- fixed a couple of bugs in the process
  (shift right was incorrect for negative numbers)
- added more tests and made some tests more robust
- changed pidigits back to using shifts to multiply
  by 2 instead of add

  This improves pidigit -s -n 10000 by approx. 5%:

  user 0m6.496s (old)
  user 0m6.156s (new)

R=rsc
CC=golang-dev
https://golang.org/cl/963044

R=rsc
CC=golang-dev
https://golang.org/cl/986044

Fixes #589.

R=ken2
CC=golang-dev
https://golang.org/cl/1043042

Fixes #662.

R=ken2
CC=golang-dev
https://golang.org/cl/978043

Fixes #746.

R=ken2
CC=golang-dev
https://golang.org/cl/1040042

Fixes #755.

R=ken2
CC=golang-dev
https://golang.org/cl/965047

R=rsc
CC=golang-dev
https://golang.org/cl/962046

R=rsc
CC=golang-dev
https://golang.org/cl/1044041

This results in an improvement of > 35% for the existing Mul benchmark
using the same karatsuba threshold, and an improvement of > 50% with
a slightly higher threshold (32 instead of 30):

big.BenchmarkMul           500	   6731846 ns/op (old alg.)
big.BenchmarkMul	   500	   4351122 ns/op (new alg.)
big.BenchmarkMul           500	   3133782 ns/op (new alg., new theshold)

Also:
- tweaked calibrate.go, use same benchmark as for Mul benchmark

R=rsc
CC=golang-dev
https://golang.org/cl/1037041

R=rsc
CC=golang-dev
https://golang.org/cl/974043

     provide assembly versions (for x86-64 for now)

(Not yet used - waiting for previous CL to clear)

R=rsc
CC=golang-dev
https://golang.org/cl/1040041

R=rsc
CC=go-dev, golang-dev
https://golang.org/cl/1006045

R=rsc, Joe Poirier
CC=golang-dev
https://golang.org/cl/1015043

Fixes #717.

R=adg, PeterGo
CC=golang-dev
https://golang.org/cl/1032041

R=rsc, brainman
CC=golang-dev
https://golang.org/cl/961047

Fixes #749.

R=ken2
CC=golang-dev
https://golang.org/cl/963043

Fixes #745.

R=ken2
CC=golang-dev
https://golang.org/cl/1008045

R=rsc
CC=golang-dev
https://golang.org/cl/1019043

Fixes #686.

R=rsc
CC=adg, golang-dev
https://golang.org/cl/979044

R=r, rsc
CC=golang-dev
https://golang.org/cl/946046

R=adg
CC=golang-dev
https://golang.org/cl/993042

Also, calling fmt, ioutil and os "builtin" is probably confusing.

R=adg
CC=golang-dev
https://golang.org/cl/965045

Fixes #549.

R=adg
CC=golang-dev
https://golang.org/cl/1019042

Fixes #686.

R=adg
CC=golang-dev
https://golang.org/cl/961046

Fixes #674.

R=adg
CC=golang-dev
https://golang.org/cl/1018042

inaccurate in meaning.

R=gri
CC=golang-dev
https://golang.org/cl/964045

R=r
CC=golang-dev, rsc
https://golang.org/cl/1014043

earlier.

R=rsc
CC=golang-dev
https://golang.org/cl/972045

5l, 6l, 8l: change ELF header so that strip doesn't destroy binary

Fixes #261.

R=iant, r
CC=golang-dev
https://golang.org/cl/994044

R=r
CC=golang-dev
https://golang.org/cl/972044

R=r, rsc
CC=golang-dev
https://golang.org/cl/963042

noticed this as I was submitting the previous CL :(

R=r
CC=golang-dev
https://golang.org/cl/989045

R=r, rsc, gri
CC=golang-dev
https://golang.org/cl/887045

Plus:
- calibration "test" - include in tests with gotest -calibrate
- basic Mul benchmark
- extra multiplication tests
- various cleanups

This change improves multiplication speed of numbers >= 30 words
in length (current threshold; found empirically with calibrate):

The multiplication benchmark (multiplication of a variety of long numbers)
improves by ~35%, individual multiplies can be significantly faster.

gotest -benchmarks=Mul
big.BenchmarkMul	     500	   6829290 ns/op (w/ Karatsuba)
big.BenchmarkMul	     100	  10600760 ns/op

There's no impact on pidigits for -n=10000 or -n=20000
because the operands are are too small.

R=rsc
CC=golang-dev
https://golang.org/cl/1004042

Fixes #610.

R=adg
CC=golang-dev
https://golang.org/cl/1006044

Fixes #483.

R=dsymonds, dsymonds1
CC=golang-dev
https://golang.org/cl/965044

Fixes #750.

R=adg
CC=golang-dev
https://golang.org/cl/988043

Fixes #743.

R=r, rsc
CC=golang-dev
https://golang.org/cl/944044

R=r
CC=golang-dev
https://golang.org/cl/1013042

R=ken2
CC=golang-dev
https://golang.org/cl/944043