[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Guile-commits] 23/27: Fix slot representation computation for fadd, fmu
From: |
Andy Wingo |
Subject: |
[Guile-commits] 23/27: Fix slot representation computation for fadd, fmul, etc |
Date: |
Wed, 11 Nov 2015 11:39:15 +0000 |
wingo pushed a commit to branch master
in repository guile.
commit f0594be035ebc53813a9a4c5d09cf8a3e61c8835
Author: Andy Wingo <address@hidden>
Date: Thu Oct 29 14:06:11 2015 +0000
Fix slot representation computation for fadd, fmul, etc
* module/language/cps/slot-allocation.scm (compute-var-representations):
fadd, fmul and so on also define f64 values.
---
module/language/cps/slot-allocation.scm | 3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/module/language/cps/slot-allocation.scm
b/module/language/cps/slot-allocation.scm
index 6fc2a53..8d865d7 100644
--- a/module/language/cps/slot-allocation.scm
+++ b/module/language/cps/slot-allocation.scm
@@ -790,7 +790,8 @@ are comparable with eqv?. A tmp slot may be used."
(($ $values (arg))
(intmap-add representations var
(intmap-ref representations arg)))
- (($ $primcall (or 'scm->f64 'bv-f32-ref 'bv-f64-ref))
+ (($ $primcall (or 'scm->f64 'bv-f32-ref 'bv-f64-ref
+ 'fadd 'fsub 'fmul 'fdiv))
(intmap-add representations var 'f64))
(_
(intmap-add representations var 'scm))))
- [Guile-commits] 05/27: Fix slot-allocation to make 'return' not need to alloc-frame, (continued)
- [Guile-commits] 05/27: Fix slot-allocation to make 'return' not need to alloc-frame, Andy Wingo, 2015/11/11
- [Guile-commits] 07/27: Don't emit redundant reset-frame before return, Andy Wingo, 2015/11/11
- [Guile-commits] 11/27: Remove use of return in disassembler.scm, Andy Wingo, 2015/11/11
- [Guile-commits] 10/27: rtl.test uses return-values, Andy Wingo, 2015/11/11
- [Guile-commits] 13/27: Treat tail $values as generating lazy allocations, Andy Wingo, 2015/11/11
- [Guile-commits] 14/27: VM support for raw slots, Andy Wingo, 2015/11/11
- [Guile-commits] 15/27: Reflection support for unboxed f64 slots, Andy Wingo, 2015/11/11
- [Guile-commits] 06/27: return-values opcode resets the frame, Andy Wingo, 2015/11/11
- [Guile-commits] 20/27: Scalar replacement for f64->scm, Andy Wingo, 2015/11/11
- [Guile-commits] 21/27: Add fadd, fsub, fmul, fdiv instructions, Andy Wingo, 2015/11/11
- [Guile-commits] 23/27: Fix slot representation computation for fadd, fmul, etc,
Andy Wingo <=
- [Guile-commits] 19/27: bv-{f32, f64}-{ref, set!} operate on raw f64 values, Andy Wingo, 2015/11/11
- [Guile-commits] 25/27: Better f64 unboxing for loop vars that might flow to $ktail, Andy Wingo, 2015/11/11
- [Guile-commits] 17/27: Add VM ops to pack and unpack raw f64 values., Andy Wingo, 2015/11/11
- [Guile-commits] 03/27: CSE can run on first-order CPS, Andy Wingo, 2015/11/11
- [Guile-commits] 16/27: Stack slots can hold a double, Andy Wingo, 2015/11/11
- [Guile-commits] 12/27: Remove return opcode, Andy Wingo, 2015/11/11
- [Guile-commits] 09/27: Always emit return-values, Andy Wingo, 2015/11/11
- [Guile-commits] 08/27: Replace return primcalls with $values, Andy Wingo, 2015/11/11
- [Guile-commits] 24/27: The compiler can unbox float64 loop variables, Andy Wingo, 2015/11/11
- [Guile-commits] 26/27: Remove debug printout in specialize-numbers, Andy Wingo, 2015/11/11
- Prev by Date:
[Guile-commits] 21/27: Add fadd, fsub, fmul, fdiv instructions
- Next by Date:
[Guile-commits] 19/27: bv-{f32, f64}-{ref, set!} operate on raw f64 values
- Previous by thread:
[Guile-commits] 21/27: Add fadd, fsub, fmul, fdiv instructions
- Next by thread:
[Guile-commits] 19/27: bv-{f32, f64}-{ref, set!} operate on raw f64 values
- Index(es):