uint256: optimize AddMod

[AaronChen0]

Test

go test ./...

ok  	github.com/holiman/uint256	1.367s

Benchmark

goos: linux
goarch: amd64
pkg: github.com/holiman/uint256
cpu: AMD Ryzen 7 7735H with Radeon Graphics         
                         │     old     │                 new                 │
                         │   sec/op    │   sec/op     vs base                │
AddMod/small/uint256-16    6.292n ± 1%   6.309n ± 1%        ~ (p=0.123 n=10)
AddMod/mod64/uint256-16    10.07n ± 1%   10.02n ± 1%   -0.55% (p=0.049 n=10)
AddMod/mod128/uint256-16   19.14n ± 1%   19.16n ± 1%        ~ (p=0.271 n=10)
AddMod/mod192/uint256-16   20.88n ± 2%   21.09n ± 1%   +1.03% (p=0.050 n=10)
AddMod/mod256/uint256-16   9.728n ± 2%   6.460n ± 1%  -33.59% (p=0.000 n=10)
AddMod/small/big-16        25.80n ± 0%   25.85n ± 0%        ~ (p=0.071 n=10)
AddMod/mod64/big-16        30.18n ± 1%   30.11n ± 0%        ~ (p=0.224 n=10)
AddMod/mod128/big-16       67.25n ± 1%   68.10n ± 0%   +1.26% (p=0.000 n=10)
AddMod/mod192/big-16       68.95n ± 0%   69.28n ± 0%   +0.47% (p=0.005 n=10)
AddMod/mod256/big-16       68.94n ± 1%   69.47n ± 0%   +0.76% (p=0.005 n=10)
geomean                    23.92n        23.04n        -3.69%

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 100.00%. Comparing base (09331ab) to head (0b682e9).

Additional details and impacted files

@@            Coverage Diff            @@
##            master      #165   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            5         5           
  Lines         1628      1626    -2     
=========================================
- Hits          1628      1626    -2     

[holiman]

Can confirm:

[user@work uint256]$ benchstat addmod.1 addmod.2
name                     old time/op  new time/op  delta
AddMod/small/uint256-8   21.0ns ±44%  17.0ns ±43%     ~     (p=0.052 n=10+10)
AddMod/mod64/uint256-8   28.6ns ±36%  26.6ns ±44%     ~     (p=0.481 n=10+10)
AddMod/mod128/uint256-8  56.1ns ±43%  53.7ns ±49%     ~     (p=0.810 n=10+10)
AddMod/mod192/uint256-8  62.0ns ±44%  62.2ns ±44%     ~     (p=0.971 n=10+10)
AddMod/mod256/uint256-8  25.4ns ±25%  12.0ns ±44%  -52.74%  (p=0.000 n=10+9)

Neat!

[holiman]

Good catch, thanks!