From d2af79a77ed2936ff0ed90cadf8e48637d774d4c Mon Sep 17 00:00:00 2001 From: Xavier Leroy Date: Tue, 4 Oct 2016 15:52:16 +0200 Subject: Turn 64-bit integer division and modulus by constants into multiply-high This trick was already implemented for 32-bit integer division and modulus. Here we extend it to the 64-bit case. For 32-bit target processors, the runtime library must implement 64-bit multiply-high (signed and unsigned). Tentative implementations are provided for IA32 and PowerPC, but need testing. --- runtime/powerpc/i64_smul.s | 76 ++++++++++++++++++++++++++++++++++++++++++++++ runtime/powerpc/i64_umul.s | 64 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 140 insertions(+) create mode 100644 runtime/powerpc/i64_smul.s create mode 100644 runtime/powerpc/i64_umul.s (limited to 'runtime/powerpc') diff --git a/runtime/powerpc/i64_smul.s b/runtime/powerpc/i64_smul.s new file mode 100644 index 00000000..9eb453d4 --- /dev/null +++ b/runtime/powerpc/i64_smul.s @@ -0,0 +1,76 @@ +# ***************************************************************** +# +# The Compcert verified compiler +# +# Xavier Leroy, INRIA Paris +# +# Copyright (c) 2016 Institut National de Recherche en Informatique et +# en Automatique. +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are met: +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# * Neither the name of the nor the +# names of its contributors may be used to endorse or promote products +# derived from this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, +# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, +# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR +# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF +# LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING +# NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +# +# ********************************************************************* + +# Helper functions for 64-bit integer arithmetic. PowerPC version. + + .text + +# Signed multiply high + +# Reference C implementation in ../c/i64_smul.c + + .balign 16 + .globl __i64_smulh +__i64_smulh: + # u1 in r3; u0 in r4; v1 in r5; v0 in r6 + # First compute unsigned product (see i64_umul.s) + mulhwu r0, r4, r6 # k (in r0) = high((u64) u0 * (u64) v0) + mullw r8, r3, r6 + mulhwu r7, r3, r6 # t (in r8:r7) = (u64) u1 * (u64) v0 + addc r0, r8, r0 # w1 (in r0) = low (t + k) + addze r9, r7 # w2 (in r9) = high (t + k) + mullw r8, r4, r5 + mulhwu r7, r4, r5 # t (in r8:r7) = (u64) u0 * (u64) v1 + addc r0, r8, r0 # tmp (in r0) = low (t + w1) + addze r0, r7 # k (in r0) = high(t + w1) + mullw r8, r3, r5 + mulhwu r7, r3, r5 # t (in r8:r7) = (u64) u1 * (u64) v1 + addc r8, r8, r9 # add w2 + addze r7, r7 + addc r8, r8, r0 # add k + addze r7, r7 + # Here r8:r7 contains the high 64 bits of the unsigned product + srawi r0, r3, 31 # r0 = 0 if U >= 0, -1 if U < 0 + srawi r9, r5, 31 # r9 = 0 if V >= 0, -1 if V < 0 + and r3, r3, r9 + and r4, r4, r9 # r3:r4 = U if V < 0, = 0 if V >= 0 + and r5, r5, r0 + and r6, r6, r0 # r5:r6 = V if U < 0, = 0 if U >= 0 + subfc r8, r4, r8 + subfe r7, r3, r7 + subfc r4, r6, r8 + subfe r3, r5, r7 # result is r8:r7 - r3:r4 - r5:r6 + blr + .type __i64_umulh, @function + .size __i64_umulh, .-__i64_umulh diff --git a/runtime/powerpc/i64_umul.s b/runtime/powerpc/i64_umul.s new file mode 100644 index 00000000..e734b93c --- /dev/null +++ b/runtime/powerpc/i64_umul.s @@ -0,0 +1,64 @@ +# ***************************************************************** +# +# The Compcert verified compiler +# +# Xavier Leroy, INRIA Paris +# +# Copyright (c) 2016 Institut National de Recherche en Informatique et +# en Automatique. +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are met: +# * Redistributions of source code must retain the above copyright +# notice, this list of conditions and the following disclaimer. +# * Redistributions in binary form must reproduce the above copyright +# notice, this list of conditions and the following disclaimer in the +# documentation and/or other materials provided with the distribution. +# * Neither the name of the nor the +# names of its contributors may be used to endorse or promote products +# derived from this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, +# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, +# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR +# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF +# LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING +# NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS +# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. +# +# ********************************************************************* + +# Helper functions for 64-bit integer arithmetic. PowerPC version. + + .text + +# Unsigned multiply high + +# Reference C implementation in ../c/i64_umul.c + + .balign 16 + .globl __i64_umulh +__i64_umulh: + # u1 in r3; u0 in r4; v1 in r5; v0 in r6 + mulhwu r0, r4, r6 # k (in r0) = high((u64) u0 * (u64) v0) + mullw r8, r3, r6 + mulhwu r7, r3, r6 # t (in r8:r7) = (u64) u1 * (u64) v0 + addc r0, r8, r0 # w1 (in r0) = low (t + k) + addze r9, r7 # w2 (in r9) = high (t + k) + mullw r8, r4, r5 + mulhwu r7, r4, r5 # t (in r8:r7) = (u64) u0 * (u64) v1 + addc r0, r8, r0 # tmp (in r0) = low (t + w1) + addze r0, r7 # k (in r0) = high(t + w1) + mullw r8, r3, r5 + mulhwu r7, r3, r5 # t (in r8:r7) = (u64) u1 * (u64) v1 + addc r4, r8, r9 # add w2 + addze r3, r7 + addc r4, r4, r0 # add k + addze r3, r3 + blr + .type __i64_umulh, @function + .size __i64_umulh, .-__i64_umulh -- cgit