Skip to content
Snippets Groups Projects
Select Git revision
  • 4b15a3818dc3736528b603c5c5eaf121d1956899
  • master default protected
  • btypes-fixes
  • freetypes
  • embedded-codegen
  • rust-support
  • z3solver
7 results

README.md

Blame
  • user avatar
    8d8f2122
    History
    Code owners
    Assign users and groups as approvers for specific file changes. Learn more.

    Benchmarks

    Execution

    Executing benchmarks is done by executing make in each of the directories.

    Technical Data

    • Java: Java HotSpot(TM) 64-Bit Server VM (build 12.0.1+12,mixed mode, sharing)
    • C++ Compiler: Apple LLVM version 10.0.1 (clang-1001.0.46.4) with -O1 and -O2
    • ProB: 1.9.0-nightly (c5a6e9d31022d0bfe40cbcdf68e910041665ec41)
    • B2Program Commit: edce3cc7
    • Device: MacBook Air with 8 GBof RAM and a 1.6 GHz Intel i5 processor with two cores

    Results

    Machines ProB Java BI Java PI C++ PI -O1 C++ PI -O2
    Lift Runtime > 1800 156.63 27.43 78.42 0.00
    Speedup 1 > 11.49 > 65.62 > 22.95 > 180 000
    Memory - 735 188 785 628 756 736
    TrafficLight Runtime > 1800 47.04 9.05 69.09 0.00
    Speedup 1 > 38.27 > 198.9 > 26.05 > 180 000
    Memory - 855 112 447 828 756 736
    Sieve Runtime 76.31 7.71 6.49 14.63 8.94
    Speedup 1 9.9 11.76 5.22 8.54
    Memory 398 980 1 415 428 1 096 284 32 472 35 732
    Scheduler Runtime 786.74 10.62 10.49 21.57 10.32
    Speedup 1 74.08 74.99 36.47 76.23
    Memory 5 341 316 414 772 398 924 816 820
    sort_m2_data1000 Runtime 17.27 3.27 2.10 0.2 0.03
    Speedup 1 5.28 8.22 86.35 575.67
    Memory 577 808 191 280 143 864 1192 1104
    CAN Bus Runtime 273.58 7.23 6.81 7.23 2.91
    Speedup 1 37.84 40.17 37.84 94.01
    Memory 167 284 428 084 402 432 968 952
    Train Runtime 241.16 13.31 12.83 18.55 8.10
    Speedup 1 18.11 18.8 13 29.77
    Memory 163 476 377 292 376 540 984 1016
    Cruise Controller Runtime > 1800 21.26 15.26 11.90 0.30
    Speedup 1 > 84.67 > 117.96 > 151.26 > 6000
    Memory - 750 816 484 948 864 820

    Results (Rust)

    Rust development was done on a different system, therefore benchmarks are not comparable to the old table. These benchmarks where run on a Ryzen 7 1700 under Debian via wsl (Win 10) (Memory values seem pretty high. GNU time had a bug for reporting 4x the RSS, but I'm using version 1.9 here, which apparently has it fixed...)

    Machines Java PI Java BI C++ PI -O1 C++ PI -O2 Rust PI O1 Rust PI O2 Rust PI O3 Rust BI O2 Rust BI O3
    Lift Runtime 11.85 102.67 51.05 00.00 20.55 18.43 18.49 685.03 705.46
    Speedup
    Memory 2 220 976 2 247 676 3120 3060 2160 2276 2168 2572 2584
    TrafficLight Runtime 05.21 17.40 51.35 00.00 21.31 10.87 10.78 78.33 80.27
    Speedup
    Memory 1 400 352 2 202 788 3188 3188 2312 2256 2312 2496 2664
    Sieve Runtime 07.41 08.04 26.15 06.86 53.04 09.03 09.06 39.27 39.54
    Speedup
    Memory 2 094 128 2 079 056 165 528 165 548 104 764 104 608 104 604 450 804 450 688
    Scheduler Runtime 06.39 06.39 14.25 01.98 34.97 06.14 06.04 06.28 06.00
    Speedup
    Memory 1 422 120 1 415 844 3080 3136 2384 2228 2324 2528 2584
    sort_m2_data1000 Runtime 02.02 02.15 00.20 00.03 02.77 00.50 00.51 01.11 01.15
    Speedup
    Memory 204 220 206 676 3412 3420 5312 5216 5328 10072 9872
    CAN Bus Runtime 05.17 05.15 05.80 00.59 19.83 03.16 03.18 06.36 06.47
    Speedup
    Memory 930 816 917 552 3364 3208 2668 2540 2508 2912 2928
    Train Runtime 07.40 07.57 13.19 01.64 31.52 04.76 04.71 02.78 02.77
    Speedup
    Memory 1 420 528 1419832 3224 3372 2740 2636 2612 2856 2812
    Cruise Controller Runtime 13.91 14.66 08.34 00.04 05.15 05.43 05.33 14.20 14.64
    Speedup
    Memory 1 401 192 1 400 232 3160 3092 2376 2540 2352 2668 2668

    Runtime in Seconds

    Memory Usage in KB

    BI = Big Integer

    PI = Primitive Integer