[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Guile-commits] 25/41: Add unsigned 64-bit arithmetic operators: uadd, u
From: |
Andy Wingo |
Subject: |
[Guile-commits] 25/41: Add unsigned 64-bit arithmetic operators: uadd, usub, umul |
Date: |
Wed, 02 Dec 2015 08:06:54 +0000 |
wingo pushed a commit to branch master
in repository guile.
commit d294d5d1e19f589dd910ec269ef360484ad754e5
Author: Andy Wingo <address@hidden>
Date: Fri Nov 20 10:58:21 2015 +0100
Add unsigned 64-bit arithmetic operators: uadd, usub, umul
* libguile/vm-engine.c (uadd, usub, umul): New ops.
* module/language/cps/effects-analysis.scm (uadd, usub, umul): Add
effects analysis.
* module/language/cps/slot-allocation.scm (compute-var-representations):
The new ops define 'u64 values.
* module/language/cps/types.scm (uadd, usub, umul): Add type checkers
and inferrers.
* module/system/vm/assembler.scm (emit-uadd, emit-usub, emit-umul): New
assemblers.
---
libguile/vm-engine.c | 45 ++++++++++++++++++++++++++++--
module/language/cps/effects-analysis.scm | 3 ++
module/language/cps/slot-allocation.scm | 3 +-
module/language/cps/types.scm | 21 ++++++++++++++
module/system/vm/assembler.scm | 3 ++
5 files changed, 71 insertions(+), 4 deletions(-)
diff --git a/libguile/vm-engine.c b/libguile/vm-engine.c
index e7994cd..d615af1 100644
--- a/libguile/vm-engine.c
+++ b/libguile/vm-engine.c
@@ -3402,9 +3402,48 @@ VM_NAME (scm_i_thread *thread, struct scm_vm *vp,
BR_U64_ARITHMETIC (<=, scm_leq_p);
}
- VM_DEFINE_OP (149, unused_149, NULL, NOP)
- VM_DEFINE_OP (150, unused_150, NULL, NOP)
- VM_DEFINE_OP (151, unused_151, NULL, NOP)
+ /* uadd dst:8 a:8 b:8
+ *
+ * Add A to B, and place the result in DST. The operands and the
+ * result are unboxed unsigned 64-bit integers. Overflow will wrap
+ * around.
+ */
+ VM_DEFINE_OP (149, uadd, "uadd", OP1 (X8_S8_S8_S8) | OP_DST)
+ {
+ scm_t_uint8 dst, a, b;
+ UNPACK_8_8_8 (op, dst, a, b);
+ SP_SET_U64 (dst, SP_REF_U64 (a) + SP_REF_U64 (b));
+ NEXT (1);
+ }
+
+ /* usub dst:8 a:8 b:8
+ *
+ * Subtract B from A, and place the result in DST. The operands and
+ * the result are unboxed unsigned 64-bit integers. Overflow will
+ * wrap around.
+ */
+ VM_DEFINE_OP (150, usub, "usub", OP1 (X8_S8_S8_S8) | OP_DST)
+ {
+ scm_t_uint8 dst, a, b;
+ UNPACK_8_8_8 (op, dst, a, b);
+ SP_SET_U64 (dst, SP_REF_U64 (a) - SP_REF_U64 (b));
+ NEXT (1);
+ }
+
+ /* umul dst:8 a:8 b:8
+ *
+ * Multiply A and B, and place the result in DST. The operands and
+ * the result are unboxed unsigned 64-bit integers. Overflow will
+ * wrap around.
+ */
+ VM_DEFINE_OP (151, umul, "umul", OP1 (X8_S8_S8_S8) | OP_DST)
+ {
+ scm_t_uint8 dst, a, b;
+ UNPACK_8_8_8 (op, dst, a, b);
+ SP_SET_U64 (dst, SP_REF_U64 (a) * SP_REF_U64 (b));
+ NEXT (1);
+ }
+
VM_DEFINE_OP (152, unused_152, NULL, NOP)
VM_DEFINE_OP (153, unused_153, NULL, NOP)
VM_DEFINE_OP (154, unused_154, NULL, NOP)
diff --git a/module/language/cps/effects-analysis.scm
b/module/language/cps/effects-analysis.scm
index fc82293..9112c42 100644
--- a/module/language/cps/effects-analysis.scm
+++ b/module/language/cps/effects-analysis.scm
@@ -425,6 +425,9 @@ is or might be a read or a write to the same location as A."
((fsub . _))
((fmul . _))
((fdiv . _))
+ ((uadd . _))
+ ((usub . _))
+ ((umul . _))
((sub1 . _) &type-check)
((add1 . _) &type-check)
((quo . _) &type-check)
diff --git a/module/language/cps/slot-allocation.scm
b/module/language/cps/slot-allocation.scm
index e540780..e8519f0 100644
--- a/module/language/cps/slot-allocation.scm
+++ b/module/language/cps/slot-allocation.scm
@@ -793,7 +793,8 @@ are comparable with eqv?. A tmp slot may be used."
(($ $primcall (or 'scm->f64 'bv-f32-ref 'bv-f64-ref
'fadd 'fsub 'fmul 'fdiv))
(intmap-add representations var 'f64))
- (($ $primcall (or 'scm->u64 'bv-length))
+ (($ $primcall (or 'scm->u64 'bv-length
+ 'uadd 'usub 'umul))
(intmap-add representations var 'u64))
(_
(intmap-add representations var 'scm))))
diff --git a/module/language/cps/types.scm b/module/language/cps/types.scm
index 81d2eb1..41d4f56 100644
--- a/module/language/cps/types.scm
+++ b/module/language/cps/types.scm
@@ -916,6 +916,7 @@ minimum, and maximum."
(define-simple-type-checker (add &number &number))
(define-type-checker (fadd a b) #t)
+(define-type-checker (uadd a b) #t)
(define-type-inferrer (add a b result)
(define-binary-result! a b result #t
(+ (&min a) (&min b))
@@ -924,9 +925,16 @@ minimum, and maximum."
(define! result &f64
(+ (&min a) (&min b))
(+ (&max a) (&max b))))
+(define-type-inferrer (uadd a b result)
+ ;; Handle wraparound.
+ (let ((max (+ (&max a) (&max b))))
+ (if (<= max #xffffffffffffffff)
+ (define! result &u64 (+ (&min a) (&min b)) max)
+ (define! result &u64 0 #xffffffffffffffff))))
(define-simple-type-checker (sub &number &number))
(define-type-checker (fsub a b) #t)
+(define-type-checker (usub a b) #t)
(define-type-inferrer (sub a b result)
(define-binary-result! a b result #t
(- (&min a) (&max b))
@@ -935,9 +943,16 @@ minimum, and maximum."
(define! result &f64
(- (&min a) (&max b))
(- (&max a) (&min b))))
+(define-type-inferrer (usub a b result)
+ ;; Handle wraparound.
+ (let ((min (- (&min a) (&max b))))
+ (if (< min 0)
+ (define! result &u64 0 #xffffffffffffffff)
+ (define! result &u64 min (- (&max a) (&min b))))))
(define-simple-type-checker (mul &number &number))
(define-type-checker (fmul a b) #t)
+(define-type-checker (umul a b) #t)
(define (mul-result-range same? nan-impossible? min-a max-a min-b max-b)
(define (nan* a b)
(if (and (or (and (inf? a) (zero? b))
@@ -980,6 +995,12 @@ minimum, and maximum."
min-a max-a min-b max-b))
(lambda (min max)
(define! result &f64 min max)))))
+(define-type-inferrer (umul a b result)
+ ;; Handle wraparound.
+ (let ((max (* (&max a) (&max b))))
+ (if (<= max #xffffffffffffffff)
+ (define! result &u64 (* (&min a) (&min b)) max)
+ (define! result &u64 0 #xffffffffffffffff))))
(define-type-checker (div a b)
(and (check-type a &number -inf.0 +inf.0)
diff --git a/module/system/vm/assembler.scm b/module/system/vm/assembler.scm
index 0ee3918..76ae892 100644
--- a/module/system/vm/assembler.scm
+++ b/module/system/vm/assembler.scm
@@ -142,6 +142,9 @@
(emit-fsub* . emit-fsub)
(emit-fmul* . emit-fmul)
(emit-fdiv* . emit-fdiv)
+ (emit-uadd* . emit-uadd)
+ (emit-usub* . emit-usub)
+ (emit-umul* . emit-umul)
(emit-logand* . emit-logand)
(emit-logior* . emit-logior)
(emit-logxor* . emit-logxor)
- [Guile-commits] 14/41: VM traps don't match on value of slot 0, (continued)
- [Guile-commits] 14/41: VM traps don't match on value of slot 0, Andy Wingo, 2015/12/02
- [Guile-commits] 17/41: Remove frame->module, Andy Wingo, 2015/12/02
- [Guile-commits] 19/41: Add bv-length instruction, Andy Wingo, 2015/12/02
- [Guile-commits] 18/41: Range inference over the full U64+S64 range, Andy Wingo, 2015/12/02
- [Guile-commits] 20/41: bv-f{32, 64}-{ref, set!} take unboxed u64 index, Andy Wingo, 2015/12/02
- [Guile-commits] 23/41: Beginning of u64 phi unboxing, Andy Wingo, 2015/12/02
- [Guile-commits] 16/41: Add low-level support for unboxed 64-bit unsigned ints, Andy Wingo, 2015/12/02
- [Guile-commits] 22/41: Specialize u64 comparisons, Andy Wingo, 2015/12/02
- [Guile-commits] 26/41: Slower range saturation in type inference, Andy Wingo, 2015/12/02
- [Guile-commits] 21/41: Add instructions to branch on u64 comparisons, Andy Wingo, 2015/12/02
- [Guile-commits] 25/41: Add unsigned 64-bit arithmetic operators: uadd, usub, umul,
Andy Wingo <=
- [Guile-commits] 24/41: Unbox u64 phi values, Andy Wingo, 2015/12/02
- [Guile-commits] 28/41: Specialize u64 arithmetic, Andy Wingo, 2015/12/02
- [Guile-commits] 37/41: Disable warnings on bootstrap build, Andy Wingo, 2015/12/02
- [Guile-commits] 35/41: Add current-thread VM op, Andy Wingo, 2015/12/02
- [Guile-commits] 27/41: Better range inference for indexes of vector-ref, string-ref et al, Andy Wingo, 2015/12/02
- [Guile-commits] 29/41: Remove add1 and sub1, Andy Wingo, 2015/12/02
- [Guile-commits] 30/41: Add tagged and untagged arithmetic ops with immediate operands, Andy Wingo, 2015/12/02
- [Guile-commits] 32/41: Add support for unboxed s64 values, Andy Wingo, 2015/12/02
- [Guile-commits] 39/41: Specialize u64 bit operations, Andy Wingo, 2015/12/02
- [Guile-commits] 31/41: New instructions load-f64, load-u64, Andy Wingo, 2015/12/02