maandree/libzahl

Suckless big integer library

libzahl's Issues

Branch-less macros?

You may be able (probably unlikely) to sip a little more performance if you replace the logical operator in the following (from libzahl/zahl/inlines.h) with a bitwise operator. If you decide to change zzero(), there are other candidates in inlines.h worth reviewing.

ZAHL_INLINE int zzero(z_t a) { return !a->sign; }

... {return a->sign & 0;}
But, I haven't looked at the assembler code.

In the unlikely event that something is likely unlikely.

Can you provide rationale on your choices of likely() and unlikely()?

void
zsub_nonnegative_assign(z_t a, z_t b)
{
if (unlikely(zzero(b)))
zabs(a, a);
else if (unlikely(!zcmpmag(a, b)))
SET_SIGNUM(a, 0);
else
zsub_impl(a, b, b->used);
}

What if I issue zsub_nonnegative_assign(x, "0") in an iterative process? In this example, the unlikely is very, very, very unlikely (zzero(0) is not just likely, but totally likely). (I see that zsub_nonnegative_assign() may be removed.) How do these branch predictions behave if they are more likely to be incorrect than correct for certain use cases? Perhaps you could allow the user of the library a choice on the use of likely()/unlikely().

Benchmark of libzahl in fractions vs mpdecimal

I wonder how fast would it be to use fractions with high(er) number of places with libzahl as "backend" compared to using plain mpdecimal for precise calculations.

In other words, first find out how to set precision of the fractions to match the precision of mpdecimal and then benchmark some real-world (or at least artifiicial) scenarios for speed.

What would be your guess to be faster?

libzahl/src/allocator.c

Please initialize i.

Doesn't Compile with C++ compilers

C++ doesn't support register and restrict.

To fix this we can use macros to only activate these keywordsin C mode and not in C++ mode.

I have a commit ready to fix this. Do you accept pull requests here on github ? Or on suckless.org ?