| @@ -13,6 +13,7 @@ CAMAXKERNEL = camax_lsx.S | |||
| SAMINKERNEL = amin_lsx.S | |||
| DAMINKERNEL = amin_lsx.S | |||
| CAMINKERNEL = camin_lsx.S | |||
| SMAXKERNEL = max_lsx.S | |||
| DMAXKERNEL = max_lsx.S | |||
| @@ -13,6 +13,7 @@ CAMAXKERNEL = camax_lasx.S | |||
| SAMINKERNEL = amin_lasx.S | |||
| DAMINKERNEL = amin_lasx.S | |||
| CAMINKERNEL = camin_lasx.S | |||
| SMAXKERNEL = max_lsx.S | |||
| DMAXKERNEL = max_lsx.S | |||
| @@ -0,0 +1,199 @@ | |||
| /*************************************************************************** | |||
| Copyright (c) 2023, The OpenBLAS Project | |||
| All rights reserved. | |||
| Redistribution and use in source and binary forms, with or without | |||
| modification, are permitted provided that the following conditions are | |||
| met: | |||
| 1. Redistributions of source code must retain the above copyright | |||
| notice, this list of conditions and the following disclaimer. | |||
| 2. 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. | |||
| 3. Neither the name of the OpenBLAS project 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 THE OPENBLAS PROJECT OR CONTRIBUTORS 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. | |||
| *****************************************************************************/ | |||
| #define ASSEMBLER | |||
| #include "common.h" | |||
| #define N $r4 | |||
| #define X $r5 | |||
| #define INCX $r6 | |||
| #define I $r12 | |||
| #define TEMP $r16 | |||
| #define t1 $f14 | |||
| #define t2 $f18 | |||
| #define t3 $f15 | |||
| #define t4 $f17 | |||
| #define s1 $f22 | |||
| #define s2 $f9 | |||
| #define s3 $f10 | |||
| #define s4 $f11 | |||
| #define a0 $f20 | |||
| #define a1 $f21 | |||
| #define x1 $xr9 | |||
| #define x2 $xr10 | |||
| #define x3 $xr11 | |||
| #define x4 $xr12 | |||
| #define VT0 $xr13 | |||
| #define VT1 $xr14 | |||
| #define res0 $xr18 | |||
| #define neg1 $xr19 | |||
| #define VX0 $xr20 | |||
| #define VX1 $xr21 | |||
| #define VM0 $xr22 | |||
| #define VM1 $xr23 | |||
| PROLOGUE | |||
| MTC s1, $r0 | |||
| xvxor.v res0, res0, res0 | |||
| bge $r0, N, .L999 | |||
| bge $r0, INCX, .L999 | |||
| fld.s a0, X, 0 * SIZE | |||
| fld.s a1, X, 1 * SIZE | |||
| fabs.s a0, a0 | |||
| fabs.s a1, a1 | |||
| fadd.s s1, a1, a0 | |||
| xvreplve0.w VM0, VM0 | |||
| li.d TEMP, 1 | |||
| li.w I, -1 | |||
| slli.d TEMP, TEMP, ZBASE_SHIFT | |||
| slli.d INCX, INCX, ZBASE_SHIFT | |||
| xvreplgr2vr.w neg1, I | |||
| xvffint.s.w neg1, neg1 | |||
| srai.d I, N, 3 | |||
| bne INCX, TEMP, .L20 | |||
| bge $r0, I, .L23 | |||
| .align 3 | |||
| .L10: | |||
| xvld VX0, X, 0 * SIZE | |||
| xvld VX1, X, 8 * SIZE | |||
| addi.d I, I, -1 | |||
| xvpickev.w x1, VX1, VX0 | |||
| xvpickod.w x2, VX1, VX0 | |||
| xvfmul.s x3, neg1, x1 | |||
| xvfmul.s x4, neg1, x2 | |||
| xvfcmp.clt.s VT0, x1, res0 | |||
| xvfcmp.clt.s VT1, x2, res0 | |||
| xvbitsel.v x1, x1, x3, VT0 | |||
| xvbitsel.v x2, x2, x4, VT1 | |||
| addi.d X, X, 16 * SIZE | |||
| xvfadd.s VM1, x1, x2 | |||
| xvfmin.s VM0, VM0, VM1 | |||
| blt $r0, I, .L10 | |||
| .align 3 | |||
| .L11: | |||
| xvpickve.w x1, VM0, 0 | |||
| xvpickve.w x2, VM0, 1 | |||
| xvpickve.w x3, VM0, 2 | |||
| xvpickve.w x4, VM0, 3 | |||
| xvfmin.s VM1, x1, x2 | |||
| xvfmin.s VM0, x3, x4 | |||
| xvfmin.s VM0, VM0, VM1 | |||
| b .L23 | |||
| .align 3 | |||
| .L20: // INCX!=1 | |||
| bge $r0, I, .L23 | |||
| .align 3 | |||
| .L21: | |||
| fld.s t1, X, 0 * SIZE | |||
| fld.s t2, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fld.s t3, X, 0 * SIZE | |||
| fld.s t4, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fabs.s t1, t1 | |||
| fabs.s t2, t2 | |||
| fabs.s t3, t3 | |||
| fabs.s t4, t4 | |||
| fadd.s t1, t1, t2 | |||
| fadd.s t3, t3, t4 | |||
| fmin.s s1, t1, t3 | |||
| fld.s t1, X, 0 * SIZE | |||
| fld.s t2, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fld.s t3, X, 0 * SIZE | |||
| fld.s t4, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fabs.s t1, t1 | |||
| fabs.s t2, t2 | |||
| fabs.s t3, t3 | |||
| fabs.s t4, t4 | |||
| fadd.s t1, t1, t2 | |||
| fadd.s t3, t3, t4 | |||
| fmin.s s1, t1, t3 | |||
| fld.s t1, X, 0 * SIZE | |||
| fld.s t2, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fld.s t3, X, 0 * SIZE | |||
| fld.s t4, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fabs.s t1, t1 | |||
| fabs.s t2, t2 | |||
| fabs.s t3, t3 | |||
| fabs.s t4, t4 | |||
| addi.d I, I, -1 | |||
| fadd.s t1, t1, t2 | |||
| fadd.s t3, t3, t4 | |||
| fmin.s s3, t1, t3 | |||
| fld.s t1, X, 0 * SIZE | |||
| fld.s t2, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fld.s t3, X, 0 * SIZE | |||
| fld.s t4, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fabs.s t1, t1 | |||
| fabs.s t2, t2 | |||
| fabs.s t3, t3 | |||
| fabs.s t4, t4 | |||
| fadd.s t1, t1, t2 | |||
| fadd.s t3, t3, t4 | |||
| fmin.s s4, t1, t3 | |||
| blt $r0, I, .L21 | |||
| .align 3 | |||
| .L22: | |||
| fmin.s s1, s1, s2 | |||
| fmin.s s3, s3, s4 | |||
| fmin.s s1, s1, s3 | |||
| .align 3 | |||
| .L23: //N<8 | |||
| andi I, N, 7 | |||
| bge $r0, I, .L999 | |||
| .align 3 | |||
| .L24: | |||
| LD a0, X, 0 * SIZE | |||
| LD a1, X, 1 * SIZE | |||
| addi.d I, I, -1 | |||
| FABS a0, a0 | |||
| FABS a1, a1 | |||
| ADD a0, a0, a1 | |||
| add.d X, X, INCX | |||
| fmin.s s1, a0, s1 | |||
| blt $r0, I, .L24 | |||
| .align 3 | |||
| .L999: | |||
| fmov.s $f0, $f22 | |||
| jirl $r0, $r1, 0x0 | |||
| .align 3 | |||
| EPILOGUE | |||
| @@ -0,0 +1,211 @@ | |||
| /*************************************************************************** | |||
| Copyright (c) 2023, The OpenBLAS Project | |||
| All rights reserved. | |||
| Redistribution and use in source and binary forms, with or without | |||
| modification, are permitted provided that the following conditions are | |||
| met: | |||
| 1. Redistributions of source code must retain the above copyright | |||
| notice, this list of conditions and the following disclaimer. | |||
| 2. 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. | |||
| 3. Neither the name of the OpenBLAS project 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 THE OPENBLAS PROJECT OR CONTRIBUTORS 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. | |||
| *****************************************************************************/ | |||
| #define ASSEMBLER | |||
| #include "common.h" | |||
| #define N $r4 | |||
| #define X $r5 | |||
| #define INCX $r6 | |||
| #define I $r12 | |||
| #define t1 $f14 | |||
| #define t2 $f18 | |||
| #define t3 $f15 | |||
| #define t4 $f17 | |||
| #define s1 $f22 | |||
| #define s2 $f9 | |||
| #define s3 $f10 | |||
| #define s4 $f11 | |||
| #define TEMP $r16 | |||
| #define a0 $f20 | |||
| #define a1 $f21 | |||
| #define x1 $vr9 | |||
| #define x2 $vr10 | |||
| #define x3 $vr11 | |||
| #define x4 $vr12 | |||
| #define VT0 $vr13 | |||
| #define VT1 $vr14 | |||
| #define res0 $vr18 | |||
| #define neg1 $vr19 | |||
| #define VX0 $vr20 | |||
| #define VX1 $vr21 | |||
| #define VM0 $vr22 | |||
| #define VM1 $vr23 | |||
| PROLOGUE | |||
| MTC s1, $r0 | |||
| vxor.v res0, res0, res0 | |||
| bge $r0, N, .L999 | |||
| bge $r0, INCX, .L999 | |||
| fld.s a0, X, 0 * SIZE | |||
| fld.s a1, X, 1 * SIZE | |||
| fabs.s a0, a0 | |||
| fabs.s a1, a1 | |||
| fadd.s s1, a1, a0 | |||
| vreplvei.w VM0, VM0, 0 | |||
| li.d TEMP, 1 | |||
| li.w I, -1 | |||
| slli.d TEMP, TEMP, ZBASE_SHIFT | |||
| slli.d INCX, INCX, ZBASE_SHIFT | |||
| vreplgr2vr.w neg1, I | |||
| vffint.s.w neg1, neg1 | |||
| srai.d I, N, 3 | |||
| bne INCX, TEMP, .L20 | |||
| bge $r0, I, .L23 | |||
| .align 3 | |||
| .L10: | |||
| vld VX0, X, 0 * SIZE | |||
| vld VX1, X, 4 * SIZE | |||
| addi.d I, I, -1 | |||
| vpickev.w x1, VX1, VX0 | |||
| vpickod.w x2, VX1, VX0 | |||
| vfmul.s x3, neg1, x1 | |||
| vfmul.s x4, neg1, x2 | |||
| vfcmp.clt.s VT0, x1, res0 | |||
| vfcmp.clt.s VT1, x2, res0 | |||
| vld VX0, X, 8 * SIZE | |||
| vbitsel.v x1, x1, x3, VT0 | |||
| vbitsel.v x2, x2, x4, VT1 | |||
| vld VX1, X, 12 * SIZE | |||
| vfadd.s VM1, x1, x2 | |||
| vpickev.w x1, VX1, VX0 | |||
| vpickod.w x2, VX1, VX0 | |||
| vfmul.s x3, neg1, x1 | |||
| vfmul.s x4, neg1, x2 | |||
| vfcmp.clt.s VT0, x1, res0 | |||
| vfcmp.clt.s VT1, x2, res0 | |||
| addi.d X, X, 16 * SIZE | |||
| vbitsel.v x1, x1, x3, VT0 | |||
| vbitsel.v x2, x2, x4, VT1 | |||
| vfadd.s x1, x1, x2 | |||
| vfmin.s VM1, x1, VM1 | |||
| vfmin.s VM0, VM0, VM1 | |||
| blt $r0, I, .L10 | |||
| .align 3 | |||
| .L11: | |||
| vreplvei.w x1, VM0, 0 | |||
| vreplvei.w x2, VM0, 1 | |||
| vreplvei.w x3, VM0, 2 | |||
| vreplvei.w x4, VM0, 3 | |||
| vfmin.s VM1, x1, x2 | |||
| vfmin.s VM0, x3, x4 | |||
| vfmin.s VM0, VM0, VM1 | |||
| b .L23 | |||
| .align 3 | |||
| .L20: // INCX!=1 | |||
| bge $r0, I, .L23 | |||
| .align 3 | |||
| .L21: | |||
| fld.s t1, X, 0 * SIZE | |||
| fld.s t2, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fld.s t3, X, 0 * SIZE | |||
| fld.s t4, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fabs.s t1, t1 | |||
| fabs.s t2, t2 | |||
| fabs.s t3, t3 | |||
| fabs.s t4, t4 | |||
| fadd.s t1, t1, t2 | |||
| fadd.s t3, t3, t4 | |||
| fmin.s s1, t1, t3 | |||
| fld.s t1, X, 0 * SIZE | |||
| fld.s t2, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fld.s t3, X, 0 * SIZE | |||
| fld.s t4, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fabs.s t1, t1 | |||
| fabs.s t2, t2 | |||
| fabs.s t3, t3 | |||
| fabs.s t4, t4 | |||
| fadd.s t1, t1, t2 | |||
| fadd.s t3, t3, t4 | |||
| fmin.s s1, t1, t3 | |||
| fld.s t1, X, 0 * SIZE | |||
| fld.s t2, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fld.s t3, X, 0 * SIZE | |||
| fld.s t4, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fabs.s t1, t1 | |||
| fabs.s t2, t2 | |||
| fabs.s t3, t3 | |||
| fabs.s t4, t4 | |||
| addi.d I, I, -1 | |||
| fadd.s t1, t1, t2 | |||
| fadd.s t3, t3, t4 | |||
| fmin.s s3, t1, t3 | |||
| fld.s t1, X, 0 * SIZE | |||
| fld.s t2, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fld.s t3, X, 0 * SIZE | |||
| fld.s t4, X, 1 * SIZE | |||
| add.d X, X, INCX | |||
| fabs.s t1, t1 | |||
| fabs.s t2, t2 | |||
| fabs.s t3, t3 | |||
| fabs.s t4, t4 | |||
| fadd.s t1, t1, t2 | |||
| fadd.s t3, t3, t4 | |||
| fmin.s s4, t1, t3 | |||
| blt $r0, I, .L21 | |||
| .align 3 | |||
| .L22: | |||
| fmin.s s1, s1, s2 | |||
| fmin.s s3, s3, s4 | |||
| fmin.s s1, s1, s3 | |||
| .align 3 | |||
| .L23: //N<8 | |||
| andi I, N, 7 | |||
| bge $r0, I, .L999 | |||
| .align 3 | |||
| .L24: | |||
| fld.s a0, X, 0 * SIZE | |||
| fld.s a1, X, 1 * SIZE | |||
| addi.d I, I, -1 | |||
| fabs.s a0, a0 | |||
| fabs.s a1, a1 | |||
| fadd.s a0, a0, a1 | |||
| add.d X, X, INCX | |||
| fmin.s s1, a0, s1 | |||
| blt $r0, I, .L24 | |||
| .align 3 | |||
| .L999: | |||
| fmov.s $f0, $f22 | |||
| jirl $r0, $r1, 0x0 | |||
| .align 3 | |||
| EPILOGUE | |||