ruby-changes:74469
From: Kenta <ko1@a...>
Date: Sun, 13 Nov 2022 15:01:42 +0900 (JST)
Subject: [ruby-changes:74469] 802dce4acd (master): [ruby/bigdecimal] Add and use specific value allocators
https://git.ruby-lang.org/ruby.git/commit/?id=802dce4acd From 802dce4acd97234bef88fe4bbb98db5ab90c45d9 Mon Sep 17 00:00:00 2001 From: Kenta Murata <mrkn@m...> Date: Sun, 13 Nov 2022 14:59:55 +0900 Subject: [ruby/bigdecimal] Add and use specific value allocators * Add rbd_allocate_struct_zero for making 0.0 * Add rbd_allocate_struct_one for making 1.0 * Use them to replace VpAlloc calls * Renmae VpPt5 to VpConstPt5 https://github.com/ruby/bigdecimal/commit/40c826f5e6 --- ext/bigdecimal/bigdecimal.c | 69 ++++++++++++++++++++++++++++++++------------- 1 file changed, 50 insertions(+), 19 deletions(-) diff --git a/ext/bigdecimal/bigdecimal.c b/ext/bigdecimal/bigdecimal.c index bad10ae832..a914d99589 100644 --- a/ext/bigdecimal/bigdecimal.c +++ b/ext/bigdecimal/bigdecimal.c @@ -144,7 +144,8 @@ PUREFUNC(static inline size_t rbd_struct_size(size_t const)); https://github.com/ruby/ruby/blob/trunk/ext/bigdecimal/bigdecimal.c#L144 static inline size_t rbd_struct_size(size_t const internal_digits) { - return offsetof(Real, frac) + sizeof(DECDIG) * internal_digits; + size_t const frac_len = (internal_digits == 0) ? 1 : internal_digits; + return offsetof(Real, frac) + frac_len * sizeof(DECDIG); } static inline Real * @@ -153,6 +154,7 @@ rbd_allocate_struct(size_t const internal_digits) https://github.com/ruby/ruby/blob/trunk/ext/bigdecimal/bigdecimal.c#L154 size_t const size = rbd_struct_size(internal_digits); Real *real = ruby_xcalloc(1, size); atomic_allocation_count_inc(); + real->MaxPrec = internal_digits; return real; } @@ -164,6 +166,7 @@ rbd_reallocate_struct(Real *real, size_t const internal_digits) https://github.com/ruby/ruby/blob/trunk/ext/bigdecimal/bigdecimal.c#L166 size_t const size = rbd_struct_size(internal_digits); VALUE obj = real ? real->obj : 0; Real *new_real = (Real *)ruby_xrealloc(real, size); + new_real->MaxPrec = internal_digits; if (obj) { new_real->obj = 0; BigDecimal_wrap_struct(obj, new_real); @@ -181,6 +184,23 @@ rbd_free_struct(Real *real) https://github.com/ruby/ruby/blob/trunk/ext/bigdecimal/bigdecimal.c#L184 } } +static Real * +rbd_allocate_struct_zero(size_t const digits, int sign) +{ + size_t const len = roomof(digits, BASE_FIG); + Real *real = rbd_allocate_struct(len); + VpSetZero(real, sign); + return real; +} + +static Real * +rbd_allocate_struct_one(size_t const digits, int sign) +{ + Real *real = rbd_allocate_struct_zero(digits, sign); + VpSetOne(real); + return real; +} + /* * ================== Ruby Interface part ========================== */ @@ -4517,7 +4537,7 @@ static int gfCheckVal = 1; /* Value checking flag in VpNmlz() */ https://github.com/ruby/ruby/blob/trunk/ext/bigdecimal/bigdecimal.c#L4537 #endif /* BIGDECIMAL_DEBUG */ static Real *VpConstOne; /* constant 1.0 */ -static Real *VpPt5; /* constant 0.5 */ +static Real *VpConstPt5; /* constant 0.5 */ #define maxnr 100UL /* Maximum iterations for calculating sqrt. */ /* used in VpSqrt() */ @@ -4936,9 +4956,13 @@ VpInit(DECDIG BaseVal) https://github.com/ruby/ruby/blob/trunk/ext/bigdecimal/bigdecimal.c#L4956 /* Setup +/- Inf NaN -0 */ VpGetDoubleNegZero(); - /* Allocates Vp constants. */ - VpConstOne = VpAlloc(1UL, "1", 1, 1); - VpPt5 = VpAlloc(1UL, ".5", 1, 1); + /* Const 1.0 */ + VpConstOne = rbd_allocate_struct_one(1, 1); + + /* Const 0.5 */ + VpConstPt5 = rbd_allocate_struct_one(1, 1); + VpConstPt5->exponent = 0; + VpConstPt5->frac[0] = 5*BASE1; #ifdef BIGDECIMAL_DEBUG gnAlloc = 0; @@ -5838,7 +5862,7 @@ VpMult(Real *c, Real *a, Real *b) https://github.com/ruby/ruby/blob/trunk/ext/bigdecimal/bigdecimal.c#L5862 if (MxIndC < MxIndAB) { /* The Max. prec. of c < Prec(a)+Prec(b) */ w = c; - c = VpAlloc((size_t)((MxIndAB + 1) * BASE_FIG), "#0", 1, 1); + c = rbd_allocate_struct_zero((size_t)((MxIndAB + 1) * BASE_FIG), 1); MxIndC = MxIndAB; } @@ -7003,8 +7027,9 @@ VpSqrt(Real *y, Real *x) https://github.com/ruby/ruby/blob/trunk/ext/bigdecimal/bigdecimal.c#L7027 if (x->MaxPrec > (size_t)n) n = (ssize_t)x->MaxPrec; /* allocate temporally variables */ - f = VpAlloc(y->MaxPrec * (BASE_FIG + 2), "#1", 1, 1); - r = VpAlloc((n + n) * (BASE_FIG + 2), "#1", 1, 1); + /* TODO: reconsider MaxPrec of f and r */ + f = rbd_allocate_struct_one(y->MaxPrec * (BASE_FIG + 2), 1); + r = rbd_allocate_struct_one((n + n) * (BASE_FIG + 2), 1); nr = 0; y_prec = y->MaxPrec; @@ -7029,16 +7054,21 @@ VpSqrt(Real *y, Real *x) https://github.com/ruby/ruby/blob/trunk/ext/bigdecimal/bigdecimal.c#L7054 f->MaxPrec = y->MaxPrec + 1; n = (SIGNED_VALUE)(y_prec * BASE_FIG); if (n < (SIGNED_VALUE)maxnr) n = (SIGNED_VALUE)maxnr; + + /* + * Perform: y_{n+1} = (y_n - x/y_n) / 2 + */ do { - y->MaxPrec *= 2; - if (y->MaxPrec > y_prec) y->MaxPrec = y_prec; - f->MaxPrec = y->MaxPrec; - VpDivd(f, r, x, y); /* f = x/y */ - VpAddSub(r, f, y, -1); /* r = f - y */ - VpMult(f, VpPt5, r); /* f = 0.5*r */ - if (VpIsZero(f)) goto converge; - VpAddSub(r, f, y, 1); /* r = y + f */ - VpAsgn(y, r, 1); /* y = r */ + y->MaxPrec *= 2; + if (y->MaxPrec > y_prec) y->MaxPrec = y_prec; + f->MaxPrec = y->MaxPrec; + VpDivd(f, r, x, y); /* f = x/y */ + VpAddSub(r, f, y, -1); /* r = f - y */ + VpMult(f, VpConstPt5, r); /* f = 0.5*r */ + if (VpIsZero(f)) + goto converge; + VpAddSub(r, f, y, 1); /* r = y + f */ + VpAsgn(y, r, 1); /* y = r */ } while (++nr < n); #ifdef BIGDECIMAL_DEBUG @@ -7455,9 +7485,10 @@ VpPowerByInt(Real *y, Real *x, SIGNED_VALUE n) https://github.com/ruby/ruby/blob/trunk/ext/bigdecimal/bigdecimal.c#L7485 } /* Allocate working variables */ + /* TODO: reconsider MaxPrec of w1 and w2 */ + w1 = rbd_allocate_struct_zero((y->MaxPrec + 2) * BASE_FIG, 1); + w2 = rbd_allocate_struct_zero((w1->MaxPrec * 2 + 1) * BASE_FIG, 1); - w1 = VpAlloc((y->MaxPrec + 2) * BASE_FIG, "#0", 1, 1); - w2 = VpAlloc((w1->MaxPrec * 2 + 1) * BASE_FIG, "#0", 1, 1); /* calculation start */ VpAsgn(y, x, 1); -- cgit v1.2.3 -- ML: ruby-changes@q... Info: http://www.atdot.net/~ko1/quickml/