[issue27255] More opcode predictions

Serhiy Storchaka added the comment: OK, I'm only +0 on adding these predictions and fine with status quo. The ratio 28:1 looks compelling to me, but all this is not enough important. -- resolution: -> fixed stage: patch review -> resolved status: open -> closed ___

[issue27255] More opcode predictions

Raymond Hettinger added the comment: The BUILD_SLICE pairing with BINARY_SUBSCR has a tight conceptual linkage with the lookup case dominating use patterns over slice deletion and slice assignment. IIRC, I passed on that pairing long ago because even though the predictive power was good, it d

[issue27255] More opcode predictions

Serhiy Storchaka added the comment: Thank you Raymond. Committed without the prediction from DUP_TOP_TWO to BINARY_SUBSCR. What are you think about following highly predictive pairs? 1. From UNPACK_SEQUENCE to STORE_FAST with 96.5% probability. This is the 15th of most common pairs. It is mor

[issue27255] More opcode predictions

Roundup Robot added the comment: New changeset f19c2b28710e by Serhiy Storchaka in branch 'default': Issue #27255: Added more predictions in ceval.c. https://hg.python.org/cpython/rev/f19c2b28710e -- nosy: +python-dev ___ Python tracker

[issue27255] More opcode predictions

Raymond Hettinger added the comment: The second version of the patch mostly looks fine. The prediction from DUP_TOP_TWO to BINARY_SUBSCR is questionable. Because compile.c only uses it one context, the prediction rate is 100%; however, there is no intrinsic relationship between the two opco

[issue27255] More opcode predictions

Serhiy Storchaka added the comment: Since existing predictions were added many new opcodes and opcode combinations were added. Predictions need to be updated to reflect current status. The patch contained only predictions for highly predictive (>=94%) pairs of commonly used opcodes or 100% pre

[issue27255] More opcode predictions

Raymond Hettinger added the comment: Serhiy, please slow down and stop rewriting every single thing you see. Your rate of patches is prolific and hard to digest. Please give some consideration that the people who came before you (lots of them) put a lot of thought into what was done and were

[issue27255] More opcode predictions

Changes by Raymond Hettinger : -- assignee: -> rhettinger ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: http

[issue27255] More opcode predictions

New submission from Serhiy Storchaka: Currently the PREDICT() macros is used for predicticting following pair of opcodes in the ceval loop: LIST_APPEND JUMP_ABSOLUTE SET_ADD JUMP_ABSOLUTE MAP_ADD JUMP_ABSOLUTE COMPARE_OP POP_JUMP_IF_FALSE COMPARE_OP POP_JUMP_IF_TRUE GET_ITER FOR_ITER FOR_ITER S

[issue27255] More opcode predictions

Changes by Serhiy Storchaka : -- priority: normal -> low ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https: