ALTree/bigfloat

Issues

• sqrt: switch to the stdlib Sqrt and tag 0.2

  #34 opened a year ago by ALTree
  1

• all: tag version 0.1

  #33 opened a year ago by ALTree
  2

• Panic in Pow for negative base with integer exponent

  #35 opened 6 years ago by vsivsi
  1

• Pow(Inf, -1) == 0 not Inf. True for all negative exponents

  #36 opened 6 years ago by vsivsi
  2

• Not very accurate with numbers close to 1

  #32 opened 6 years ago by BenthamEJ
  1

• pow.go: fastpath for Pow(x, 0.5) using Sqrt

  #31 opened 8 years ago by ALTree
  0

• all: switch to sub{test, benchmarks}

  #30 opened 8 years ago by ALTree
  0

• sqrt.go: remove useless x decl in Sqrt

  #29 opened 8 years ago by ALTree
  0

• all: check alias on Sub and Add calls

  #28 opened 8 years ago by ALTree
  0

• misc.go: a big pi cache slows down log for low precs

  #8 opened 9 years ago by ALTree
  0

• all: unifom prec use

  #25 opened 9 years ago by ALTree
  0

• misc.go: make agm reuse its arguments

  #27 opened 8 years ago by ALTree
  1

• log.go: small optimizations

  #26 opened 8 years ago by ALTree
  0

• misc.go: clean up pi and agm

  #22 opened 9 years ago by ALTree
  0

• readme.md: extend example

  #24 opened 9 years ago by ALTree
  0

• misc.go: merge newton and newton2

  #20 opened 9 years ago by ALTree
  0

• exp.go: change exp to use newton

  #21 opened 9 years ago by ALTree
  0

• misc.go: tune newton's last step, no need to overdo it

  #23 opened 9 years ago by ALTree
  0

• sqrt.go: figure out if we can do computations with reduced prec in newton iterations

  #7 opened 9 years ago by ALTree
  0

• misc.go: general newton function

  #11 opened 9 years ago by ALTree
  0

• all: fill the pi cache before benchmarking anything

  #12 opened 9 years ago by ALTree
  0

• sqrt.go: test zfs == 0 path

  #9 opened 9 years ago by ALTree
  1

• all: uniform names

  #17 opened 9 years ago by ALTree
  0

• misc.go: no need to SetPrec on constants

  #5 opened 9 years ago by ALTree
  0

• all: check we do not modify args by mistake

  #16 opened 9 years ago by ALTree
  0

• all: test special cases carefully

  #18 opened 9 years ago by ALTree
  0

• all: increase accuracy required in IEEE tests

  #19 opened 9 years ago by ALTree
  0

• log_test.go: figure out a good permitted error for TestLog64

  #6 opened 9 years ago by ALTree
  1

• all: implement x ** y

  #13 opened 9 years ago by ALTree
  0

• all: uniform IEEE-754 tests

  #14 opened 9 years ago by ALTree
  0

• sqrt.go: test with very big z (causing small 1/sqrt(z) approx)

  #10 opened 9 years ago by ALTree
  0

• all: better checks around initial IEEE-754 estimates

  #15 opened 9 years ago by ALTree
  0

• misc.go: always cache pi

  #4 opened 9 years ago by ALTree
  0

• Serious testing and benchmarking of agm( ) and pi( )

  #2 opened 9 years ago by ALTree
  0

• Investigate computing sqrt via 1/sqrt

  #3 opened 9 years ago by ALTree
  0

• misc.go: agm computation is off for very big floats

  #1 opened 9 years ago by ALTree
  0