Self numbers: Difference between revisions
Content added Content deleted
m (→{{header|Phix}}: s->sv) |
m (→{{header|Phix}}: added escuse) |
||
Line 508: | Line 508: | ||
{{trans|Go}} |
{{trans|Go}} |
||
Replacing the problematic sv[a+b+... line with a bit of dirty inline assembly improved performance by 90%<br> |
Replacing the problematic sv[a+b+... line with a bit of dirty inline assembly improved performance by 90%<br> |
||
(Of course you lose bounds checking, type checking, negative subscripts, fraction handling, and all that jazz.)<br> |
|||
We use a string of Y/N for the sieve to force one byte per element ('\0' and 1 would be equally valid). |
We use a string of Y/N for the sieve to force one byte per element ('\0' and 1 would be equally valid). |
||
<lang Phix>if machine_bits()=32 then crash("requires 64 bit") end if |
<lang Phix>if machine_bits()=32 then crash("requires 64 bit") end if |