Skip to content

perf: use trunc(dval(&u)) instead of (Long)(dval(&u)) #22676

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
henderkes wants to merge 1 commit into
base: master
Choose a base branch
from
+25 −0
Open

perf: use trunc(dval(&u)) instead of (Long)(dval(&u)) #22676

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
25 changes: 25 additions & 0 deletions Zend/zend_strtod.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3700,6 +3700,15 @@ zend_freedtoa(char *s)
* calculation.
*/

/* trunc(dval(&u)) keeps the per-digit remainder chain in the FP unit instead of a
* FP->int->FP round-trip, cutting the digit loop's critical-path latency and
* speeding up fixed-precision float->string ((string)/echo/printf/number_format).
* Only a win where trunc() lowers to one instruction (aarch64 frintz, x86
* roundsd), elsewhere it's a call, so keep (Long)(dval(&u)). */
#if defined(__SSE4_1__) || defined(__aarch64__)
# define FAST_TRUNC
#endif

ZEND_API char *zend_dtoa(double dd, int mode, int ndigits, int *decpt, bool *sign, char **rve)
{
/* Arguments ndigits, decpt, sign are similar to those
Expand Down Expand Up @@ -3747,6 +3756,9 @@ ZEND_API char *zend_dtoa(double dd, int mode, int ndigits, int *decpt, bool *sig
Bigint *b, *b1, *delta, *mlo, *mhi, *S;
U d2, eps, u;
double ds;
#ifdef FAST_TRUNC
double tr;
#endif
char *s, *s0;
#ifndef No_leftright
#ifdef IEEE_Arith
Expand Down Expand Up @@ -4047,9 +4059,16 @@ ZEND_API char *zend_dtoa(double dd, int mode, int ndigits, int *decpt, bool *sig
/* Generate ilim digits, then fix them up. */
dval(&eps) *= tens[ilim-1];
for(i = 1;; i++, dval(&u) *= 10.) {
#ifdef FAST_TRUNC
tr = trunc(dval(&u));
L = (Long)tr;
if (!(dval(&u) -= tr))
ilim = i;
#else
L = (Long)(dval(&u));
if (!(dval(&u) -= L))
ilim = i;
#endif
*s++ = '0' + (int)L;
if (i == ilim) {
if (dval(&u) > 0.5 + dval(&eps))
Expand Down Expand Up @@ -4084,8 +4103,14 @@ ZEND_API char *zend_dtoa(double dd, int mode, int ndigits, int *decpt, bool *sig
goto one_digit;
}
for(i = 1;; i++, dval(&u) *= 10.) {
#ifdef FAST_TRUNC
tr = trunc(dval(&u) / ds);
L = (Long)tr;
dval(&u) -= tr*ds;
#else
L = (Long)(dval(&u) / ds);
dval(&u) -= L*ds;
#endif
#ifdef Check_FLT_ROUNDS
/* If FLT_ROUNDS == 2, L will usually be high by 1 */
if (dval(&u) < 0) {
Expand Down
Loading