Compilation and Installation
Procedure
- Use PuTTY to log in to the server as the root user.
- Download the source code to the /path/to directory:
cd /path/to git clone -b v5.0.1428 https://github.com/rcedgar/muscle.git
- Go to the src directory.
cd /path/to/muscle/src
- Modify the Makefile file.
- Modify the pairhmm.h file.
- Modify the pairhmm.cpp file.
- Modify the perturbhmm.cpp file.
- Modify the multisequence.cpp file.
- Open the multisequence.cpp file.
vi multisequence.cpp
- Press i to enter the edit mode and add #include <errno.h> to the header file list.
Before the modification:
After the modification:
- Press Esc, type :wq!, and press Enter to save the settings and exit.
- Open the multisequence.cpp file.
- Modify the myutils.cpp file.
- Open the myutils.cpp file.
vi myutils.cpp
- Press i to enter the edit mode and delete the entire GetVersionString code segment from line 2055. This code segment has defects and the deletion does not affect the program usage.
Before the modification:
After the modification:
- Press Esc, type :wq!, and press Enter to save the settings and exit.
- Open the myutils.cpp file.
- Modify the timing.h file.
- Open the timing.h file.
vi timing.h
- Press i to enter the edit mode.
- Add a macro definition at the beginning of the file.
#if defined(__aarch64__) #include "KunpengTrans.h" #endif
Before the modification:
After the modification:
- Modify the __asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi)) code in line 25 as follows:
#if defined(__x86_64__) __asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi)); #elif defined(__aarch64__) // Description: Replace with the converted code block suggested. Note: // Suggestion: // { unsigned int _kp_tmp_CPUFreq = KUNPENG_CPU_FREQUENCY_MHZ; unsigned long long _kp_tmp_TSCCount; unsigned long long _kp_tmp_TSCFreq; __asm__ __volatile__ ("mrs %0, cntfrq_el0" : "=r" (_kp_tmp_TSCFreq)); __asm__ __volatile__ ("mrs %0, cntvct_el0" : "=r" (_kp_tmp_TSCCount)); lo = (unsigned int)(_kp_tmp_TSCCount * _kp_tmp_CPUFreq / _kp_tmp_TSCFreq); hi = (unsigned int)((_kp_tmp_TSCCount * _kp_tmp_CPUFreq / _kp_tmp_TSCFreq) >> 32); // } #endifBefore the modification:
After the modification:
- Add a macro definition at the beginning of the file.
- Press Esc, type :wq!, and press Enter to save the settings and exit.
- Open the timing.h file.
- Create a KunpengTrans.h file.
- Create a KunpengTrans.h file.
vi KunpengTrans.h
- Press i to enter the edit mode and add the following content:
/* * @Description: KunpengTrans.h * @Copyright (c) Huawei Technologies Co., Ltd. 2020-2020. All rights reserved. */ #ifndef KUNPENG_TRANS_H #define KUNPENG_TRANS_H #include <string.h> /* ATTENTION: * Please set KUNPENG_CPU_FREQUENCY_MHZ to the actual cpu frequecy of your running environment. */ const int KUNPENG_CPU_FREQUENCY_MHZ = 2600; const int EAX_LEAF = 7; const int SHIFT_THREE = 3; /* CASE eax = 0; Highest Function Parameter and Manufacturer ID */ void GetCPUManuID(unsigned int *ebx, unsigned int *edx, unsigned int *ecx); // %ecx #define bit_SSE3 (1 << 0) #define bit_PCLMUL (1 << 1) #define bit_LZCNT (1 << 5) #define bit_SSSE3 (1 << 9) #define bit_FMA (1 << 12) #define bit_CMPXCHG16B (1 << 13) #define bit_SSE4_1 (1 << 19) #define bit_SSE4_2 (1 << 20) #define bit_MOVBE (1 << 22) #define bit_POPCNT (1 << 23) #define bit_AES (1 << 25) #define bit_XSAVE (1 << 26) #define bit_OSXSAVE (1 << 27) #define bit_AVX (1 << 28) #define bit_F16C (1 << 29) #define bit_RDRND (1 << 30) // %edx #define bit_CMPXCHG8B (1 << 8) #define bit_CMOV (1 << 15) #define bit_MMX (1 << 23) #define bit_FXSAVE (1 << 24) #define bit_SSE (1 << 25) #define bit_SSE2 (1 << 26) /* CASE eax = 1; Processor Info and Feature Bits * Skylake: eax = 0x00050654, ebx = 0x43400800, ecx = 0x7ffefbf7, edx = 0xbfebfbff // Sampled at 6148 * Cascade Lake : eax = 0x00050657, ebx = 0x08400800, ecx = 0x7ffefbf7, edx = 0xbfebfbff // Sampled at 6248 */ void GetCPUFeature(unsigned int *eax, unsigned int *ebx, unsigned int *ecx, unsigned int *edx); /* %ebx */ #define bit_FSGSBASE (1 << 0) #define bit_SGX (1 << 2) #define bit_BMI (1 << 3) #define bit_HLE (1 << 4) #define bit_AVX2 (1 << 5) #define bit_BMI2 (1 << 8) #define bit_RTM (1 << 11) #define bit_MPX (1 << 14) #define bit_AVX512F (1 << 16) #define bit_AVX512DQ (1 << 17) #define bit_RDSEED (1 << 18) #define bit_ADX (1 << 19) #define bit_AVX512IFMA (1 << 21) #define bit_CLFLUSHOPT (1 << 23) #define bit_CLWB (1 << 24) #define bit_AVX512PF (1 << 26) #define bit_AVX512ER (1 << 27) #define bit_AVX512CD (1 << 28) #define bit_SHA (1 << 29) #define bit_AVX512BW (1 << 30) #define bit_AVX512VL (1u << 31) /* %ecx */ #define bit_PREFETCHWT1 (1 << 0) #define bit_AVX512VBMI (1 << 1) #define bit_PKU (1 << 3) #define bit_OSPKE (1 << 4) #define bit_AVX512VBMI2 (1 << 6) #define bit_SHSTK (1 << 7) #define bit_GFNI (1 << 8) #define bit_VAES (1 << 9) #define bit_AVX512VNNI (1 << 11) #define bit_VPCLMULQDQ (1 << 10) #define bit_AVX512BITALG (1 << 12) #define bit_AVX512VPOPCNTDQ (1 << 14) #define bit_RDPID (1 << 22) #define bit_MOVDIRI (1 << 27) #define bit_MOVDIR64B (1 << 28) /* %edx */ #define bit_AVX5124VNNIW (1 << 2) #define bit_AVX5124FMAPS (1 << 3) #define bit_IBT (1 << 20) #define bit_PCONFIG (1 << 18) /* CASE eax = 7, ecx = 0; Extended Features * Skylake: eax = 0x00000000, ebx = 0xd39ffffb, ecx = 0x00000018, edx = 0x9c002400 // Sampled at 6148 * Cascade Lake : eax = 0x00000000, ebx = 0xd39ffffb, ecx = 0x00000818, edx = 0xbc000400 // Sampled at 6248 */ void GetExtendCPUFeature(unsigned int *eax, unsigned int *ebx, unsigned int *ecx, unsigned int *edx); /* * GetSupportedCPUID only supported while eax = 0, eax = 1 , eax = 7 and ecx = 0 * In other cases always return 0x0; * chipID support 1,2 -- 1 for skylake server eg. 6148; 2 for Cascade Lake 6248 */ void GetSupportedCPUID(unsigned int *eax, unsigned int *ebx, unsigned int *ecx, unsigned int *edx); void RepStos(void *dest, unsigned long long src, unsigned long long len, unsigned width, int df); void RepStosB(void *dest, unsigned long long src, unsigned long long len, int DF); void RepStosW(void *dest, unsigned long long src, unsigned long long len, int DF); void RepStosD(void *dest, unsigned long long src, unsigned long long len, int DF); void RepStosQ(void *dest, unsigned long long src, unsigned long long len, int DF); void RepMovs(void *dest, void *src, unsigned long long len, unsigned width, int df); #include <arm_neon.h> #define KP_FORCE_INLINE static inline __attribute__((always_inline)) typedef union { int8x16_t vect_s8; int16x8_t vect_s16; int32x4_t vect_s32; int64x2_t vect_s64; uint8x16_t vect_u8; uint16x8_t vect_u16; uint32x4_t vect_u32; uint64x2_t vect_u64; } __kp_m128i; #define _KP_SIDD_NEGATIVE_POLARITY 0x10 // negate results #define _KP_SIDD_MASKED_NEGATIVE_POLARITY 0x30 // negate results only before end of string static uint16_t g_kp_mask_epi16[8] __attribute__((aligned(16))) = { 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80 }; static uint8_t g_kp_mask_epi8[16] __attribute__((aligned(16))) = { 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80 }; #define KP_PCMPSTR_EQ_16x8(a, b, mtx) \ { \ mtx[0].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 0)), a.vect_u16); \ mtx[1].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 1)), a.vect_u16); \ mtx[2].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 2)), a.vect_u16); \ mtx[3].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 3)), a.vect_u16); \ mtx[4].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 4)), a.vect_u16); \ mtx[5].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 5)), a.vect_u16); \ mtx[6].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 6)), a.vect_u16); \ mtx[7].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 7)), a.vect_u16); \ } #define KP_PCMPSTR_EQ_8x16(a, b, mtx) \ { \ mtx[0].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 0)), a.vect_u8); \ mtx[1].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 1)), a.vect_u8); \ mtx[2].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 2)), a.vect_u8); \ mtx[3].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 3)), a.vect_u8); \ mtx[4].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 4)), a.vect_u8); \ mtx[5].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 5)), a.vect_u8); \ mtx[6].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 6)), a.vect_u8); \ mtx[7].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 7)), a.vect_u8); \ mtx[8].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 8)), a.vect_u8); \ mtx[9].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 9)), a.vect_u8); \ mtx[10].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 10)), a.vect_u8); \ mtx[11].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 11)), a.vect_u8); \ mtx[12].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 12)), a.vect_u8); \ mtx[13].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 13)), a.vect_u8); \ mtx[14].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 14)), a.vect_u8); \ mtx[15].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 15)), a.vect_u8); \ } #define KP_PCMPSTR_RNG_U16x8(a, b, mtx) \ { \ uint16x8_t vect_b[8]; \ __kp_m128i mask; \ mask.vect_u32 = vdupq_n_u32(0xffff); \ vect_b[0] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 0)); \ vect_b[1] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 1)); \ vect_b[2] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 2)); \ vect_b[3] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 3)); \ vect_b[4] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 4)); \ vect_b[5] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 5)); \ vect_b[6] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 6)); \ vect_b[7] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 7)); \ int i; \ for (i = 0; i < 8; i++) { \ mtx[i].vect_u16 = vbslq_u16(mask.vect_u16, vcgeq_u16(vect_b[i], a.vect_u16), \ vcleq_u16(vect_b[i], a.vect_u16)); \ } \ } #define KP_PCMPSTR_RNG_S16x8(a, b, mtx) \ { \ int16x8_t vect_b[8]; \ __kp_m128i mask; \ mask.vect_u32 = vdupq_n_u32(0xffff); \ vect_b[0] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 0)); \ vect_b[1] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 1)); \ vect_b[2] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 2)); \ vect_b[3] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 3)); \ vect_b[4] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 4)); \ vect_b[5] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 5)); \ vect_b[6] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 6)); \ vect_b[7] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 7)); \ int i; \ for (i = 0; i < 8; i++) { \ mtx[i].vect_u16 = vbslq_u16(mask.vect_u16, vcgeq_s16(vect_b[i], a.vect_s16), \ vcleq_s16(vect_b[i], a.vect_s16)); \ } \ } #define KP_PCMPSTR_RNG_U8x16(a, b, mtx) \ { \ uint8x16_t vect_b[16]; \ __kp_m128i mask; \ mask.vect_u16 = vdupq_n_u16(0xff); \ vect_b[0] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 0)); \ vect_b[1] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 1)); \ vect_b[2] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 2)); \ vect_b[3] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 3)); \ vect_b[4] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 4)); \ vect_b[5] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 5)); \ vect_b[6] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 6)); \ vect_b[7] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 7)); \ vect_b[8] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 8)); \ vect_b[9] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 9)); \ vect_b[10] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 10)); \ vect_b[11] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 11)); \ vect_b[12] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 12)); \ vect_b[13] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 13)); \ vect_b[14] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 14)); \ vect_b[15] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 15)); \ int i; \ for (i = 0; i < 16; i++) { \ mtx[i].vect_u8 = vbslq_u8(mask.vect_u8, vcgeq_u8(vect_b[i], a.vect_u8), vcleq_u8(vect_b[i], a.vect_u8)); \ } \ } #define KP_PCMPSTR_RNG_S8x16(a, b, mtx) \ { \ int8x16_t vect_b[16]; \ __kp_m128i mask; \ mask.vect_u16 = vdupq_n_u16(0xff); \ vect_b[0] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 0)); \ vect_b[1] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 1)); \ vect_b[2] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 2)); \ vect_b[3] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 3)); \ vect_b[4] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 4)); \ vect_b[5] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 5)); \ vect_b[6] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 6)); \ vect_b[7] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 7)); \ vect_b[8] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 8)); \ vect_b[9] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 9)); \ vect_b[10] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 10)); \ vect_b[11] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 11)); \ vect_b[12] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 12)); \ vect_b[13] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 13)); \ vect_b[14] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 14)); \ vect_b[15] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 15)); \ int i; \ for (i = 0; i < 16; i++) { \ mtx[i].vect_u8 = vbslq_u8(mask.vect_u8, vcgeq_s8(vect_b[i], a.vect_s8), vcleq_s8(vect_b[i], a.vect_s8)); \ } \ } static int kp_aggregate_equal_any_8x16(int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j; int m = (1 << la) - 1; uint8x8_t vect_mask = vld1_u8(g_kp_mask_epi8); uint8x8_t t_lo = vtst_u8(vdup_n_u8(m & 0xff), vect_mask); uint8x8_t t_hi = vtst_u8(vdup_n_u8(m >> 8), vect_mask); uint8x16_t vect = vcombine_u8(t_lo, t_hi); for (j = 0; j < lb; j++) { mtx[j].vect_u8 = vandq_u8(vect, mtx[j].vect_u8); mtx[j].vect_u8 = vshrq_n_u8(mtx[j].vect_u8, 7); int tmp = vaddvq_u8(mtx[j].vect_u8) ? 1 : 0; res |= (tmp << j); } return res; } static int kp_aggregate_equal_any_16x8(int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j; int m = (1 << la) - 1; uint16x8_t vect = vtstq_u16(vdupq_n_u16(m), vld1q_u16(g_kp_mask_epi16)); for (j = 0; j < lb; j++) { mtx[j].vect_u16 = vandq_u16(vect, mtx[j].vect_u16); mtx[j].vect_u16 = vshrq_n_u16(mtx[j].vect_u16, 15); int tmp = vaddvq_u16(mtx[j].vect_u16) ? 1 : 0; res |= (tmp << j); } return res; } static int kp_cal_res_byte_equal_any(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_EQ_8x16(a, b, mtx); return kp_aggregate_equal_any_8x16(la, lb, mtx); } static int kp_cal_res_word_equal_any(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_EQ_16x8(a, b, mtx); return kp_aggregate_equal_any_16x8(la, lb, mtx); } static int kp_aggregate_ranges_16x8(int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j; int m = (1 << la) - 1; uint16x8_t vect = vtstq_u16(vdupq_n_u16(m), vld1q_u16(g_kp_mask_epi16)); for (j = 0; j < lb; j++) { mtx[j].vect_u16 = vandq_u16(vect, mtx[j].vect_u16); mtx[j].vect_u16 = vshrq_n_u16(mtx[j].vect_u16, 15); __kp_m128i tmp; tmp.vect_u32 = vshrq_n_u32(mtx[j].vect_u32, 16); uint32x4_t vect_res = vandq_u32(mtx[j].vect_u32, tmp.vect_u32); int t = vaddvq_u32(vect_res) ? 1 : 0; res |= (t << j); } return res; } static int kp_aggregate_ranges_8x16(int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j; int m = (1 << la) - 1; uint8x8_t vect_mask = vld1_u8(g_kp_mask_epi8); uint8x8_t t_lo = vtst_u8(vdup_n_u8(m & 0xff), vect_mask); uint8x8_t t_hi = vtst_u8(vdup_n_u8(m >> 8), vect_mask); uint8x16_t vect = vcombine_u8(t_lo, t_hi); for (j = 0; j < lb; j++) { mtx[j].vect_u8 = vandq_u8(vect, mtx[j].vect_u8); mtx[j].vect_u8 = vshrq_n_u8(mtx[j].vect_u8, 7); __kp_m128i tmp; tmp.vect_u16 = vshrq_n_u16(mtx[j].vect_u16, 8); uint16x8_t vect_res = vandq_u16(mtx[j].vect_u16, tmp.vect_u16); int t = vaddvq_u16(vect_res) ? 1 : 0; res |= (t << j); } return res; } static int kp_cal_res_ubyte_ranges(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_RNG_U8x16(a, b, mtx); return kp_aggregate_ranges_8x16(la, lb, mtx); } static int kp_cal_res_sbyte_ranges(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_RNG_S8x16(a, b, mtx); return kp_aggregate_ranges_8x16(la, lb, mtx); } static int kp_cal_res_uword_ranges(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_RNG_U16x8(a, b, mtx); return kp_aggregate_ranges_16x8(la, lb, mtx); } static int kp_cal_res_sword_ranges(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_RNG_S16x8(a, b, mtx); return kp_aggregate_ranges_16x8(la, lb, mtx); } static int kp_cal_res_byte_equal_each(__kp_m128i a, int la, __kp_m128i b, int lb) { uint8x16_t mtx = vceqq_u8(a.vect_u8, b.vect_u8); int m0 = (la < lb) ? 0 : ((1 << la) - (1 << lb)); int m1 = 0x10000 - (1 << la); int tb = 0x10000 - (1 << lb); uint8x8_t vect_mask, vect0_lo, vect0_hi, vect1_lo, vect1_hi; uint8x8_t tmp_lo, tmp_hi, res_lo, res_hi; vect_mask = vld1_u8(g_kp_mask_epi8); vect0_lo = vtst_u8(vdup_n_u8(m0), vect_mask); vect0_hi = vtst_u8(vdup_n_u8(m0 >> 8), vect_mask); vect1_lo = vtst_u8(vdup_n_u8(m1), vect_mask); vect1_hi = vtst_u8(vdup_n_u8(m1 >> 8), vect_mask); tmp_lo = vtst_u8(vdup_n_u8(tb), vect_mask); tmp_hi = vtst_u8(vdup_n_u8(tb >> 8), vect_mask); res_lo = vbsl_u8(vect0_lo, vdup_n_u8(0), vget_low_u8(mtx)); res_hi = vbsl_u8(vect0_hi, vdup_n_u8(0), vget_high_u8(mtx)); res_lo = vbsl_u8(vect1_lo, tmp_lo, res_lo); res_hi = vbsl_u8(vect1_hi, tmp_hi, res_hi); res_lo = vand_u8(res_lo, vect_mask); res_hi = vand_u8(res_hi, vect_mask); int res = vaddv_u8(res_lo) + (vaddv_u8(res_hi) << 8); return res; } static int kp_cal_res_word_equal_each(__kp_m128i a, int la, __kp_m128i b, int lb) { uint16x8_t mtx = vceqq_u16(a.vect_u16, b.vect_u16); int m0 = (la < lb) ? 0 : ((1 << la) - (1 << lb)); int m1 = 0x100 - (1 << la); int tb = 0x100 - (1 << lb); uint16x8_t vect_mask = vld1q_u16(g_kp_mask_epi16); uint16x8_t vect0 = vtstq_u16(vdupq_n_u16(m0), vect_mask); uint16x8_t vect1 = vtstq_u16(vdupq_n_u16(m1), vect_mask); uint16x8_t tmp = vtstq_u16(vdupq_n_u16(tb), vect_mask); mtx = vbslq_u16(vect0, vdupq_n_u16(0), mtx); mtx = vbslq_u16(vect1, tmp, mtx); mtx = vandq_u16(mtx, vect_mask); return vaddvq_u16(mtx); } static int kp_aggregate_equal_ordered_8x16(int bound, int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j, k; int m1 = 0x10000 - (1 << la); uint8x16_t vect_mask = vld1q_u8(g_kp_mask_epi8); uint8x16_t vect1 = vtstq_u8(vdupq_n_u8(m1), vect_mask); uint8x16_t vect_minusone = vdupq_n_u8(-1); uint8x16_t vect_zero = vdupq_n_u8(0); for (j = 0; j < lb; j++) { mtx[j].vect_u8 = vbslq_u8(vect1, vect_minusone, mtx[j].vect_u8); } for (j = lb; j < bound; j++) { mtx[j].vect_u8 = vbslq_u8(vect1, vect_minusone, vect_zero); } uint8_t enable[16] = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; for (j = 0; j < bound; j++) { int val = 1; uint8x16_t vect_en = vld1q_u8(enable); for (k = j; k < bound && val == 1; k++) { int t = vaddvq_u8(vandq_u8(mtx[j].vect_u8, vect_en)); val = (t == bound - j) ? 1 : 0; } res = (val << j) + res; enable[bound - 1 - j] = 0; } return res; } static int kp_aggregate_equal_ordered_16x8(int bound, int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j, k; int m1 = 0x100 - (1 << la); uint16x8_t vect_mask = vld1q_u16(g_kp_mask_epi16); uint16x8_t vect1 = vtstq_u16(vdupq_n_u16(m1), vect_mask); uint16x8_t vect_minusone = vdupq_n_u16(-1); uint16x8_t vect_zero = vdupq_n_u16(0); for (j = 0; j < lb; j++) { mtx[j].vect_u16 = vbslq_u16(vect1, vect_minusone, mtx[j].vect_u16); } for (j = lb; j < bound; j++) { mtx[j].vect_u16 = vbslq_u16(vect1, vect_minusone, vect_zero); } uint16_t enable[8] = {1, 1, 1, 1, 1, 1, 1, 1}; for (j = 0; j < bound; j++) { int val = 1; uint16x8_t vect_en = vld1q_u16(enable); for (k = j; k < bound && val == 1; k++) { int t = vaddvq_u16(vandq_u16(mtx[j].vect_u16, vect_en)); val = (t == bound - j) ? 1 : 0; } res = (val << j) + res; enable[bound - 1 - j] = 0; } return res; } static int kp_cal_res_byte_equal_ordered(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_EQ_8x16(a, b, mtx); return kp_aggregate_equal_ordered_8x16(16, la, lb, mtx); } static int kp_cal_res_word_equal_ordered(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_EQ_16x8(a, b, mtx); return kp_aggregate_equal_ordered_16x8(8, la, lb, mtx); } typedef enum { KP_CMP_UBYTE_EQUAL_ANY, KP_CMP_UWORD_EQUAL_ANY, KP_CMP_SBYTE_EQUAL_ANY, KP_CMP_SWORD_EQUAL_ANY, KP_CMP_UBYTE_RANGES, KP_CMP_UWORD_RANGES, KP_CMP_SBYTE_RANGES, KP_CMP_SWORD_RANGES, KP_CMP_UBYTE_EQUAL_EACH, KP_CMP_UWORD_EQUAL_EACH, KP_CMP_SBYTE_EQUAL_EACH, KP_CMP_SWORD_EQUAL_EACH, KP_CMP_UBYTE_EQUAL_ORDERED, KP_CMP_UWORD_EQUAL_ORDERED, KP_CMP_SBYTE_EQUAL_ORDERED, KP_CMP_SWORD_EQUAL_ORDERED } _KP_MM_CMPESTR_ENUM; typedef int (*KP_CMPESTR)(__kp_m128i a, int la, __kp_m128i b, int lb); typedef struct { _KP_MM_CMPESTR_ENUM cmpintEnum; KP_CMPESTR cmpFun; } KP_CmpestrFuncList; static KP_CmpestrFuncList g_kp_cmpestrFuncList[] = {{KP_CMP_UBYTE_EQUAL_ANY, kp_cal_res_byte_equal_any}, {KP_CMP_UWORD_EQUAL_ANY, kp_cal_res_word_equal_any}, {KP_CMP_SBYTE_EQUAL_ANY, kp_cal_res_byte_equal_any}, {KP_CMP_SWORD_EQUAL_ANY, kp_cal_res_word_equal_any}, {KP_CMP_UBYTE_RANGES, kp_cal_res_ubyte_ranges}, {KP_CMP_UWORD_RANGES, kp_cal_res_uword_ranges}, {KP_CMP_SBYTE_RANGES, kp_cal_res_sbyte_ranges}, {KP_CMP_SWORD_RANGES, kp_cal_res_sword_ranges}, {KP_CMP_UBYTE_EQUAL_EACH, kp_cal_res_byte_equal_each}, {KP_CMP_UWORD_EQUAL_EACH, kp_cal_res_word_equal_each}, {KP_CMP_SBYTE_EQUAL_EACH, kp_cal_res_byte_equal_each}, {KP_CMP_SWORD_EQUAL_EACH, kp_cal_res_word_equal_each}, {KP_CMP_UBYTE_EQUAL_ORDERED, kp_cal_res_byte_equal_ordered}, {KP_CMP_UWORD_EQUAL_ORDERED, kp_cal_res_word_equal_ordered}, {KP_CMP_SBYTE_EQUAL_ORDERED, kp_cal_res_byte_equal_ordered}, {KP_CMP_SWORD_EQUAL_ORDERED, kp_cal_res_word_equal_ordered}}; KP_FORCE_INLINE int kp_neg_fun(int res, int lb, int imm8, int bound) { int m; switch (imm8 & 0x30) { case _KP_SIDD_NEGATIVE_POLARITY: res ^= 0xffffffff; break; case _KP_SIDD_MASKED_NEGATIVE_POLARITY: m = (1 << lb) - 1; res ^= m; break; default: break; } return res & ((bound == 8) ? 0xFF : 0xFFFF); } int __remill_simd__mm_cmpestri(__kp_m128i a, int la, __kp_m128i b, int lb, const int imm8, int *intRes2); __kp_m128i __remill_simd__mm_cmpestrm(__kp_m128i a, int la, __kp_m128i b, int lb, const int imm8, int *intRes2); #endif // KUNPENG_TRANS_H - Press Esc, type :wq!, and press Enter to save the settings and exit.
- Create a KunpengTrans.h file.
- Create a KunpengTrans.cpp file.
- Create a KunpengTrans.cpp file.
vi KunpengTrans.cpp
- Press i to enter the edit mode and add the following content:
/* * @Description: KunpengTrans.h * @Copyright (c) Huawei Technologies Co., Ltd. 2020-2020. All rights reserved. */ #ifndef KUNPENG_TRANS_H #define KUNPENG_TRANS_H #include <string.h> #include <KunpengTrans.h> /* ATTENTION: * Please set KUNPENG_CPU_FREQUENCY_MHZ to the actual cpu frequecy of your running environment. */ const int KUNPENG_CPU_FREQUENCY_MHZ = 2600; const int EAX_LEAF = 7; const int SHIFT_THREE = 3; /* CASE eax = 0; Highest Function Parameter and Manufacturer ID */ void GetCPUManuID(unsigned int *ebx, unsigned int *edx, unsigned int *ecx) { // ID str = "KunpengHisil" char b_str[] = "Kunp"; char d_str[] = "engH"; char c_str[] = "isil"; *ebx = *(unsigned int *)b_str; *edx = *(unsigned int *)d_str; *ecx = *(unsigned int *)c_str; } // %ecx #define bit_SSE3 (1 << 0) #define bit_PCLMUL (1 << 1) #define bit_LZCNT (1 << 5) #define bit_SSSE3 (1 << 9) #define bit_FMA (1 << 12) #define bit_CMPXCHG16B (1 << 13) #define bit_SSE4_1 (1 << 19) #define bit_SSE4_2 (1 << 20) #define bit_MOVBE (1 << 22) #define bit_POPCNT (1 << 23) #define bit_AES (1 << 25) #define bit_XSAVE (1 << 26) #define bit_OSXSAVE (1 << 27) #define bit_AVX (1 << 28) #define bit_F16C (1 << 29) #define bit_RDRND (1 << 30) // %edx #define bit_CMPXCHG8B (1 << 8) #define bit_CMOV (1 << 15) #define bit_MMX (1 << 23) #define bit_FXSAVE (1 << 24) #define bit_SSE (1 << 25) #define bit_SSE2 (1 << 26) /* CASE eax = 1; Processor Info and Feature Bits * Skylake: eax = 0x00050654, ebx = 0x43400800, ecx = 0x7ffefbf7, edx = 0xbfebfbff // Sampled at 6148 * Cascade Lake : eax = 0x00050657, ebx = 0x08400800, ecx = 0x7ffefbf7, edx = 0xbfebfbff // Sampled at 6248 */ void GetCPUFeature(unsigned int *eax, unsigned int *ebx, unsigned int *ecx, unsigned int *edx) { *eax = 0x0; // Processor Info not defined in kunpeng; *ebx = 0x0; // Additional Info not defined in kunpeng; *ecx = bit_SSE3 | bit_LZCNT | bit_SSSE3 | bit_SSE4_1 | bit_SSE4_2 | bit_POPCNT; *edx = bit_MMX | bit_SSE | bit_SSE2; } /* %ebx */ #define bit_FSGSBASE (1 << 0) #define bit_SGX (1 << 2) #define bit_BMI (1 << 3) #define bit_HLE (1 << 4) #define bit_AVX2 (1 << 5) #define bit_BMI2 (1 << 8) #define bit_RTM (1 << 11) #define bit_MPX (1 << 14) #define bit_AVX512F (1 << 16) #define bit_AVX512DQ (1 << 17) #define bit_RDSEED (1 << 18) #define bit_ADX (1 << 19) #define bit_AVX512IFMA (1 << 21) #define bit_CLFLUSHOPT (1 << 23) #define bit_CLWB (1 << 24) #define bit_AVX512PF (1 << 26) #define bit_AVX512ER (1 << 27) #define bit_AVX512CD (1 << 28) #define bit_SHA (1 << 29) #define bit_AVX512BW (1 << 30) #define bit_AVX512VL (1u << 31) /* %ecx */ #define bit_PREFETCHWT1 (1 << 0) #define bit_AVX512VBMI (1 << 1) #define bit_PKU (1 << 3) #define bit_OSPKE (1 << 4) #define bit_AVX512VBMI2 (1 << 6) #define bit_SHSTK (1 << 7) #define bit_GFNI (1 << 8) #define bit_VAES (1 << 9) #define bit_AVX512VNNI (1 << 11) #define bit_VPCLMULQDQ (1 << 10) #define bit_AVX512BITALG (1 << 12) #define bit_AVX512VPOPCNTDQ (1 << 14) #define bit_RDPID (1 << 22) #define bit_MOVDIRI (1 << 27) #define bit_MOVDIR64B (1 << 28) /* %edx */ #define bit_AVX5124VNNIW (1 << 2) #define bit_AVX5124FMAPS (1 << 3) #define bit_IBT (1 << 20) #define bit_PCONFIG (1 << 18) /* CASE eax = 7, ecx = 0; Extended Features * Skylake: eax = 0x00000000, ebx = 0xd39ffffb, ecx = 0x00000018, edx = 0x9c002400 // Sampled at 6148 * Cascade Lake : eax = 0x00000000, ebx = 0xd39ffffb, ecx = 0x00000818, edx = 0xbc000400 // Sampled at 6248 */ void GetExtendCPUFeature(unsigned int *eax, unsigned int *ebx, unsigned int *ecx, unsigned int *edx) { *eax = 0x0; *ebx = 0x0; *ecx = 0x0; *edx = 0x0; } /* * GetSupportedCPUID only supported while eax = 0, eax = 1 , eax = 7 and ecx = 0 * In other cases always return 0x0; * chipID support 1,2 -- 1 for skylake server eg. 6148; 2 for Cascade Lake 6248 */ void GetSupportedCPUID(unsigned int *eax, unsigned int *ebx, unsigned int *ecx, unsigned int *edx) { unsigned int leaf = *eax; unsigned int count = *ecx; if (leaf == 0) { *eax = 0x7; GetCPUManuID(ebx, edx, ecx); } else if (leaf == 1) { GetCPUFeature(eax, ebx, ecx, edx); } else if (leaf == EAX_LEAF && count == 0) { GetExtendCPUFeature(eax, ebx, ecx, edx); } else { *eax = 0x0; *ebx = 0x0; *ecx = 0x0; *edx = 0x0; } } void RepStos(void *dest, unsigned long long src, unsigned long long len, unsigned width, int df) { unsigned n = width >> SHIFT_THREE; unsigned char *d = (unsigned char *)dest; unsigned i; unsigned j; if (df == 1) { for (i = 0; i < len; i++) { for (j = 0; j < n; j++) { d[j] = src >> (j << SHIFT_THREE); } d -= n; } } else { for (i = 0; i < len; i++) { for (j = 0; j < n; j++) { d[j] = src >> (j << SHIFT_THREE); } d += n; } } return; } void RepStosB(void *dest, unsigned long long src, unsigned long long len, int DF) { unsigned char *s = (unsigned char *)dest; unsigned char *e = (unsigned char *)dest; if (DF) { s = s - len; } else { e = e + len; } while (s < e) { *s++ = src; } } void RepStosW(void *dest, unsigned long long src, unsigned long long len, int DF) { unsigned short *s = (unsigned short *)dest; unsigned short *e = (unsigned short *)dest; if (DF) { s = s - len; } else { e = e + len; } while (s < e) { *s++ = src; } } void RepStosD(void *dest, unsigned long long src, unsigned long long len, int DF) { unsigned int *s = (unsigned int *)dest; unsigned int *e = (unsigned int *)dest; if (DF) { s = s - len; } else { e = e + len; } while (s < e) { *s++ = src; } } void RepStosQ(void *dest, unsigned long long src, unsigned long long len, int DF) { unsigned long long *s = (unsigned long long *)dest; unsigned long long *e = (unsigned long long *)dest; if (DF) { s = s - len; } else { e = e + len; } while (s < e) { *s++ = src; } } void RepMovs(void *dest, void *src, unsigned long long len, unsigned width, int df) { unsigned n = len * (width >> SHIFT_THREE); char *d = NULL; char *s = NULL; if (df == 1) { d = (char *)dest - (n - 1); s = (char *)src - (n - 1); } else { d = (char *)dest; s = (char *)src; } memcpy(d, s, n); // This is an insecure function. If necessary, change it to a secure function. return; } #include <arm_neon.h> #define KP_FORCE_INLINE static inline __attribute__((always_inline)) typedef union { int8x16_t vect_s8; int16x8_t vect_s16; int32x4_t vect_s32; int64x2_t vect_s64; uint8x16_t vect_u8; uint16x8_t vect_u16; uint32x4_t vect_u32; uint64x2_t vect_u64; } __kp_m128i; #define _KP_SIDD_NEGATIVE_POLARITY 0x10 // negate results #define _KP_SIDD_MASKED_NEGATIVE_POLARITY 0x30 // negate results only before end of string static uint16_t g_kp_mask_epi16[8] __attribute__((aligned(16))) = { 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80 }; static uint8_t g_kp_mask_epi8[16] __attribute__((aligned(16))) = { 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80 }; #define KP_PCMPSTR_EQ_16x8(a, b, mtx) \ { \ mtx[0].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 0)), a.vect_u16); \ mtx[1].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 1)), a.vect_u16); \ mtx[2].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 2)), a.vect_u16); \ mtx[3].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 3)), a.vect_u16); \ mtx[4].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 4)), a.vect_u16); \ mtx[5].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 5)), a.vect_u16); \ mtx[6].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 6)), a.vect_u16); \ mtx[7].vect_u16 = vceqq_u16(vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 7)), a.vect_u16); \ } #define KP_PCMPSTR_EQ_8x16(a, b, mtx) \ { \ mtx[0].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 0)), a.vect_u8); \ mtx[1].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 1)), a.vect_u8); \ mtx[2].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 2)), a.vect_u8); \ mtx[3].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 3)), a.vect_u8); \ mtx[4].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 4)), a.vect_u8); \ mtx[5].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 5)), a.vect_u8); \ mtx[6].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 6)), a.vect_u8); \ mtx[7].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 7)), a.vect_u8); \ mtx[8].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 8)), a.vect_u8); \ mtx[9].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 9)), a.vect_u8); \ mtx[10].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 10)), a.vect_u8); \ mtx[11].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 11)), a.vect_u8); \ mtx[12].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 12)), a.vect_u8); \ mtx[13].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 13)), a.vect_u8); \ mtx[14].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 14)), a.vect_u8); \ mtx[15].vect_u8 = vceqq_u8(vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 15)), a.vect_u8); \ } #define KP_PCMPSTR_RNG_U16x8(a, b, mtx) \ { \ uint16x8_t vect_b[8]; \ __kp_m128i mask; \ mask.vect_u32 = vdupq_n_u32(0xffff); \ vect_b[0] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 0)); \ vect_b[1] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 1)); \ vect_b[2] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 2)); \ vect_b[3] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 3)); \ vect_b[4] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 4)); \ vect_b[5] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 5)); \ vect_b[6] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 6)); \ vect_b[7] = vdupq_n_u16(vgetq_lane_u16(b.vect_u16, 7)); \ int i; \ for (i = 0; i < 8; i++) { \ mtx[i].vect_u16 = vbslq_u16(mask.vect_u16, vcgeq_u16(vect_b[i], a.vect_u16), \ vcleq_u16(vect_b[i], a.vect_u16)); \ } \ } #define KP_PCMPSTR_RNG_S16x8(a, b, mtx) \ { \ int16x8_t vect_b[8]; \ __kp_m128i mask; \ mask.vect_u32 = vdupq_n_u32(0xffff); \ vect_b[0] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 0)); \ vect_b[1] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 1)); \ vect_b[2] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 2)); \ vect_b[3] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 3)); \ vect_b[4] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 4)); \ vect_b[5] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 5)); \ vect_b[6] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 6)); \ vect_b[7] = vdupq_n_s16(vgetq_lane_s16(b.vect_s16, 7)); \ int i; \ for (i = 0; i < 8; i++) { \ mtx[i].vect_u16 = vbslq_u16(mask.vect_u16, vcgeq_s16(vect_b[i], a.vect_s16), \ vcleq_s16(vect_b[i], a.vect_s16)); \ } \ } #define KP_PCMPSTR_RNG_U8x16(a, b, mtx) \ { \ uint8x16_t vect_b[16]; \ __kp_m128i mask; \ mask.vect_u16 = vdupq_n_u16(0xff); \ vect_b[0] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 0)); \ vect_b[1] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 1)); \ vect_b[2] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 2)); \ vect_b[3] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 3)); \ vect_b[4] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 4)); \ vect_b[5] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 5)); \ vect_b[6] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 6)); \ vect_b[7] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 7)); \ vect_b[8] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 8)); \ vect_b[9] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 9)); \ vect_b[10] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 10)); \ vect_b[11] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 11)); \ vect_b[12] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 12)); \ vect_b[13] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 13)); \ vect_b[14] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 14)); \ vect_b[15] = vdupq_n_u8(vgetq_lane_u8(b.vect_u8, 15)); \ int i; \ for (i = 0; i < 16; i++) { \ mtx[i].vect_u8 = vbslq_u8(mask.vect_u8, vcgeq_u8(vect_b[i], a.vect_u8), vcleq_u8(vect_b[i], a.vect_u8)); \ } \ } #define KP_PCMPSTR_RNG_S8x16(a, b, mtx) \ { \ int8x16_t vect_b[16]; \ __kp_m128i mask; \ mask.vect_u16 = vdupq_n_u16(0xff); \ vect_b[0] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 0)); \ vect_b[1] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 1)); \ vect_b[2] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 2)); \ vect_b[3] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 3)); \ vect_b[4] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 4)); \ vect_b[5] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 5)); \ vect_b[6] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 6)); \ vect_b[7] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 7)); \ vect_b[8] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 8)); \ vect_b[9] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 9)); \ vect_b[10] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 10)); \ vect_b[11] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 11)); \ vect_b[12] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 12)); \ vect_b[13] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 13)); \ vect_b[14] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 14)); \ vect_b[15] = vdupq_n_s8(vgetq_lane_s8(b.vect_s8, 15)); \ int i; \ for (i = 0; i < 16; i++) { \ mtx[i].vect_u8 = vbslq_u8(mask.vect_u8, vcgeq_s8(vect_b[i], a.vect_s8), vcleq_s8(vect_b[i], a.vect_s8)); \ } \ } static int kp_aggregate_equal_any_8x16(int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j; int m = (1 << la) - 1; uint8x8_t vect_mask = vld1_u8(g_kp_mask_epi8); uint8x8_t t_lo = vtst_u8(vdup_n_u8(m & 0xff), vect_mask); uint8x8_t t_hi = vtst_u8(vdup_n_u8(m >> 8), vect_mask); uint8x16_t vect = vcombine_u8(t_lo, t_hi); for (j = 0; j < lb; j++) { mtx[j].vect_u8 = vandq_u8(vect, mtx[j].vect_u8); mtx[j].vect_u8 = vshrq_n_u8(mtx[j].vect_u8, 7); int tmp = vaddvq_u8(mtx[j].vect_u8) ? 1 : 0; res |= (tmp << j); } return res; } static int kp_aggregate_equal_any_16x8(int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j; int m = (1 << la) - 1; uint16x8_t vect = vtstq_u16(vdupq_n_u16(m), vld1q_u16(g_kp_mask_epi16)); for (j = 0; j < lb; j++) { mtx[j].vect_u16 = vandq_u16(vect, mtx[j].vect_u16); mtx[j].vect_u16 = vshrq_n_u16(mtx[j].vect_u16, 15); int tmp = vaddvq_u16(mtx[j].vect_u16) ? 1 : 0; res |= (tmp << j); } return res; } static int kp_cal_res_byte_equal_any(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_EQ_8x16(a, b, mtx); return kp_aggregate_equal_any_8x16(la, lb, mtx); } static int kp_cal_res_word_equal_any(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_EQ_16x8(a, b, mtx); return kp_aggregate_equal_any_16x8(la, lb, mtx); } static int kp_aggregate_ranges_16x8(int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j; int m = (1 << la) - 1; uint16x8_t vect = vtstq_u16(vdupq_n_u16(m), vld1q_u16(g_kp_mask_epi16)); for (j = 0; j < lb; j++) { mtx[j].vect_u16 = vandq_u16(vect, mtx[j].vect_u16); mtx[j].vect_u16 = vshrq_n_u16(mtx[j].vect_u16, 15); __kp_m128i tmp; tmp.vect_u32 = vshrq_n_u32(mtx[j].vect_u32, 16); uint32x4_t vect_res = vandq_u32(mtx[j].vect_u32, tmp.vect_u32); int t = vaddvq_u32(vect_res) ? 1 : 0; res |= (t << j); } return res; } static int kp_aggregate_ranges_8x16(int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j; int m = (1 << la) - 1; uint8x8_t vect_mask = vld1_u8(g_kp_mask_epi8); uint8x8_t t_lo = vtst_u8(vdup_n_u8(m & 0xff), vect_mask); uint8x8_t t_hi = vtst_u8(vdup_n_u8(m >> 8), vect_mask); uint8x16_t vect = vcombine_u8(t_lo, t_hi); for (j = 0; j < lb; j++) { mtx[j].vect_u8 = vandq_u8(vect, mtx[j].vect_u8); mtx[j].vect_u8 = vshrq_n_u8(mtx[j].vect_u8, 7); __kp_m128i tmp; tmp.vect_u16 = vshrq_n_u16(mtx[j].vect_u16, 8); uint16x8_t vect_res = vandq_u16(mtx[j].vect_u16, tmp.vect_u16); int t = vaddvq_u16(vect_res) ? 1 : 0; res |= (t << j); } return res; } static int kp_cal_res_ubyte_ranges(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_RNG_U8x16(a, b, mtx); return kp_aggregate_ranges_8x16(la, lb, mtx); } static int kp_cal_res_sbyte_ranges(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_RNG_S8x16(a, b, mtx); return kp_aggregate_ranges_8x16(la, lb, mtx); } static int kp_cal_res_uword_ranges(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_RNG_U16x8(a, b, mtx); return kp_aggregate_ranges_16x8(la, lb, mtx); } static int kp_cal_res_sword_ranges(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_RNG_S16x8(a, b, mtx); return kp_aggregate_ranges_16x8(la, lb, mtx); } static int kp_cal_res_byte_equal_each(__kp_m128i a, int la, __kp_m128i b, int lb) { uint8x16_t mtx = vceqq_u8(a.vect_u8, b.vect_u8); int m0 = (la < lb) ? 0 : ((1 << la) - (1 << lb)); int m1 = 0x10000 - (1 << la); int tb = 0x10000 - (1 << lb); uint8x8_t vect_mask, vect0_lo, vect0_hi, vect1_lo, vect1_hi; uint8x8_t tmp_lo, tmp_hi, res_lo, res_hi; vect_mask = vld1_u8(g_kp_mask_epi8); vect0_lo = vtst_u8(vdup_n_u8(m0), vect_mask); vect0_hi = vtst_u8(vdup_n_u8(m0 >> 8), vect_mask); vect1_lo = vtst_u8(vdup_n_u8(m1), vect_mask); vect1_hi = vtst_u8(vdup_n_u8(m1 >> 8), vect_mask); tmp_lo = vtst_u8(vdup_n_u8(tb), vect_mask); tmp_hi = vtst_u8(vdup_n_u8(tb >> 8), vect_mask); res_lo = vbsl_u8(vect0_lo, vdup_n_u8(0), vget_low_u8(mtx)); res_hi = vbsl_u8(vect0_hi, vdup_n_u8(0), vget_high_u8(mtx)); res_lo = vbsl_u8(vect1_lo, tmp_lo, res_lo); res_hi = vbsl_u8(vect1_hi, tmp_hi, res_hi); res_lo = vand_u8(res_lo, vect_mask); res_hi = vand_u8(res_hi, vect_mask); int res = vaddv_u8(res_lo) + (vaddv_u8(res_hi) << 8); return res; } static int kp_cal_res_word_equal_each(__kp_m128i a, int la, __kp_m128i b, int lb) { uint16x8_t mtx = vceqq_u16(a.vect_u16, b.vect_u16); int m0 = (la < lb) ? 0 : ((1 << la) - (1 << lb)); int m1 = 0x100 - (1 << la); int tb = 0x100 - (1 << lb); uint16x8_t vect_mask = vld1q_u16(g_kp_mask_epi16); uint16x8_t vect0 = vtstq_u16(vdupq_n_u16(m0), vect_mask); uint16x8_t vect1 = vtstq_u16(vdupq_n_u16(m1), vect_mask); uint16x8_t tmp = vtstq_u16(vdupq_n_u16(tb), vect_mask); mtx = vbslq_u16(vect0, vdupq_n_u16(0), mtx); mtx = vbslq_u16(vect1, tmp, mtx); mtx = vandq_u16(mtx, vect_mask); return vaddvq_u16(mtx); } static int kp_aggregate_equal_ordered_8x16(int bound, int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j, k; int m1 = 0x10000 - (1 << la); uint8x16_t vect_mask = vld1q_u8(g_kp_mask_epi8); uint8x16_t vect1 = vtstq_u8(vdupq_n_u8(m1), vect_mask); uint8x16_t vect_minusone = vdupq_n_u8(-1); uint8x16_t vect_zero = vdupq_n_u8(0); for (j = 0; j < lb; j++) { mtx[j].vect_u8 = vbslq_u8(vect1, vect_minusone, mtx[j].vect_u8); } for (j = lb; j < bound; j++) { mtx[j].vect_u8 = vbslq_u8(vect1, vect_minusone, vect_zero); } uint8_t enable[16] = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; for (j = 0; j < bound; j++) { int val = 1; uint8x16_t vect_en = vld1q_u8(enable); for (k = j; k < bound && val == 1; k++) { int t = vaddvq_u8(vandq_u8(mtx[j].vect_u8, vect_en)); val = (t == bound - j) ? 1 : 0; } res = (val << j) + res; enable[bound - 1 - j] = 0; } return res; } static int kp_aggregate_equal_ordered_16x8(int bound, int la, int lb, __kp_m128i mtx[16]) { int res = 0; int j, k; int m1 = 0x100 - (1 << la); uint16x8_t vect_mask = vld1q_u16(g_kp_mask_epi16); uint16x8_t vect1 = vtstq_u16(vdupq_n_u16(m1), vect_mask); uint16x8_t vect_minusone = vdupq_n_u16(-1); uint16x8_t vect_zero = vdupq_n_u16(0); for (j = 0; j < lb; j++) { mtx[j].vect_u16 = vbslq_u16(vect1, vect_minusone, mtx[j].vect_u16); } for (j = lb; j < bound; j++) { mtx[j].vect_u16 = vbslq_u16(vect1, vect_minusone, vect_zero); } uint16_t enable[8] = {1, 1, 1, 1, 1, 1, 1, 1}; for (j = 0; j < bound; j++) { int val = 1; uint16x8_t vect_en = vld1q_u16(enable); for (k = j; k < bound && val == 1; k++) { int t = vaddvq_u16(vandq_u16(mtx[j].vect_u16, vect_en)); val = (t == bound - j) ? 1 : 0; } res = (val << j) + res; enable[bound - 1 - j] = 0; } return res; } static int kp_cal_res_byte_equal_ordered(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_EQ_8x16(a, b, mtx); return kp_aggregate_equal_ordered_8x16(16, la, lb, mtx); } static int kp_cal_res_word_equal_ordered(__kp_m128i a, int la, __kp_m128i b, int lb) { __kp_m128i mtx[16]; KP_PCMPSTR_EQ_16x8(a, b, mtx); return kp_aggregate_equal_ordered_16x8(8, la, lb, mtx); } typedef enum { KP_CMP_UBYTE_EQUAL_ANY, KP_CMP_UWORD_EQUAL_ANY, KP_CMP_SBYTE_EQUAL_ANY, KP_CMP_SWORD_EQUAL_ANY, KP_CMP_UBYTE_RANGES, KP_CMP_UWORD_RANGES, KP_CMP_SBYTE_RANGES, KP_CMP_SWORD_RANGES, KP_CMP_UBYTE_EQUAL_EACH, KP_CMP_UWORD_EQUAL_EACH, KP_CMP_SBYTE_EQUAL_EACH, KP_CMP_SWORD_EQUAL_EACH, KP_CMP_UBYTE_EQUAL_ORDERED, KP_CMP_UWORD_EQUAL_ORDERED, KP_CMP_SBYTE_EQUAL_ORDERED, KP_CMP_SWORD_EQUAL_ORDERED } _KP_MM_CMPESTR_ENUM; typedef int (*KP_CMPESTR)(__kp_m128i a, int la, __kp_m128i b, int lb); typedef struct { _KP_MM_CMPESTR_ENUM cmpintEnum; KP_CMPESTR cmpFun; } KP_CmpestrFuncList; static KP_CmpestrFuncList g_kp_cmpestrFuncList[] = {{KP_CMP_UBYTE_EQUAL_ANY, kp_cal_res_byte_equal_any}, {KP_CMP_UWORD_EQUAL_ANY, kp_cal_res_word_equal_any}, {KP_CMP_SBYTE_EQUAL_ANY, kp_cal_res_byte_equal_any}, {KP_CMP_SWORD_EQUAL_ANY, kp_cal_res_word_equal_any}, {KP_CMP_UBYTE_RANGES, kp_cal_res_ubyte_ranges}, {KP_CMP_UWORD_RANGES, kp_cal_res_uword_ranges}, {KP_CMP_SBYTE_RANGES, kp_cal_res_sbyte_ranges}, {KP_CMP_SWORD_RANGES, kp_cal_res_sword_ranges}, {KP_CMP_UBYTE_EQUAL_EACH, kp_cal_res_byte_equal_each}, {KP_CMP_UWORD_EQUAL_EACH, kp_cal_res_word_equal_each}, {KP_CMP_SBYTE_EQUAL_EACH, kp_cal_res_byte_equal_each}, {KP_CMP_SWORD_EQUAL_EACH, kp_cal_res_word_equal_each}, {KP_CMP_UBYTE_EQUAL_ORDERED, kp_cal_res_byte_equal_ordered}, {KP_CMP_UWORD_EQUAL_ORDERED, kp_cal_res_word_equal_ordered}, {KP_CMP_SBYTE_EQUAL_ORDERED, kp_cal_res_byte_equal_ordered}, {KP_CMP_SWORD_EQUAL_ORDERED, kp_cal_res_word_equal_ordered}}; KP_FORCE_INLINE int kp_neg_fun(int res, int lb, int imm8, int bound) { int m; switch (imm8 & 0x30) { case _KP_SIDD_NEGATIVE_POLARITY: res ^= 0xffffffff; break; case _KP_SIDD_MASKED_NEGATIVE_POLARITY: m = (1 << lb) - 1; res ^= m; break; default: break; } return res & ((bound == 8) ? 0xFF : 0xFFFF); } int __remill_simd__mm_cmpestri(__kp_m128i a, int la, __kp_m128i b, int lb, const int imm8, int *intRes2) { int bound = (imm8 & 0x01) ? 8 : 16; __asm__ __volatile__ ( "eor w0, %w[a], %w[a], asr31 \n\t" "sub %w[a], w0, %w[a], asr31 \n\t" "eor w1, %w[b], %w[b], asr31 \n\t" "sub %w[b], w1, %w[b], asr31 \n\t" "cmp %w[a], %w[bd] \n\t" "csel %w[a], %w[bd], %w[a], gt \n\t" "cmp %w[b], %w[bd] \n\t" "csel %w[b], %w[bd], %w[b], gt \n\t" :[a]"+r"(la), [b]"+r"(lb) :[bd]"r"(bound) :"w0", "w1"); int r2 = g_kp_cmpestrFuncList[imm8 & 0x0f].cmpFun(a, la, b, lb); r2 = kp_neg_fun(r2, lb, imm8, bound); *intRes2 = r2; return (r2 == 0) ? bound : ((imm8 & 0x40) ? (31 - __builtin_clz(r2)) : __builtin_ctz(r2)); } __kp_m128i __remill_simd__mm_cmpestrm(__kp_m128i a, int la, __kp_m128i b, int lb, const int imm8, int *intRes2) { __kp_m128i dst; int bound = (imm8 & 0x01) ? 8 : 16; __asm__ __volatile__ ( "eor w0, %w[a], %w[a], asr31 \n\t" "sub %w[a], w0, %w[a], asr31 \n\t" "eor w1, %w[b], %w[b], asr31 \n\t" "sub %w[b], w1, %w[b], asr31 \n\t" "cmp %w[a], %w[bd] \n\t" "csel %w[a], %w[bd], %w[a], gt \n\t" "cmp %w[b], %w[bd] \n\t" "csel %w[b], %w[bd], %w[b], gt \n\t" :[a]"+r"(la), [b]"+r"(lb) :[bd]"r"(bound) :"w0", "w1"); int r2 = g_kp_cmpestrFuncList[imm8 & 0x0f].cmpFun(a, la, b, lb); r2 = kp_neg_fun(r2, lb, imm8, bound); *intRes2 = r2; dst.vect_u8 = vdupq_n_u8(0); if (imm8 & 0x40) { if (bound == 8) { uint16x8_t tmp = vtstq_u16(vdupq_n_u16(r2), vld1q_u16(g_kp_mask_epi16)); dst.vect_u16 = vbslq_u16(tmp, vdupq_n_u16(-1), dst.vect_u16); } else { uint8x16_t vect_r2 = vcombine_u8(vdup_n_u8(r2), vdup_n_u8(r2 >> 8)); uint8x16_t tmp = vtstq_u8(vect_r2, vld1q_u8(g_kp_mask_epi8)); dst.vect_u8 = vbslq_u8(tmp, vdupq_n_u8(-1), dst.vect_u8); } } else { if (bound == 16) { dst.vect_u16 = vsetq_lane_u16(r2 & 0xffff, dst.vect_u16, 0); } else { dst.vect_u8 = vsetq_lane_u8(r2 & 0xff, dst.vect_u8, 0); } } return dst; } #endif // KUNPENG_TRANS_H - Press Esc, type :wq!, and press Enter to save the settings and exit.
- Create a KunpengTrans.cpp file.
- Upload KunpengTrans.h and KunpengTrans.cpp to the src directory of the MUSCLE source code.
- Perform the compilation.
make
After the compilation is successful, the executable file is generated in the /muscle/src/o directory.
Parent topic: Muscle 5.0.1428 Porting Guide (CentOS 7.9)