aboutsummaryrefslogtreecommitdiffstats
path: root/backend
Commit message (Collapse)AuthorAgeFilesLines
...
* | | | | Flushing debug outputCyril SIX2020-12-091-1/+1
| | | | |
* | | | | The last fix for get_loop_info was giving false positives. Fixing that.Cyril SIX2020-12-081-2/+12
| | | | |
* | | | | Fixing get_loop_info : part 2Cyril SIX2020-12-081-8/+28
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Loops with multiple CBs were only partially predicted for some cases (the header CB would get predicted, but not the CBs inside) This would happen in the case of breaks leading to another loop, such as: ```c void lift_check_level() { int i; int middle = lift_one_level >> 2; if ( lift_cntValid ) { for ( lift_level = 1; lift_level < 14; ++lift_level ) { if ( lift_cnt < lift_levelPos[lift_level] - middle ) break; /* This break */ } } else lift_level = 0; for ( i = 0; i < 14; ++i ) lift_ctrl_io_led[i] = ( i == lift_level - 1 ); } ```
* | | | | Fixing loop detection in get_loop_info - part 1Cyril SIX2020-12-081-16/+8
| | | | | | | | | | | | | | | | | | | | Various typos made the code fail silently for certain loops
* | | | | Moving codeCyril SIX2020-12-081-71/+72
| | | | |
* | | | | More debugCyril SIX2020-12-081-6/+21
| | | | |
* | | | | More debugCyril SIX2020-12-081-3/+3
| | | | |
* | | | | Do not duplicate nodes that don't need to when unrolling the bodyCyril SIX2020-12-081-4/+18
| | | | |
* | | | | Fix on find_last_node_before_loopCyril SIX2020-12-081-3/+7
| |_|/ / |/| | | | | | | | | | | It was too permissive
* | | | Merge branch 'kvx-work' into kvx-work-merge3.8Cyril SIX2020-12-0411-286/+1393
|\ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | Conflicts: Makefile configure
| * | | | CommentCyril SIX2020-12-041-0/+1
| | | | |
| * | | | Less aggressive tail duplicationCyril SIX2020-12-041-6/+11
| | | | | | | | | | | | | | | | | | | | | | | | | In some cases of two imbricated loops, we would tail-duplicate too much, because of the input trace traversing both loop headers.
| * | | | Clean-up debugCyril SIX2020-12-041-4/+2
| | | | |
| * | | | Fixed infinite loop on find_last_node_before_loopCyril SIX2020-12-041-3/+6
| | | | | | | | | | | | | | | | | | | | Happened when a loop was predicted not to be taken
| * | | | Slight perf improvementCyril SIX2020-12-021-2/+2
| | | | |
| * | | | [expensive] Behavior change when the loop has two final instructionsCyril SIX2020-12-021-9/+57
| | | | | | | | | | | | | | | | | | | | | | | | | Right now though the compilation time is too high for glpk, I need to figure out why
| * | | | Duplicateaux: Generalization of look_aheadCyril SIX2020-12-011-3/+5
| | |/ / | |/| |
| * | | Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassCyril SIX2020-12-013-8/+161
| |\| | | | | | | | | | | | | | | | | | | | | | | | | | Conflicts: arm/Op.v common/Values.v kvx/Op.v
| | * | store2_soundDavid Monniaux2020-11-251-1/+1
| | | |
| | * | clever_kill_store_soundDavid Monniaux2020-11-251-7/+5
| | | |
| | * | kill_store_soundDavid Monniaux2020-11-251-0/+49
| | | |
| | * | two lemmas admittedDavid Monniaux2020-11-253-9/+115
| | | |
| * | | Ignore loopback edges on tail-duplicateCyril SIX2020-12-011-0/+2
| | | |
| * | | Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-11-241-2/+2
| |\| |
| | * | tiny simplification in Tunnelingaux.mlSylvain Boulmé2020-11-241-2/+2
| | | |
| * | | Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-11-194-257/+672
| |\| |
| | * | minor fix in coq2html commentSylvain Boulmé2020-11-161-1/+2
| | | |
| | * | Tunneling: improved elimination of conditionsSylvain Boulmé2020-11-164-256/+670
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously, the elimination of useless conditions done in the 2nd pass of the tunneling introduced some new "nop" instructions that were not eliminated. This commit solves this issue by anticipating the elimination of conditions in the 1st pass of the tunneling, the 2nd pass being unchanged. In case of cycles in the CFG, the full elimination of useless conditions may require several iterations in the 1st pass. Moreover, in the simulation proof, the measure counting the number of eliminated steps from each node, needs to be adapted according to the modifications on the 1st pass. Hence, this measure results from a quite complex fixpoint computation: proving its properties would be very difficult. This leads us to introduce an oracle, implementing the first pass and producing the expected measure. A certified verifier directly checks that the measure provided by the oracle satisfies the properties expected by the simulation proof. Introducing this oracle/verifier pair has here the following advantage: - the proof is simpler than the original one (e.g. having a certified union-find structure is no more necessary for this pass). - the oracle is very efficient, by using imperative data-structure to compute memoized fixpoints. At runtime, the overhead induced by the verifier computations (and the actual computation of the measure) seems largely compensated by the gains obtained through the imperative oracle.
| * | | Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-11-062-27/+32
| |\| |
| * | | Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-11-053-28/+109
| |\ \ \
| * \ \ \ Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-11-041-1/+2
| |\ \ \ \
| * \ \ \ \ Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-11-034-35/+114
| |\ \ \ \ \
| * \ \ \ \ \ Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-10-313-17/+118
| |\ \ \ \ \ \
| * \ \ \ \ \ \ Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-10-304-168/+78
| |\ \ \ \ \ \ \
| * \ \ \ \ \ \ \ Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-10-294-48/+107
| |\ \ \ \ \ \ \ \
| * \ \ \ \ \ \ \ \ Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-10-283-40/+24
| |\ \ \ \ \ \ \ \ \
| * \ \ \ \ \ \ \ \ \ Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-10-274-19/+110
| |\ \ \ \ \ \ \ \ \ \
| * | | | | | | | | | | improved CSE3David Monniaux2020-10-271-12/+12
| | | | | | | | | | | |
| * | | | | | | | | | | progress in proofs on new CSE3David Monniaux2020-10-271-3/+34
| | | | | | | | | | | |
| * | | | | | | | | | | deactivate LICM by defaultDavid Monniaux2020-10-271-20/+11
| | | | | | | | | | | |
| * | | | | | | | | | | begin fixing CSE3 to keep more inductive stuffDavid Monniaux2020-10-272-10/+19
| | | | | | | | | | | |
| * | | | | | | | | | | invariant printing more aligned with RTL dumpsDavid Monniaux2020-10-271-2/+2
| | | | | | | | | | | |
| * | | | | | | | | | | print invariantsDavid Monniaux2020-10-271-11/+46
| | | | | | | | | | | |
| * | | | | | | | | | | attempt at store -> load.sDavid Monniaux2020-10-261-2/+3
| | | | | | | | | | | |
| * | | | | | | | | | | Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-10-188-143/+845
| |\ \ \ \ \ \ \ \ \ \ \
| * \ \ \ \ \ \ \ \ \ \ \ Merge remote-tracking branch 'origin/kvx-work' into kvx-test-prepassDavid Monniaux2020-10-021-2/+9
| |\ \ \ \ \ \ \ \ \ \ \ \
| * \ \ \ \ \ \ \ \ \ \ \ \ Merge remote-tracking branch 'origin/kvx-work-riscV' into kvx-test-prepassDavid Monniaux2020-09-211-1/+482
| |\ \ \ \ \ \ \ \ \ \ \ \ \
| | * | | | | | | | | | | | | risc-V now without trapping instructionsDavid Monniaux2020-09-211-0/+24
| | | | | | | | | | | | | | |
| | * | | | | | | | | | | | | moved Risc-V div ValueAOp to central locationDavid Monniaux2020-09-211-0/+215
| | | | | | | | | | | | | | |
| | * | | | | | | | | | | | | moved some "total" value domain functions to a central locationDavid Monniaux2020-09-211-1/+243
| | | | | | | | | | | | | | |
generated by cgit (git 2.34.1) at 2024-07-04 10:09:00 +0000