Lhs.1 'cls', Lhs.2 'read', Lhs.0 'func'
162046 ~0% {3} r4 = JOIN r3 WITH py_Classes ON FIRST 1 OUTPUT Lhs.1 'read', Lhs.2 'func', Lhs.0 'cls'
161935 ~5% {4} r5 = JOIN r4 WITH Attributes::AttrRef::getAttributeName_dispred#bf ON FIRST 1 OUTPUT Rhs.1, Lhs.0 'read', Lhs.1 'func', Lhs.2 'cls'
688526 ~1% {4} r6 = JOIN r5 WITH Function::Function::getName_dispred#ff_10#join_rhs ON FIRST 1 OUTPUT Lhs.2 'func', Lhs.3 'cls', Lhs.1 'read', Rhs.1 'readFunction'
return r6
[2021-11-22 11:54:25] (415s) Tuple counts for PoorMansFunctionResolution::getSimpleMethodReferenceWithinClass#ff/2@f73ae6dq after 58ms:
688526 ~0% {4} r1 = SCAN PoorMansFunctionResolution::getSimpleMethodReferenceWithinClass_helper2#ffff OUTPUT In.1, In.0, In.3 'func', In.2 'result'
688526 ~0% {3} r2 = JOIN r1 WITH Class::Class::getAMethod_dispred#ff ON FIRST 2 OUTPUT Rhs.0, Lhs.2 'func', Lhs.3 'result'
20913 ~0% {2} r3 = JOIN r2 WITH Class::Class::getAMethod_dispred#ff ON FIRST 2 OUTPUT Lhs.1 'func', Lhs.2 'result'
return r3
We need the `pragma[only_bind_into]` in getSimpleMethodReferenceWithinClass_helper2, otherwise the tuple counts would look like, which is needlessly big.
[2021-11-22 17:14:34] (2s) Tuple counts for PoorMansFunctionResolution::getSimpleMethodReferenceWithinClass_helper2#ffff/4@5f0505h7 after 711ms:
13570510 ~3% {2} r1 = JOIN Function::Function::getName_dispred#ff_10#join_rhs WITH Attributes::AttrRef::getAttributeName_dispred#ff_10#join_rhs ON FIRST 1 OUTPUT Rhs.1 'read', Lhs.1 'readFunction'
688526 ~1% {4} r2 = JOIN r1 WITH PoorMansFunctionResolution::getSimpleMethodReferenceWithinClass_helper#fff_201#join_rhs ON FIRST 1 OUTPUT Rhs.1 'func', Rhs.2 'cls', Lhs.0 'read', Lhs.1 'readFunction'
return r2
2021-11-22 17:22:39 +01:00
