Skip to content

Calculate division and multiplication with Ruby's Integer for large precision #380

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 1 commit into
base: master
Choose a base branch
from
Draft

Calculate division and multiplication with Ruby's Integer for large precision #380

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
79 changes: 79 additions & 0 deletions ext/bigdecimal/bigdecimal.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -145,6 +145,9 @@ bdvalue_nullable(BDVALUE v)
#define BIGDECIMAL_POSITIVE_P(bd) ((bd)->sign > 0)
#define BIGDECIMAL_NEGATIVE_P(bd) ((bd)->sign < 0)

#define MULT_BY_RB_INTEGER_THRESHOLD 150
#define DIV_BY_RB_INTEGER_THRESHOLD 300

/*
* ================== Memory allocation ============================
*/
Expand Down Expand Up @@ -5873,6 +5876,32 @@ VpSetPTR(Real *a, Real *b, Real *c, size_t *a_pos, size_t *b_pos, size_t *c_pos,
return word_shift;
}

static int
VpMultWithRubyInteger(Real *c, Real *a, Real *b)
{
Real *ap, *bp;
VALUE a2, b2, ab;
BDVALUE c2;
ap = VpCopy(NULL, a);
bp = VpCopy(NULL, b);
ap->exponent = a->Prec;
bp->exponent = b->Prec;
a2 = TypedData_Wrap_Struct(rb_cBigDecimal, &BigDecimal_data_type, 0);
b2 = TypedData_Wrap_Struct(rb_cBigDecimal, &BigDecimal_data_type, 0);
BigDecimal_wrap_struct(a2, ap);
BigDecimal_wrap_struct(b2, bp);
ab = rb_funcall(BigDecimal_to_i(a2), '*', 1, BigDecimal_to_i(b2));

c2 = GetBDValueMust(ab);
VpAsgn(c, c2.real, 1);
if (!AddExponent(c, a->exponent)) return 0;
if (!AddExponent(c, b->exponent)) return 0;
if (!AddExponent(c, -a->Prec)) return 0;
if (!AddExponent(c, -b->Prec)) return 0;
RB_GC_GUARD(c2.bigdecimal);
return 1;
}

/*
* Return number of significant digits
* c = a * b , Where a = a0a1a2 ... an
Expand Down Expand Up @@ -5926,6 +5955,11 @@ VpMult(Real *c, Real *a, Real *b)
MxIndC = c->MaxPrec - 1;
MxIndAB = a->Prec + b->Prec - 1;

if (a->Prec >= MULT_BY_RB_INTEGER_THRESHOLD && b->Prec >= MULT_BY_RB_INTEGER_THRESHOLD) {
if (!VpMultWithRubyInteger(c, a, b)) return 0;
goto Exit;
}

if (MxIndC < MxIndAB) { /* The Max. prec. of c < Prec(a)+Prec(b) */
w = c;
c = NewZeroNolimit(1, (size_t)((MxIndAB + 1) * BASE_FIG));
Expand Down Expand Up @@ -6000,6 +6034,46 @@ VpMult(Real *c, Real *a, Real *b)
return c->Prec*BASE_FIG;
}

static void
VpDivdWithRubyInteger(Real *c, Real *r, Real *a, Real *b)
{
Real *ap, *bp;
BDVALUE c2, r2;
VALUE a2, b2, divmod, div, mod;
size_t div_prec = c->MaxPrec - 1;
size_t base_prec = b->Prec;

ap = VpCopy(NULL, a);
bp = VpCopy(NULL, b);
VpSetSign(ap, 1);
VpSetSign(bp, 1);
ap->exponent = base_prec + div_prec;
bp->exponent = base_prec;
a2 = TypedData_Wrap_Struct(rb_cBigDecimal, &BigDecimal_data_type, 0);
b2 = TypedData_Wrap_Struct(rb_cBigDecimal, &BigDecimal_data_type, 0);
BigDecimal_wrap_struct(a2, ap);
BigDecimal_wrap_struct(b2, bp);
divmod = rb_funcall(BigDecimal_to_i(a2), rb_intern("divmod"), 1, BigDecimal_to_i(b2));
if (RB_TYPE_P(divmod, T_ARRAY) && RARRAY_LEN(divmod) == 2) {
div = RARRAY_AREF(divmod, 0);
mod = RARRAY_AREF(divmod, 1);
} else {
div = mod = Qnil;
}

c2 = GetBDValueMust(div);
r2 = GetBDValueMust(mod);
VpAsgn(c, c2.real, VpGetSign(a) * VpGetSign(b));
VpAsgn(r, r2.real, VpGetSign(a));
RB_GC_GUARD(c2.bigdecimal);
RB_GC_GUARD(r2.bigdecimal);
AddExponent(c, a->exponent);
AddExponent(c, -b->exponent);
AddExponent(c, -div_prec);
AddExponent(r, a->exponent);
AddExponent(r, -base_prec - div_prec);
}

/*
* c = a / b, remainder = r
* XXXX_YYYY_ZZZZ / 0001 = XXXX_YYYY_ZZZZ
Expand Down Expand Up @@ -6041,6 +6115,11 @@ VpDivd(Real *c, Real *r, Real *a, Real *b)

if (word_a > word_r || word_b + word_c - 2 >= word_r) goto space_error;

if (word_c >= DIV_BY_RB_INTEGER_THRESHOLD && word_b >= DIV_BY_RB_INTEGER_THRESHOLD) {
VpDivdWithRubyInteger(c, r, a, b);
goto Exit;
}

for (i = 0; i < word_a; ++i) r->frac[i] = a->frac[i];
for (i = word_a; i < word_r; ++i) r->frac[i] = 0;
for (i = 0; i < word_c; ++i) c->frac[i] = 0;
Expand Down
44 changes: 41 additions & 3 deletions test/bigdecimal/test_vp_operation.rb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,9 @@
class TestVpOperation < Test::Unit::TestCase
include TestBigDecimalBase

INTEGER_MULT_THRESHOLD = 150
INTEGER_DIV_THRESHOLD = 300

def setup
super
unless BigDecimal.instance_methods.include?(:vpdivd)
Expand All @@ -16,9 +19,15 @@ def setup
def test_vpmult
assert_equal(BigDecimal('121932631112635269'), BigDecimal('123456789').vpmult(BigDecimal('987654321')))
assert_equal(BigDecimal('12193263.1112635269'), BigDecimal('123.456789').vpmult(BigDecimal('98765.4321')))
x = 123**456
y = 987**123
assert_equal(BigDecimal("#{x * y}e-300"), BigDecimal("#{x}e-100").vpmult(BigDecimal("#{y}e-200")))

[10, 50, INTEGER_MULT_THRESHOLD, INTEGER_MULT_THRESHOLD + 50].each do |prec|
x = (BASE + BASE / 7) ** prec
y = (BASE + BASE / 3) ** prec
assert_equal(BigDecimal("#{x * y}e-168"), BigDecimal("#{x}e-123").vpmult(BigDecimal("#{y}e-45")))
assert_equal(BigDecimal("-#{x * y}e-46"), BigDecimal("#{x}e-12").vpmult(BigDecimal("-#{y}e-34")))
assert_equal(BigDecimal("-#{x * y}e-333"), BigDecimal("-#{x}e12").vpmult(BigDecimal("#{y}e-345")))
assert_equal(BigDecimal("#{x * y}e46"), BigDecimal("-#{x}e12").vpmult(BigDecimal("-#{y}e34")))
end
end

def test_vpdivd
Expand Down Expand Up @@ -84,6 +93,35 @@ def test_vpdivd_precisions
end
end

def test_vpdivd_by_ruby_integer
# Exponent check
integer_div_prec = INTEGER_DIV_THRESHOLD + 50
x = (BASE + BASE / 7) ** integer_div_prec
y = (BASE + BASE / 3) ** integer_div_prec
bx = BigDecimal("#{x}e-123")
by = BigDecimal("#{y}e-456")
div, mod = bx.vpdivd(by, integer_div_prec)
assert_include(0...by, mod)
assert_equal(bx, div * by + mod)

# Precision should consistent around DIV_BY_RB_INTEGER_THRESHOLD threshold
[2, 3, 4, integer_div_prec, integer_div_prec + 1, integer_div_prec + 2].each do |prec|
a = BigDecimal('1' + '2' * BASE_FIG * prec)
b = BigDecimal('9' * BASE_FIG + '9' * BASE_FIG * prec)
assert_equal(BASE_FIG * (prec - 2) + 1, a.vpdivd(b, prec).first.n_significant_digits)
assert_equal(BASE_FIG * prec, b.vpdivd(a, prec).first.n_significant_digits)

x = BigDecimal("1#{'0' * BASE_FIG * prec}1")
y = BigDecimal("1#{'0' * BASE_FIG * prec}2")
div = BigDecimal("0.#{'9' * BASE_FIG * (prec - 1)}")
mod = BigDecimal("#{'9' * (BASE_FIG + 1)}.#{'0' * (BASE_FIG * (prec - 1) - 1)}2")
assert_equal([div, mod], x.vpdivd(y, prec))
assert_equal([-div, mod], x.vpdivd(-y, prec))
assert_equal([-div, -mod], (-x).vpdivd(y, prec))
assert_equal([div, -mod], (-x).vpdivd(-y, prec))
end
end

def test_vpdivd_borrow
y_small = BASE / 7 * BASE ** 4
y_large = (4 * BASE_FIG).times.map {|i| i % 9 + 1 }.join.to_i
Expand Down
Loading