python - Numba function slower than C++ and loop re-order further slows down x10 -


the following code simulates extracting binary words different locations within set of images.

the numba wrapped function, wordcalc in code below, has 2 problems:

  1. it 3 times slower compared similar implementation in c++.
  2. most strangely, if switch order of "ibase" , "ibit" for-loops, speed drops factor of 10 (!). not happen in c++ implementation remains unaffected.

i'm using numba 0.18.2 winpython 2.7

what causing this?

imdim = 80 numinsts = 10**4 numinstssub = 10**4/4 bitsnum = 13;  xs = np.random.rand(numinsts, imdim**2)        iinstinds = np.array(range(numinsts)[::4]) baseinds = np.arange(imdim**2 - imdim*20 + 1) ofst1 = np.random.randint(0, imdim*20, bitsnum) ofst2 = np.random.randint(0, imdim*20, bitsnum)  @nb.jit(nopython=true) def wordcalc(xs, iinstinds, baseinds, ofst, bitsnum, newxz):     count = 0     in iinstinds:         xi = xs[i]                 ibit in range(bitsnum):             ibase in range(baseinds.shape[0]):                                     u = xi[baseinds[ibase] + ofst[0, ibit]] > xi[baseinds[ibase] + ofst[1, ibit]]                 newxz[count, ibase] = newxz[count, ibase] | np.uint16(u * (2**ibit))         count += 1     return newxz  ret = wordcalc(xs, iinstinds, baseinds, np.array([ofst1, ofst2]), bitsnum, np.zeros((iinstinds.size, baseinds.size), dtype=np.uint16))

i 4x speed-up changing np.uint16(u * (2**ibit)) np.uint16(u << ibit); i.e. replace power of 2 bitshift, should equivalent (for integers).

it seems reasonably c++ compiler might making substitution itself.

swapping order of 2 loops makes small difference me both original version (5%) , optimized version (15%), can't think can make useful comment on that.

if wanted compare numba , c++ can @ compiled numba function doing os.environ['numba_dump_assembly']='1' before import numba. (that's quite involved though).

for reference, i'm using numba 0.19.1.


Comments

Post a Comment

Popular posts from this blog

powershell Start-Process exit code -1073741502 when used with Credential from a windows service environment -

i'm running strange behavior powershell start-process call. here call: $process = start-process ` "c:\somepath\mybinary.exe" ` -passthru ` -credential $defaultcredential ` -wait ` -workingdirectory "c:\somepath" ` -loaduserprofile if ($process.exitcode -ne 0) { #do } this call return exit code - 1073741502 . after quick search, exit code seems related generic error when program not load required dll (aka. status_dll_init_failed ). when run without -credential $credential program runs correctly. in order isolate problem, manually launched some.exe in prompt target credential , runs smoothly. so problem seems come way start-process cmdlet launch process. i found potential solutions problem tried apply no luck : link , link . would have idea of what's going on here ? edit 1: run proc mon monitoring program activities when launched directly or via powershell script. problem seems occur when loading kernelbase
Read more

twig - Using Twigbridge in a Laravel 5.1 Package -

i'm trying create laravel 5.1 package uses twigbridge instead of .php views. can normal .php views display fine, when try make use of .twig templates i'm met error. possible use twigbridge templates when creating views laravel package or pretty stuck blade , php ? edit 1: loading twig template without extending works. it's when start using {% extends 'file' %} error comes out. i error error loading /virtual/laravel/packages/username/mypackage/src/views/sample.twig: template "templates.master" not defined () in "/virtual/laravel/packages/username/mypackage/src/views/sample.twig" @ line 1. my folders username |_ mypackage |_ src |_ controllers - samplecontroller.php |_ views |_ templates - master.twig - sample.twig twig template sample.twig {% extends 'templates.master' %} {% block content %} {{ parent()
Read more

c# - LINQ join Entities from HashSet's, Join vs Dictionary vs HashSet performance -

i have hashset's each stores t, have written test application compares different relation algorithms can think of, i'm not pleased results getting. does there exist more efficient ways of achieving object relations, once testing? using system; using system.collections.generic; using system.linq; using system.text; using system.threading.tasks; using system.diagnostics; namespace linqtests { class program { static void main(string[] args) { hashset<user> usertable = new hashset<user>(); hashset<userproperty> userpropertytable = new hashset<userproperty>(); #region lets create dummy data. console.writeline("please wait while creating dummy data, can take while..."); console.writeline(""); int rows = 1000000; for(int x = 0; x < rows; x++) { random rnd = new random(); // ad
Read more