arkcompiler_runtime_core/tests/benchmarks/math-cordic.pa

# Copyright (c) 2021-2022 Huawei Device Co., Ltd.
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

.function u1 main(){
    movi v0, 2500000
    fmovi.64 v1, 1036257.0468539958
    call.short test, v0, v1
    return
}

.function u1 test(i32 a0, f64 a1){
    movi v3, 12
    newarr v0, v3, f64[]                            #Angles
    call.short init, v0, v0
    call.short cordic, a0, v0
    sta.64 v2                  #total
    fsub2.64 a1
    sta.64 v1
    fldai.64 1e-13
    fcmpl.64 v1
    jltz exit_failure
    ldai 0
    return
exit_failure:
    ldai 1
    return
}

.function f64 cordic(i32 a0, f64[] a1){
    fmovi.64 v1, 0.0         #total
    movi v2, 0           #loop_counter
    fmovi.64 v3, 28.027      #TARGET_ANGLE
loop:
    lda v2
    jeq a0, loop_exit
    call.short cordicsincos, v3, a1
    fadd2.64 v1
    sta.64 v1
    inci v2, 1
    jmp loop
loop_exit:
    lda.64 v1
    return.64
}

.function f64 cordicsincos(f64 a0, f64[] a1){
    fmovi.64 v1, 0.0           #x
    fmovi.64 v2, 0.0           #y
    fmovi.64 v3, 0.0           #targetAngle
    fmovi.64 v4, 0.0           #currAngle
    movi.64 v5, 0              #step
    fmovi.64 v6, 0.6072529350      #AG_CONST
    movi v12, 12
    call.short FIXED, v6, v1   #x
    sta.64 v1
    call.short FIXED, a0, a0
    sta.64 v3
    movi v8, 0                 #step
loop:
    lda v8
    jeq v12, loop_exit
    lda.64 v3
    fcmpl.64 v4
    jgtz if
    lda.64 v2
    f64toi32
    shr2 v8
    i32tof64
    fadd2.64 v1
    sta.64 v10                 #newX
    lda.64 v1
    f64toi32
    shr2 v8
    i32tof64
    fneg.64
    fadd2.64 v2
    sta.64 v2
    mov.64 v1, v10
    lda.64 v8
    fldarr.64 a1
    fneg.64
    fadd2.64 v4
    sta.64 v4
    inci v8, 1
    jmp loop
loop_exit:
    call.short FLOAT, v1, v1
    sta.64 v12
    call.short FLOAT, v2, v1
    sta.64 v13
    fmul2.64 v12
    return.64
if:
    lda.64 v2
    f64toi32
    shr2 v8
    i32tof64
    fneg.64
    fadd2.64 v1
    sta.64 v10              #newX
    lda.64 v1
    f64toi32
    shr2 v8
    i32tof64
    fadd2.64 v2
    sta.64 v2
    mov.64 v1, v10
    lda v8
    fldarr.64 a1
    fadd2.64 v4
    sta.64 v4
    inci v8, 1
    lda v8
    jmp loop
}

.function f64 FLOAT(f64 a0){
    fmovi.64 v1, 65536.0
    lda.64 a0
    fdiv2.64 v1
    return.64
}

.function f64 FIXED(f64 a0){
    fldai.64 65536.0
    fmul2.64 a0
    return.64
}

.function void init(f64[] a0){
    movi.64 v2, 0
    fmovi.64 v1, 45.0
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    inci v2, 1
    fmovi.64 v1, 26.565
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    inci v2, 1
    fmovi.64 v1, 14.0362
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    inci v2, 1
    fmovi.64 v1, 7.12502
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    inci v2, 1
    fmovi.64 v1, 3.57633
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    inci v2, 1
    fmovi.64 v1, 1.78991
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    inci v2, 1
    fmovi.64 v1, 0.895174
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    inci v2, 1
    fmovi.64 v1, 0.447614
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    inci v2, 1
    fmovi.64 v1, 0.223811
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    inci v2, 1
    fmovi.64 v1, 0.111906
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    inci v2, 1
    fmovi.64 v1, 0.055953
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    inci v2, 1
    fmovi.64 v1, 0.027977
    call.short FIXED, v1, v1
    fstarr.64 a0, v2
    return.void
}