derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
72 lines
2.6 KiB
ArmAsm
72 lines
2.6 KiB
ArmAsm
//===-- unorddf2.S - double-precision floating point comparison -----------===//
|
|
//
|
|
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
|
|
// See https://llvm.org/LICENSE.txt for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
//
|
|
// Return 1 if the result of comparing x with y is 'unordered', i.e.
|
|
// one of x and y is NaN.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#include "../assembly.h"
|
|
#include "crt_endian.h"
|
|
|
|
|
|
.syntax unified
|
|
.text
|
|
.p2align 2
|
|
|
|
|
|
#if __ARM_PCS_VFP
|
|
DEFINE_COMPILERRT_FUNCTION(__unorddf2)
|
|
push {r4, lr}
|
|
VMOV_FROM_DOUBLE(r0, r1, d0)
|
|
VMOV_FROM_DOUBLE(r2, r3, d1)
|
|
bl __aeabi_dcmpun
|
|
pop {r4, pc}
|
|
#else
|
|
DEFINE_COMPILERRT_FUNCTION_ALIAS(__unorddf2, __aeabi_dcmpun)
|
|
#endif
|
|
|
|
DEFINE_COMPILERRT_FUNCTION(__aeabi_dcmpun)
|
|
|
|
// This function isn't based on the general-purpose code in dcmp.h, because
|
|
// it's more effort than needed. Here we just need to identify whether or not
|
|
// there's at least one NaN in the inputs. There's no need to vary that check
|
|
// based on the sign bit, so we might as well just do the NaN test as quickly
|
|
// as possible.
|
|
//
|
|
// We do this by shifting the high word of each input left to get rid of the
|
|
// sign bit, shifting a bit in at the bottom which is 1 if any bit is set in
|
|
// the low word. Then we check if the result is _greater_ than 0xFFE00000
|
|
// (but not equal), via adding 0x00200000 to it and testing for the HI
|
|
// condition (carry flag set, but Z clear).
|
|
//
|
|
// Once we've done that transformation to the first input xh:xl, we
|
|
// free up xl to contain our constant 0x00200000, so there's no need
|
|
// to push any registers.
|
|
cmp xl, #1 // set C if xl is nonzero
|
|
adc xh, xh, xh // shift xh left, bringing in the C bit
|
|
cmp yl, #1 // set C if yl is nonzero
|
|
adc yh, yh, yh // shift yh left, bringing in the C bit
|
|
cmn xh, #1 << 21 // if HI, then x is a NaN
|
|
cmnls yh, #1 << 21 // if not HI, then do the same check for y
|
|
|
|
// If LS, then we have no NaNs and return false. We do this as quickly as we
|
|
// can (not stopping to take two instructions setting up r0 for both
|
|
// possibilities), on the assumption that NaNs are rare and we want to
|
|
// optimize for the non-NaN path.
|
|
movls r0, #0
|
|
bxls lr
|
|
|
|
// Otherwise, we have at least one NaN, and return true.
|
|
mov r0, #1
|
|
bx lr
|
|
|
|
END_COMPILERRT_FUNCTION(__aeabi_dcmpun)
|
|
|
|
NO_EXEC_STACK_DIRECTIVE
|