floraachy
/
OpenBLAS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Replace use of FLT_MAX in x86_64 zscal.c by isinf() 

Commit def4996 fixed issues with inf and nan values in zscal,
but used FLT_MAX, where DBL_MAX or isinf() is more appropriate,
as FLT_MAX is for single precision only.
Using FLT_MAX caused test case failures in the LAPACK tests.

isinf() is consistent with the later fix 969601a1
This commit is contained in:
Bart Oldeman 2024-05-24 17:20:27 +00:00
parent 700ea74a37
commit 62f7b244ff
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
kernel/x86_64/zscal.c
View File

@ -394,7 +394,7 @@ int CNAME(BLASLONG n, BLASLONG dummy0, BLASLONG dummy1, FLOAT da_r, FLOAT da_i,
}
}
else if (da_r < -FLT_MAX || da_r > FLT_MAX) {
else if (isinf(da_r)) {
while(j < n)
{
x[i]= NAN;
@ -410,7 +410,7 @@ int CNAME(BLASLONG n, BLASLONG dummy0, BLASLONG dummy1, FLOAT da_r, FLOAT da_i,
while(j < n)
{
temp0 = -da_i * x[i+1];
if (x[i] < -FLT_MAX || x[i] > FLT_MAX)
if (isinf(x[i]))
temp0 = NAN;
x[i+1] = da_i * x[i];
if ( x[i] == x[i]) //preserve NaN