Databases
StackOverflow2013
Postsdbo
PO
 

Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    primarykey
    data
    text
    <p>I think <a href="https://stackoverflow.com/a/16667228/912144">this answer</a> describes the reason sufficiently, but I'll expand a bit here.</p> <p>Before, however, here's <a href="http://gcc.gnu.org/onlinedocs/gcc-4.8.0/gcc/C-Dialect-Options.html#C-Dialect-Options" rel="nofollow noreferrer">gcc 4.8's documentation on <code>-fopenmp</code></a>:</p> <blockquote> <p><code>-fopenmp</code><br> Enable handling of OpenMP directives #pragma omp in C/C++ and !$omp in Fortran. When -fopenmp is specified, the compiler generates parallel code according to the OpenMP Application Program Interface v3.0 <a href="http://www.openmp.org/" rel="nofollow noreferrer">http://www.openmp.org/</a>. This option implies -pthread, and thus is only supported on targets that have support for -pthread.</p> </blockquote> <p>Note that it doesn't specify disabling of any features. Indeed, there is no reason for gcc to disable any optimization.</p> <p>The reason however why openmp with 1 thread has overhead with respect to no openmp is the fact that the compiler needs to convert the code, adding functions so it would be ready for cases with openmp with n>1 threads. So let's think of a simple example:</p> <pre><code>int *b = ... int *c = ... int a = 0; #omp parallel for reduction(+:a) for (i = 0; i &lt; 100; ++i) a += b[i] + c[i]; </code></pre> <p>This code should be converted to something like this:</p> <pre><code>struct __omp_func1_data { int start; int end; int *b; int *c; int a; }; void *__omp_func1(void *data) { struct __omp_func1_data *d = data; int i; d-&gt;a = 0; for (i = d-&gt;start; i &lt; d-&gt;end; ++i) d-&gt;a += d-&gt;b[i] + d-&gt;c[i]; return NULL; } ... for (t = 1; t &lt; nthreads; ++t) /* create_thread with __omp_func1 function */ /* for master thread, don't create a thread */ struct master_data md = { .start = /*...*/, .end = /*...*/ .b = b, .c = c }; __omp_func1(&amp;md); a += md.a; for (t = 1; t &lt; nthreads; ++t) { /* join with thread */ /* add thread_data-&gt;a to a */ } </code></pre> <p>Now if we run this with <code>nthreads==1</code>, the code effectively gets reduced to:</p> <pre><code>struct __omp_func1_data { int start; int end; int *b; int *c; int a; }; void *__omp_func1(void *data) { struct __omp_func1_data *d = data; int i; d-&gt;a = 0; for (i = d-&gt;start; i &lt; d-&gt;end; ++i) d-&gt;a += d-&gt;b[i] + d-&gt;c[i]; return NULL; } ... struct master_data md = { .start = 0, .end = 100 .b = b, .c = c }; __omp_func1(&amp;md); a += md.a; </code></pre> <p>So what are the differences between the no openmp version and the single threaded openmp version?</p> <p>One difference is that there is extra glue code. The variables that need to be passed to the function created by openmp need to be put together to form one argument. So there is some overhead preparing for the function call (and later retrieving data)</p> <p>More importantly however, is that now the code is not in one piece any more. Cross-function optimization is not so advanced yet and most optimizations are done within each function. Smaller functions means there is smaller possibility to optimize.</p> <hr> <p>To finish this answer, I'd like to show you exactly how <code>-fopenmp</code> affects <code>gcc</code>'s options. (Note: I'm on an old computer now, so I have gcc 4.4.3)</p> <p>Running <code>gcc -Q -v some_file.c</code> gives this (relevant) output:</p> <pre><code>GGC heuristics: --param ggc-min-expand=98 --param ggc-min-heapsize=128106 options passed: -v a.c -D_FORTIFY_SOURCE=2 -mtune=generic -march=i486 -fstack-protector options enabled: -falign-loops -fargument-alias -fauto-inc-dec -fbranch-count-reg -fcommon -fdwarf2-cfi-asm -fearly-inlining -feliminate-unused-debug-types -ffunction-cse -fgcse-lm -fident -finline-functions-called-once -fira-share-save-slots -fira-share-spill-slots -fivopts -fkeep-static-consts -fleading-underscore -fmath-errno -fmerge-debug-strings -fmove-loop-invariants -fpcc-struct-return -fpeephole -fsched-interblock -fsched-spec -fsched-stalled-insns-dep -fsigned-zeros -fsplit-ivs-in-unroller -fstack-protector -ftrapping-math -ftree-cselim -ftree-loop-im -ftree-loop-ivcanon -ftree-loop-optimize -ftree-parallelize-loops= -ftree-reassoc -ftree-scev-cprop -ftree-switch-conversion -ftree-vect-loop-version -funit-at-a-time -fvar-tracking -fvect-cost-model -fzero-initialized-in-bss -m32 -m80387 -m96bit-long-double -maccumulate-outgoing-args -malign-stringops -mfancy-math-387 -mfp-ret-in-387 -mfused-madd -mglibc -mieee-fp -mno-red-zone -mno-sse4 -mpush-args -msahf -mtls-direct-seg-refs </code></pre> <p>and running <code>gcc -Q -v -fopenmp some_file.c</code> gives this (relevant) output:</p> <pre><code>GGC heuristics: --param ggc-min-expand=98 --param ggc-min-heapsize=128106 options passed: -v -D_REENTRANT a.c -D_FORTIFY_SOURCE=2 -mtune=generic -march=i486 -fopenmp -fstack-protector options enabled: -falign-loops -fargument-alias -fauto-inc-dec -fbranch-count-reg -fcommon -fdwarf2-cfi-asm -fearly-inlining -feliminate-unused-debug-types -ffunction-cse -fgcse-lm -fident -finline-functions-called-once -fira-share-save-slots -fira-share-spill-slots -fivopts -fkeep-static-consts -fleading-underscore -fmath-errno -fmerge-debug-strings -fmove-loop-invariants -fpcc-struct-return -fpeephole -fsched-interblock -fsched-spec -fsched-stalled-insns-dep -fsigned-zeros -fsplit-ivs-in-unroller -fstack-protector -ftrapping-math -ftree-cselim -ftree-loop-im -ftree-loop-ivcanon -ftree-loop-optimize -ftree-parallelize-loops= -ftree-reassoc -ftree-scev-cprop -ftree-switch-conversion -ftree-vect-loop-version -funit-at-a-time -fvar-tracking -fvect-cost-model -fzero-initialized-in-bss -m32 -m80387 -m96bit-long-double -maccumulate-outgoing-args -malign-stringops -mfancy-math-387 -mfp-ret-in-387 -mfused-madd -mglibc -mieee-fp -mno-red-zone -mno-sse4 -mpush-args -msahf -mtls-direct-seg-refs </code></pre> <p>Taking a diff, we can see that the only difference is that with <code>-fopenmp</code>, we have <code>-D_REENTRANT</code> defined (and of course <code>-fopenmp</code> enabled). So, rest assured, gcc wouldn't produce worse code. It's just that it needs to add preparation code for when number of threads is greater than 1 and that has some overhead.</p> <hr> <p><strong>Update:</strong> I really should have tested this with optimization enabled. Anyway, with gcc 4.7.3, the output of the same commands, added <code>-O3</code> will give the same difference. So, even with <code>-O3</code>, there are no optimization's disabled.</p>
    singulars
    1. This table or related slice is empty.
    1. POMay compiler optimizations be inhibited by multi-threading?
      singulars
      1. PTQuestion
    1. PTAnswer
    1. USCommunity
    1. USShahbaz
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. POMay compiler optimizations be inhibited by multi-threading?
      singulars
      1. PTQuestion
    1. This table or related slice is empty.
    1. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTAcceptedByOriginator
    3. VO
      singulars
      1. PO
      1. This table or related slice is empty.
      1. VTUpMod
    1. COGreat explanation for GCC @Shahbaz. It would be interesting to see a similar discussion regarding MSVC.
      singulars
      1. PO
      1. This table or related slice is empty.
    2. CO@raxman, thanks. I've only recently learned OpenMP, so I'm not much experienced with it. I have no idea about MSVC. However, if you can find an option that would output that would output enabled optimizations, you can test it. Otherwise, the general explanation is independent of gcc (it's more like my imagination of how OpenMP is probably implemented)
      singulars
      1. PO
      1. USShahbaz
    3. COI've only recently learned OpenMP as well (about 3 months now). I have used pthreads (on Windows as well) so your explanation is quite clear. I don't know how MS implements OpenMP in MSVC. I do know that its auto-vectorization conflicts with OpenMP so that's once difference between GCC and MSVC. I use MSVC by day and GCC by night so I'm stuck with both.
      singulars
      1. PO
      1. This table or related slice is empty.
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload