Databases
StackOverflow2013
Postsdbo
POEffectiveness of GCC optmization on bit operations
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POEffectiveness of GCC optmization on bit operations
    primarykey
    data
    text
    <p>Here are two ways to set an individual bit in C on x86-64:</p> <pre><code>inline void SetBitC(long *array, int bit) { //Pure C version *array |= 1&lt;&lt;bit; } inline void SetBitASM(long *array, int bit) { // Using inline x86 assembly asm("bts %1,%0" : "+r" (*array) : "g" (bit)); } </code></pre> <p>Using GCC 4.3 with <code>-O3 -march=core2</code> options, the C version takes about <strong>90% more time</strong> when used with a constant <code>bit</code>. (Both versions compile to exactly the same assembly code, except that the C version uses an <code>or [1&lt;&lt;num],%rax</code> instruction instead of a <code>bts [num],%rax</code> instruction)</p> <p>When used with a variable <code>bit</code>, the C version performs better but is still significantly slower than the inline assembly.</p> <p>Resetting, toggling and checking bits have similar results.</p> <p>Why does GCC optimize so poorly for such a common operation? Am I doing something wrong with the C version?</p> <p><strong>Edit:</strong> Sorry for the long wait, here is the code I used to benchmark. It actually started as a simple programming problem...</p> <pre><code>int main() { // Get the sum of all integers from 1 to 2^28 with bit 11 always set unsigned long i,j,c=0; for (i=1; i&lt;(1&lt;&lt;28); i++) { j = i; SetBit(&amp;j, 10); c += j; } printf("Result: %lu\n", c); return 0; } gcc -O3 -march=core2 -pg test.c ./a.out gprof with ASM: 101.12 0.08 0.08 main with C: 101.12 0.16 0.16 main </code></pre> <p><code>time ./a.out</code> also gives similar results.</p>
    singulars
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PTQuestion
    1. USDumb Guy
    1. USDumb Guy
    plurals
    1. This table or related slice is empty.
    1. PL
      singulars
      1. LTLinked
    1. This table or related slice is empty.
    1. PO
      singulars
      1. PTAnswer
    2. PO
      singulars
      1. PTAnswer
    3. PO
      singulars
      1. PTAnswer
    1. VO
      singulars
      1. POEffectiveness of GCC optmization on bit operations
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. POEffectiveness of GCC optmization on bit operations
      1. This table or related slice is empty.
      1. VTUpMod
    3. VO
      singulars
      1. POEffectiveness of GCC optmization on bit operations
      1. This table or related slice is empty.
      1. VTUpMod
    1. COIs it really a "common operation"? The most common time I see bit manipulation used is in premature optimization by people thinking they'll save memory by packing a bunch of flags into a single byte.
      singulars
      1. POEffectiveness of GCC optmization on bit operations
      1. USAnon.
    2. COHmm...good point. Still, it's a pretty simple operation, and common in hardware drivers. Anyways, the point is the 90% performance decrease.
      singulars
      1. POEffectiveness of GCC optmization on bit operations
      1. USDumb Guy
    3. COIn driver programming, I'm pretty sure the "left-shift 1 to determine our mask" idiom isn't used much. Rather, you `or` with a predefined flag.
      singulars
      1. POEffectiveness of GCC optmization on bit operations
      1. USAnon.
 

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