sharpetronics/PostgreSQL
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-06-12 00:01:43 -04:00
Code Issues Projects Releases Wiki Activity

Frob numeric.c loop so that clang will auto-vectorize it too.

Browse Source 
Experimentation shows that clang will auto-vectorize the critical
multiplication loop if the termination condition is written "i2 < limit"
rather than "i2 <= limit".  This seems unbelievably stupid, but I've
reproduced it on both clang 9.0.1 (RHEL8) and 11.0.3 (macOS Catalina).
gcc doesn't care, so tweak the code to do it that way.

Discussion: https://postgr.es/m/CAJ3gD9evtA_vBo+WMYMyT-u=keHX7-r8p2w7OSRfXf42LTwCZQ@mail.gmail.com
This commit is contained in:
Tom Lane 2020-09-07 12:03:04 -04:00
parent 87e6ed7c8c
commit 9c79e646c6
1 changed files with 8 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/backend/utils/adt/numeric.c
View File

@ -8191,7 +8191,6 @@ mul_var(const NumericVar *var1, const NumericVar *var2, NumericVar *result,
int res_weight; int res_weight;
int maxdigits; int maxdigits;
int *dig; int *dig;
int *dig_i1_2;
int carry; int carry;
int maxdig; int maxdig;
int newdig; int newdig;
@ -8327,7 +8326,7 @@ mul_var(const NumericVar *var1, const NumericVar *var2, NumericVar *result,
* Add the appropriate multiple of var2 into the accumulator. * Add the appropriate multiple of var2 into the accumulator.
* *
* As above, digits of var2 can be ignored if they don't contribute, * As above, digits of var2 can be ignored if they don't contribute,
* so we only include digits for which i1+i2+2 <= res_ndigits - 1. * so we only include digits for which i1+i2+2 < res_ndigits.
* *
* This inner loop is the performance bottleneck for multiplication, * This inner loop is the performance bottleneck for multiplication,
* so we want to keep it simple enough so that it can be * so we want to keep it simple enough so that it can be
@ -8336,11 +8335,14 @@ mul_var(const NumericVar *var1, const NumericVar *var2, NumericVar *result,
* Since we aren't propagating carries in this loop, the order does * Since we aren't propagating carries in this loop, the order does
* not matter. * not matter.
*/ */
i = Min(var2ndigits - 1, res_ndigits - i1 - 3); {
dig_i1_2 = &dig[i1 + 2]; int i2limit = Min(var2ndigits, res_ndigits - i1 - 2);
for (i2 = 0; i2 <= i; i2++) int *dig_i1_2 = &dig[i1 + 2];
for (i2 = 0; i2 < i2limit; i2++)
dig_i1_2[i2] += var1digit * var2digits[i2]; dig_i1_2[i2] += var1digit * var2digits[i2];
} }
}
/* /*
* Now we do a final carry propagation pass to normalize the result, which * Now we do a final carry propagation pass to normalize the result, which