BDD/RelPrev/Renaming (Preliminary Work) #676
Conversation
…extend comments to column 100
Currently, every stream can be negated and so it ends up adding lots of conditional statements where none is necessary. Hence, if we move the '_negate' test into 'node_stream' only, then we can regain some performance
This 'dd' independent interface has bothered me for years...
This is the only place it is used, and there is no need to add the additional instructions to every other instance
SSoelvsten
added
✨ optimisation
SSoelvsten
changed the title
| level_info | ||
| operator!() const | ||
| { | ||
| return *this; | ||
| } |
There was a problem hiding this comment.
This function is dead, right?
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #676 +/- ##
=============================================
+ Coverage 97.871% 97.937% +0.066%
=============================================
Files 93 93
Lines 6999 7029 +30
=============================================
+ Hits 6850 6884 +34
+ Misses 149 145 -4 ☔ View full report in Codecov by Sentry. |
🟡 Regression Test (Picotrav 'adder')'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of -0.10% (stdev: 1.35%).
Number of samples: 3 |
🟡 Regression Test (12-Queens)'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of -1.22% (stdev: 0.81%).
Number of samples: 3 |
🟢 Regression Test (Picotrav 'arbiter')'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of 1.14% (stdev: 1.35%).
Number of samples: 3 |
🟡 Regression Test (QBF 'ep_dual/8x8_6_e-8-1_p-2-3_bwnib')'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of -0.68% (stdev: 0.67%).
Number of samples: 3 |
🟢 Regression Test (QBF 'breakthrough/3x4_19_bwnib')'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of 0.47% (stdev: 0.77%).
Number of samples: 3 |
🟢 Regression Test (QBF 'connect4/6x6_11_connect4_bwnib')'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of 0.59% (stdev: 0.42%).
Number of samples: 3 |
🟢 Regression Test (QBF 'breakthrough_dual/3x6_10_bwnib')'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of 1.21% (stdev: 1.93%).
Number of samples: 3 |
🟢 Regression Test (QBF 'httt/4x4_9_tippy_bwnib')'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of 0.92% (stdev: 2.94%).
Number of samples: 3 |
🟢 Regression Test (QBF 'domineering/5x5_13_bwnib')'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of 1.09% (stdev: 0.36%).
Number of samples: 3 |
🟢 Regression Test (QBF 'hex/hein_08_5x5-11_bwnib')'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of 2.53% (stdev: 1.16%).
Number of samples: 3 |
🟢 Regression Test (QBF 'ep/8x8_7_e-8-1_p-3-4_bwnib')'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of 1.67% (stdev: 0.76%).
Number of samples: 3 |
🟡 Regression Test (Picotrav 'mem_ctrl')'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of -0.92% (stdev: 0.83%).
Number of samples: 3 |
🟡 Regression Test (14-Queens)'ssoelvsten/adiar/bdd/relprev/renaming_prework' is a change in performance of -0.42% (stdev: 0.52%).
Number of samples: 3 |
While working on adding O(1) variable replacement if m(x) = x + b ( #502 ), there was a lot of clean-up and few optimisations done. This PR includes all but the actual feature - these ought to be included on
mainregardless of whether the later PR is acceptable or not.Doing it this way can also make us better able to measure the overhead (independently of the possible speedup/slowdown due to the
cnot(...)optimisation)