[Tim]
> BTW, if we know we're not interested in anything other than
> 32-bit products,
>
> double _prod = (double)src1 * (double)src2;
> if (_prod < 4294967296.0) /* 2**32 */
> result = (size_t)_prod;
> else
> overflow;
[Oren Tirosh]
> s/_prod;/src1 * src2;/
>
> Re-calculating the product with integer multiplication is probably faster
> than converting a double to integer on any sane CPU.
This is so. Better, because it allows more instruction-level parallelism:
double dprod = (double)src1 * (double)src2;
size_t iprod = src1 * src2;
if (dprod < 4294967296.0) /* 2**32 */
result = iprod;
else
/* overflow */
> I wonder if the old SPARCs that don't have an integer multiply
> instruction actually have a floating point to integer conversion
> that is faster than integer mul.
I'll let someone with an old SPARC worry about that.
> Wasting "expensive" multiplications this way still feels blasphemous :)
Well, we could simulate multiplication by hand, with a bit-at-a-time
shift-and-add loop. This reduces the detect-mul-overflow case to the
detect-add-overflow case. Bonus: on some odd boxes, it would be faster
<wink/sigh>.
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4