Introduction
Introduction Statistics Contact Development Disclaimer Help
TODO - libzahl - big integer library
git clone git://git.suckless.org/libzahl
Log
Files
Refs
README
LICENSE
---
TODO (1598B)
---
1 GMP has mpz_divexact(q,n,d), we should have zdiv_exact(q,n,d).
2 It uses optimised division algorithm that requires that d|n.
3
4 Add zsets_radix
5 Add zstr_radix
6
7 Can zmodpowu and zmodpow be improved using some other algorithm?
8 Is it worth implementing precomputed optimal
9 addition-chain exponentiation in zpowu?
10
11 Test big endian
12 Test always having .used > 0 for zero
13 Test negative/non-negative instead of sign
14 Test long .sign
15 Test always having .chars % 4 == 0
16 Test reusing objects in the temp-stack
17
18 Test optimisation of zmul:
19 bc = [(Hb * Hc) << (m2 << 1)]
20 + [(Hb * Hc) << m2]
21 - [(Hb - Lb)(Hc - Lc) << m2]
22 + [(Lb * Lc) << m2]
23 + (Lb * Lc)
24
25 Would zmul be faster if we split only one of the
26 factors until they are both approximately the same
27 size?
28
29 Add entropy test for zrand.
30
31 Should zmodpowu, zmodpow, zmodmul, and zmodsqr be removed?
32 I need to research how important these are.
33 They are important for cryptography, but we do not care about that.
34 They are important for discrete/abstract mathematics, but bignum proba…
35
36 Add CPU-warmup loop to benchmarks.
37 If ondemand scaling is available but not set, set it.
38 If the current frequency is not the minimum, run a
39 catch-fire loop until the CPU is throttled to the
40 minimum frequency.
41 This loop shall be done after help variables have
42 been assigned values, as this can help the warmup.
43
44 benchmark with worst case, average case, and best case input.
45
46 zadd, zsub: benchmark both dense and sparse integers.
47
48 Feedback on error handling:
49 http://bbs.progrider.org/prog/read/1457215529/31,47
You are viewing proxied material from suckless.org. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.