Author: Michael R. Crusoe <michael.crusoe@gmail.com>
Description: Add support for more architectures

using the SIMD Everywhere library
--- minimap2.orig/ksw2_extd2_sse.c
+++ minimap2/ksw2_extd2_sse.c
@@ -3,29 +3,19 @@
 #include <assert.h>
 #include "ksw2.h"
 
-#ifdef __SSE2__
-#include <emmintrin.h>
+#define SIMDE_ENABLE_NATIVE_ALIASES
+#include "simde/x86/sse4.1.h"
 
-#ifdef KSW_SSE2_ONLY
-#undef __SSE4_1__
-#endif
-
-#ifdef __SSE4_1__
-#include <smmintrin.h>
-#endif
-
-#ifdef KSW_CPU_DISPATCH
-#ifdef __SSE4_1__
+#if defined(SIMDE_SSE4_1_NATIVE)
 void ksw_extd2_sse41(void *km, int qlen, const uint8_t *query, int tlen, const uint8_t *target, int8_t m, const int8_t *mat,
 				   int8_t q, int8_t e, int8_t q2, int8_t e2, int w, int zdrop, int end_bonus, int flag, ksw_extz_t *ez)
-#else
+#elif defined(SIMDE_SSE2_NATIVE)
 void ksw_extd2_sse2(void *km, int qlen, const uint8_t *query, int tlen, const uint8_t *target, int8_t m, const int8_t *mat,
-				   int8_t q, int8_t e, int8_t q2, int8_t e2, int w, int zdrop, int end_bonus, int flag, ksw_extz_t *ez)
-#endif
+                                  int8_t q, int8_t e, int8_t q2, int8_t e2, int w, int zdrop, int end_bonus, int flag, ksw_extz_t *ez)
 #else
 void ksw_extd2_sse(void *km, int qlen, const uint8_t *query, int tlen, const uint8_t *target, int8_t m, const int8_t *mat,
-				   int8_t q, int8_t e, int8_t q2, int8_t e2, int w, int zdrop, int end_bonus, int flag, ksw_extz_t *ez)
-#endif // ~KSW_CPU_DISPATCH
+                                  int8_t q, int8_t e, int8_t q2, int8_t e2, int w, int zdrop, int end_bonus, int flag, ksw_extz_t *ez)
+#endif
 {
 #define __dp_code_block1 \
 	z = _mm_load_si128(&s[t]); \
@@ -161,13 +151,8 @@
 				st = _mm_loadu_si128((__m128i*)&qrr[t]);
 				mask = _mm_or_si128(_mm_cmpeq_epi8(sq, m1_), _mm_cmpeq_epi8(st, m1_));
 				tmp = _mm_cmpeq_epi8(sq, st);
-#ifdef __SSE4_1__
 				tmp = _mm_blendv_epi8(sc_mis_, sc_mch_, tmp);
 				tmp = _mm_blendv_epi8(tmp,     sc_N_,   mask);
-#else
-				tmp = _mm_or_si128(_mm_andnot_si128(tmp,  sc_mis_), _mm_and_si128(tmp,  sc_mch_));
-				tmp = _mm_or_si128(_mm_andnot_si128(mask, tmp),     _mm_and_si128(mask, sc_N_));
-#endif
 				_mm_storeu_si128((__m128i*)((int8_t*)s + t), tmp);
 			}
 		} else {
@@ -184,7 +169,6 @@
 			for (t = st_; t <= en_; ++t) {
 				__m128i z, a, b, a2, b2, xt1, x2t1, vt1, ut, tmp;
 				__dp_code_block1;
-#ifdef __SSE4_1__
 				z = _mm_max_epi8(z, a);
 				z = _mm_max_epi8(z, b);
 				z = _mm_max_epi8(z, a2);
@@ -195,27 +179,6 @@
 				_mm_store_si128(&y[t],  _mm_sub_epi8(_mm_max_epi8(b,  zero_), qe_));
 				_mm_store_si128(&x2[t], _mm_sub_epi8(_mm_max_epi8(a2, zero_), qe2_));
 				_mm_store_si128(&y2[t], _mm_sub_epi8(_mm_max_epi8(b2, zero_), qe2_));
-#else
-				tmp = _mm_cmpgt_epi8(a,  z);
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, a));
-				tmp = _mm_cmpgt_epi8(b,  z);
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, b));
-				tmp = _mm_cmpgt_epi8(a2, z);
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, a2));
-				tmp = _mm_cmpgt_epi8(b2, z);
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, b2));
-				tmp = _mm_cmplt_epi8(sc_mch_, z);
-				z = _mm_or_si128(_mm_and_si128(tmp, sc_mch_), _mm_andnot_si128(tmp, z));
-				__dp_code_block2;
-				tmp = _mm_cmpgt_epi8(a, zero_);
-				_mm_store_si128(&x[t],  _mm_sub_epi8(_mm_and_si128(tmp, a),  qe_));
-				tmp = _mm_cmpgt_epi8(b, zero_);
-				_mm_store_si128(&y[t],  _mm_sub_epi8(_mm_and_si128(tmp, b),  qe_));
-				tmp = _mm_cmpgt_epi8(a2, zero_);
-				_mm_store_si128(&x2[t], _mm_sub_epi8(_mm_and_si128(tmp, a2), qe2_));
-				tmp = _mm_cmpgt_epi8(b2, zero_);
-				_mm_store_si128(&y2[t], _mm_sub_epi8(_mm_and_si128(tmp, b2), qe2_));
-#endif
 			}
 		} else if (!(flag&KSW_EZ_RIGHT)) { // gap left-alignment
 			__m128i *pr = p + (size_t)r * n_col_ - st_;
@@ -223,7 +186,6 @@
 			for (t = st_; t <= en_; ++t) {
 				__m128i d, z, a, b, a2, b2, xt1, x2t1, vt1, ut, tmp;
 				__dp_code_block1;
-#ifdef __SSE4_1__
 				d = _mm_and_si128(_mm_cmpgt_epi8(a, z), _mm_set1_epi8(1));       // d = a  > z? 1 : 0
 				z = _mm_max_epi8(z, a);
 				d = _mm_blendv_epi8(d, _mm_set1_epi8(2), _mm_cmpgt_epi8(b,  z)); // d = b  > z? 2 : d
@@ -233,22 +195,6 @@
 				d = _mm_blendv_epi8(d, _mm_set1_epi8(4), _mm_cmpgt_epi8(b2, z)); // d = a2 > z? 3 : d
 				z = _mm_max_epi8(z, b2);
 				z = _mm_min_epi8(z, sc_mch_);
-#else // we need to emulate SSE4.1 intrinsics _mm_max_epi8() and _mm_blendv_epi8()
-				tmp = _mm_cmpgt_epi8(a,  z);
-				d = _mm_and_si128(tmp, _mm_set1_epi8(1));
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, a));
-				tmp = _mm_cmpgt_epi8(b,  z);
-				d = _mm_or_si128(_mm_andnot_si128(tmp, d), _mm_and_si128(tmp, _mm_set1_epi8(2)));
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, b));
-				tmp = _mm_cmpgt_epi8(a2, z);
-				d = _mm_or_si128(_mm_andnot_si128(tmp, d), _mm_and_si128(tmp, _mm_set1_epi8(3)));
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, a2));
-				tmp = _mm_cmpgt_epi8(b2, z);
-				d = _mm_or_si128(_mm_andnot_si128(tmp, d), _mm_and_si128(tmp, _mm_set1_epi8(4)));
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, b2));
-				tmp = _mm_cmplt_epi8(sc_mch_, z);
-				z = _mm_or_si128(_mm_and_si128(tmp, sc_mch_), _mm_andnot_si128(tmp, z));
-#endif
 				__dp_code_block2;
 				tmp = _mm_cmpgt_epi8(a, zero_);
 				_mm_store_si128(&x[t],  _mm_sub_epi8(_mm_and_si128(tmp, a),  qe_));
@@ -270,7 +216,6 @@
 			for (t = st_; t <= en_; ++t) {
 				__m128i d, z, a, b, a2, b2, xt1, x2t1, vt1, ut, tmp;
 				__dp_code_block1;
-#ifdef __SSE4_1__
 				d = _mm_andnot_si128(_mm_cmpgt_epi8(z, a), _mm_set1_epi8(1));    // d = z > a?  0 : 1
 				z = _mm_max_epi8(z, a);
 				d = _mm_blendv_epi8(_mm_set1_epi8(2), d, _mm_cmpgt_epi8(z, b));  // d = z > b?  d : 2
@@ -280,22 +225,6 @@
 				d = _mm_blendv_epi8(_mm_set1_epi8(4), d, _mm_cmpgt_epi8(z, b2)); // d = z > b2? d : 4
 				z = _mm_max_epi8(z, b2);
 				z = _mm_min_epi8(z, sc_mch_);
-#else // we need to emulate SSE4.1 intrinsics _mm_max_epi8() and _mm_blendv_epi8()
-				tmp = _mm_cmpgt_epi8(z, a);
-				d = _mm_andnot_si128(tmp, _mm_set1_epi8(1));
-				z = _mm_or_si128(_mm_and_si128(tmp, z), _mm_andnot_si128(tmp, a));
-				tmp = _mm_cmpgt_epi8(z, b);
-				d = _mm_or_si128(_mm_and_si128(tmp, d), _mm_andnot_si128(tmp, _mm_set1_epi8(2)));
-				z = _mm_or_si128(_mm_and_si128(tmp, z), _mm_andnot_si128(tmp, b));
-				tmp = _mm_cmpgt_epi8(z, a2);
-				d = _mm_or_si128(_mm_and_si128(tmp, d), _mm_andnot_si128(tmp, _mm_set1_epi8(3)));
-				z = _mm_or_si128(_mm_and_si128(tmp, z), _mm_andnot_si128(tmp, a2));
-				tmp = _mm_cmpgt_epi8(z, b2);
-				d = _mm_or_si128(_mm_and_si128(tmp, d), _mm_andnot_si128(tmp, _mm_set1_epi8(4)));
-				z = _mm_or_si128(_mm_and_si128(tmp, z), _mm_andnot_si128(tmp, b2));
-				tmp = _mm_cmplt_epi8(sc_mch_, z);
-				z = _mm_or_si128(_mm_and_si128(tmp, sc_mch_), _mm_andnot_si128(tmp, z));
-#endif
 				__dp_code_block2;
 				tmp = _mm_cmpgt_epi8(zero_, a);
 				_mm_store_si128(&x[t],  _mm_sub_epi8(_mm_andnot_si128(tmp, a),  qe_));
@@ -330,13 +259,8 @@
 					_mm_storeu_si128((__m128i*)&H[t], H1);
 					t_ = _mm_set1_epi32(t);
 					tmp = _mm_cmpgt_epi32(H1, max_H_);
-#ifdef __SSE4_1__
 					max_H_ = _mm_blendv_epi8(max_H_, H1, tmp);
 					max_t_ = _mm_blendv_epi8(max_t_, t_, tmp);
-#else
-					max_H_ = _mm_or_si128(_mm_and_si128(tmp, H1), _mm_andnot_si128(tmp, max_H_));
-					max_t_ = _mm_or_si128(_mm_and_si128(tmp, t_), _mm_andnot_si128(tmp, max_t_));
-#endif
 				}
 				_mm_storeu_si128((__m128i*)HH, max_H_);
 				_mm_storeu_si128((__m128i*)tt, max_t_);
@@ -391,4 +315,3 @@
 		kfree(km, mem2); kfree(km, off);
 	}
 }
-#endif // __SSE2__
--- minimap2.orig/ksw2_exts2_sse.c
+++ minimap2/ksw2_exts2_sse.c
@@ -3,29 +3,19 @@
 #include <assert.h>
 #include "ksw2.h"
 
-#ifdef __SSE2__
-#include <emmintrin.h>
+#define SIMDE_ENABLE_NATIVE_ALIASES
+#include "simde/x86/sse4.1.h"
 
-#ifdef KSW_SSE2_ONLY
-#undef __SSE4_1__
-#endif
-
-#ifdef __SSE4_1__
-#include <smmintrin.h>
-#endif
-
-#ifdef KSW_CPU_DISPATCH
-#ifdef __SSE4_1__
+#if defined(SIMDE_SSE4_1_NATIVE)
 void ksw_exts2_sse41(void *km, int qlen, const uint8_t *query, int tlen, const uint8_t *target, int8_t m, const int8_t *mat,
 				   int8_t q, int8_t e, int8_t q2, int8_t noncan, int zdrop, int8_t junc_bonus, int flag, const uint8_t *junc, ksw_extz_t *ez)
-#else
+#elif defined(SIMDE_SSE2_NATIVE)
 void ksw_exts2_sse2(void *km, int qlen, const uint8_t *query, int tlen, const uint8_t *target, int8_t m, const int8_t *mat,
 				   int8_t q, int8_t e, int8_t q2, int8_t noncan, int zdrop, int8_t junc_bonus, int flag, const uint8_t *junc, ksw_extz_t *ez)
-#endif
 #else
 void ksw_exts2_sse(void *km, int qlen, const uint8_t *query, int tlen, const uint8_t *target, int8_t m, const int8_t *mat,
 				   int8_t q, int8_t e, int8_t q2, int8_t noncan, int zdrop, int8_t junc_bonus, int flag, const uint8_t *junc, ksw_extz_t *ez)
-#endif // ~KSW_CPU_DISPATCH
+#endif
 {
 #define __dp_code_block1 \
 	z = _mm_load_si128(&s[t]); \
@@ -194,13 +184,8 @@
 				st = _mm_loadu_si128((__m128i*)&qrr[t]);
 				mask = _mm_or_si128(_mm_cmpeq_epi8(sq, m1_), _mm_cmpeq_epi8(st, m1_));
 				tmp = _mm_cmpeq_epi8(sq, st);
-#ifdef __SSE4_1__
 				tmp = _mm_blendv_epi8(sc_mis_, sc_mch_, tmp);
 				tmp = _mm_blendv_epi8(tmp,     sc_N_,   mask);
-#else
-				tmp = _mm_or_si128(_mm_andnot_si128(tmp,  sc_mis_), _mm_and_si128(tmp,  sc_mch_));
-				tmp = _mm_or_si128(_mm_andnot_si128(mask, tmp),     _mm_and_si128(mask, sc_N_));
-#endif
 				_mm_storeu_si128((__m128i*)((int8_t*)s + t), tmp);
 			}
 		} else {
@@ -217,7 +202,6 @@
 			for (t = st_; t <= en_; ++t) {
 				__m128i z, a, b, a2, a2a, xt1, x2t1, vt1, ut, tmp;
 				__dp_code_block1;
-#ifdef __SSE4_1__
 				z = _mm_max_epi8(z, a);
 				z = _mm_max_epi8(z, b);
 				z = _mm_max_epi8(z, a2a);
@@ -226,23 +210,6 @@
 				_mm_store_si128(&y[t],  _mm_sub_epi8(_mm_max_epi8(b,  zero_), qe_));
 				tmp = _mm_load_si128(&donor[t]);
 				_mm_store_si128(&x2[t], _mm_sub_epi8(_mm_max_epi8(a2, tmp), q2_));
-#else
-				tmp = _mm_cmpgt_epi8(a,  z);
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, a));
-				tmp = _mm_cmpgt_epi8(b,  z);
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, b));
-				tmp = _mm_cmpgt_epi8(a2a, z);
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, a2a));
-				__dp_code_block2;
-				tmp = _mm_cmpgt_epi8(a, zero_);
-				_mm_store_si128(&x[t],  _mm_sub_epi8(_mm_and_si128(tmp, a),  qe_));
-				tmp = _mm_cmpgt_epi8(b, zero_);
-				_mm_store_si128(&y[t],  _mm_sub_epi8(_mm_and_si128(tmp, b),  qe_));
-				tmp = _mm_load_si128(&donor[t]); // TODO: check if this is correct
-				tmp = _mm_cmpgt_epi8(a2, tmp);
-				tmp = _mm_or_si128(_mm_andnot_si128(tmp, tmp), _mm_and_si128(tmp, a2));
-				_mm_store_si128(&x2[t], _mm_sub_epi8(tmp, q2_));
-#endif
 			}
 		} else if (!(flag&KSW_EZ_RIGHT)) { // gap left-alignment
 			__m128i *pr = p + r * n_col_ - st_;
@@ -250,24 +217,12 @@
 			for (t = st_; t <= en_; ++t) {
 				__m128i d, z, a, b, a2, a2a, xt1, x2t1, vt1, ut, tmp, tmp2;
 				__dp_code_block1;
-#ifdef __SSE4_1__
 				d = _mm_and_si128(_mm_cmpgt_epi8(a, z), _mm_set1_epi8(1));       // d = a  > z? 1 : 0
 				z = _mm_max_epi8(z, a);
 				d = _mm_blendv_epi8(d, _mm_set1_epi8(2), _mm_cmpgt_epi8(b,  z)); // d = b  > z? 2 : d
 				z = _mm_max_epi8(z, b);
 				d = _mm_blendv_epi8(d, _mm_set1_epi8(3), _mm_cmpgt_epi8(a2a, z)); // d = a2 > z? 3 : d
 				z = _mm_max_epi8(z, a2a);
-#else // we need to emulate SSE4.1 intrinsics _mm_max_epi8() and _mm_blendv_epi8()
-				tmp = _mm_cmpgt_epi8(a,  z);
-				d = _mm_and_si128(tmp, _mm_set1_epi8(1));
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, a));
-				tmp = _mm_cmpgt_epi8(b,  z);
-				d = _mm_or_si128(_mm_andnot_si128(tmp, d), _mm_and_si128(tmp, _mm_set1_epi8(2)));
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, b));
-				tmp = _mm_cmpgt_epi8(a2a, z);
-				d = _mm_or_si128(_mm_andnot_si128(tmp, d), _mm_and_si128(tmp, _mm_set1_epi8(3)));
-				z = _mm_or_si128(_mm_andnot_si128(tmp, z), _mm_and_si128(tmp, a2a));
-#endif
 				__dp_code_block2;
 				tmp = _mm_cmpgt_epi8(a, zero_);
 				_mm_store_si128(&x[t],  _mm_sub_epi8(_mm_and_si128(tmp, a),  qe_));
@@ -278,11 +233,7 @@
 
 				tmp2 = _mm_load_si128(&donor[t]);
 				tmp = _mm_cmpgt_epi8(a2, tmp2);
-#ifdef __SSE4_1__
 				tmp2 = _mm_max_epi8(a2, tmp2);
-#else
-				tmp2 = _mm_or_si128(_mm_andnot_si128(tmp, tmp2), _mm_and_si128(tmp, a2));
-#endif
 				_mm_store_si128(&x2[t], _mm_sub_epi8(tmp2, q2_));
 				d = _mm_or_si128(d, _mm_and_si128(tmp, _mm_set1_epi8(0x20)));
 				_mm_store_si128(&pr[t], d);
@@ -293,24 +244,12 @@
 			for (t = st_; t <= en_; ++t) {
 				__m128i d, z, a, b, a2, a2a, xt1, x2t1, vt1, ut, tmp, tmp2;
 				__dp_code_block1;
-#ifdef __SSE4_1__
 				d = _mm_andnot_si128(_mm_cmpgt_epi8(z, a), _mm_set1_epi8(1));    // d = z > a?  0 : 1
 				z = _mm_max_epi8(z, a);
 				d = _mm_blendv_epi8(_mm_set1_epi8(2), d, _mm_cmpgt_epi8(z, b));  // d = z > b?  d : 2
 				z = _mm_max_epi8(z, b);
 				d = _mm_blendv_epi8(_mm_set1_epi8(3), d, _mm_cmpgt_epi8(z, a2a)); // d = z > a2? d : 3
 				z = _mm_max_epi8(z, a2a);
-#else // we need to emulate SSE4.1 intrinsics _mm_max_epi8() and _mm_blendv_epi8()
-				tmp = _mm_cmpgt_epi8(z, a);
-				d = _mm_andnot_si128(tmp, _mm_set1_epi8(1));
-				z = _mm_or_si128(_mm_and_si128(tmp, z), _mm_andnot_si128(tmp, a));
-				tmp = _mm_cmpgt_epi8(z, b);
-				d = _mm_or_si128(_mm_and_si128(tmp, d), _mm_andnot_si128(tmp, _mm_set1_epi8(2)));
-				z = _mm_or_si128(_mm_and_si128(tmp, z), _mm_andnot_si128(tmp, b));
-				tmp = _mm_cmpgt_epi8(z, a2a);
-				d = _mm_or_si128(_mm_and_si128(tmp, d), _mm_andnot_si128(tmp, _mm_set1_epi8(3)));
-				z = _mm_or_si128(_mm_and_si128(tmp, z), _mm_andnot_si128(tmp, a2a));
-#endif
 				__dp_code_block2;
 				tmp = _mm_cmpgt_epi8(zero_, a);
 				_mm_store_si128(&x[t],  _mm_sub_epi8(_mm_andnot_si128(tmp, a),  qe_));
@@ -321,11 +260,7 @@
 
 				tmp2 = _mm_load_si128(&donor[t]);
 				tmp = _mm_cmpgt_epi8(tmp2, a2);
-#ifdef __SSE4_1__
 				tmp2 = _mm_max_epi8(tmp2, a2);
-#else
-				tmp2 = _mm_or_si128(_mm_andnot_si128(tmp, a2), _mm_and_si128(tmp, tmp2));
-#endif
 				_mm_store_si128(&x2[t], _mm_sub_epi8(tmp2, q2_));
 				d = _mm_or_si128(d, _mm_andnot_si128(tmp, _mm_set1_epi8(0x20))); // d = a > 0? 1<<5 : 0
 				_mm_store_si128(&pr[t], d);
@@ -349,13 +284,8 @@
 					_mm_storeu_si128((__m128i*)&H[t], H1);
 					t_ = _mm_set1_epi32(t);
 					tmp = _mm_cmpgt_epi32(H1, max_H_);
-#ifdef __SSE4_1__
 					max_H_ = _mm_blendv_epi8(max_H_, H1, tmp);
 					max_t_ = _mm_blendv_epi8(max_t_, t_, tmp);
-#else
-					max_H_ = _mm_or_si128(_mm_and_si128(tmp, H1), _mm_andnot_si128(tmp, max_H_));
-					max_t_ = _mm_or_si128(_mm_and_si128(tmp, t_), _mm_andnot_si128(tmp, max_t_));
-#endif
 				}
 				_mm_storeu_si128((__m128i*)HH, max_H_);
 				_mm_storeu_si128((__m128i*)tt, max_t_);
@@ -406,4 +336,3 @@
 		kfree(km, mem2); kfree(km, off);
 	}
 }
-#endif // __SSE2__
--- minimap2.orig/ksw2_extz2_sse.c
+++ minimap2/ksw2_extz2_sse.c
@@ -2,26 +2,16 @@
 #include <assert.h>
 #include "ksw2.h"
 
-#ifdef __SSE2__
-#include <emmintrin.h>
+#define SIMDE_ENABLE_NATIVE_ALIASES
+#include "simde/x86/sse4.1.h"
 
-#ifdef KSW_SSE2_ONLY
-#undef __SSE4_1__
-#endif
-
-#ifdef __SSE4_1__
-#include <smmintrin.h>
-#endif
-
-#ifdef KSW_CPU_DISPATCH
-#ifdef __SSE4_1__
+#if defined(SIMDE_SSE4_1_NATIVE)
 void ksw_extz2_sse41(void *km, int qlen, const uint8_t *query, int tlen, const uint8_t *target, int8_t m, const int8_t *mat, int8_t q, int8_t e, int w, int zdrop, int end_bonus, int flag, ksw_extz_t *ez)
-#else
+#elif defined(SIMDE_SSE2_NATIVE)
 void ksw_extz2_sse2(void *km, int qlen, const uint8_t *query, int tlen, const uint8_t *target, int8_t m, const int8_t *mat, int8_t q, int8_t e, int w, int zdrop, int end_bonus, int flag, ksw_extz_t *ez)
-#endif
 #else
 void ksw_extz2_sse(void *km, int qlen, const uint8_t *query, int tlen, const uint8_t *target, int8_t m, const int8_t *mat, int8_t q, int8_t e, int w, int zdrop, int end_bonus, int flag, ksw_extz_t *ez)
-#endif // ~KSW_CPU_DISPATCH
+#endif
 {
 #define __dp_code_block1 \
 	z = _mm_add_epi8(_mm_load_si128(&s[t]), qe2_); \
@@ -129,13 +119,8 @@
 				st = _mm_loadu_si128((__m128i*)&qrr[t]);
 				mask = _mm_or_si128(_mm_cmpeq_epi8(sq, m1_), _mm_cmpeq_epi8(st, m1_));
 				tmp = _mm_cmpeq_epi8(sq, st);
-#ifdef __SSE4_1__
 				tmp = _mm_blendv_epi8(sc_mis_, sc_mch_, tmp);
 				tmp = _mm_blendv_epi8(tmp,     sc_N_,   mask);
-#else
-				tmp = _mm_or_si128(_mm_andnot_si128(tmp,  sc_mis_), _mm_and_si128(tmp,  sc_mch_));
-				tmp = _mm_or_si128(_mm_andnot_si128(mask, tmp),     _mm_and_si128(mask, sc_N_));
-#endif
 				_mm_storeu_si128((__m128i*)((uint8_t*)s + t), tmp);
 			}
 		} else {
@@ -151,22 +136,10 @@
 			for (t = st_; t <= en_; ++t) {
 				__m128i z, a, b, xt1, vt1, ut, tmp;
 				__dp_code_block1;
-#ifdef __SSE4_1__
 				z = _mm_max_epi8(z, a);                          // z = z > a? z : a (signed)
-#else // we need to emulate SSE4.1 intrinsics _mm_max_epi8()
-				z = _mm_and_si128(z, _mm_cmpgt_epi8(z, zero_));  // z = z > 0? z : 0;
-				z = _mm_max_epu8(z, a);                          // z = max(z, a); this works because both are non-negative
-#endif
 				__dp_code_block2;
-#ifdef __SSE4_1__
 				_mm_store_si128(&x[t], _mm_max_epi8(a, zero_));
 				_mm_store_si128(&y[t], _mm_max_epi8(b, zero_));
-#else
-				tmp = _mm_cmpgt_epi8(a, zero_);
-				_mm_store_si128(&x[t], _mm_and_si128(a, tmp));
-				tmp = _mm_cmpgt_epi8(b, zero_);
-				_mm_store_si128(&y[t], _mm_and_si128(b, tmp));
-#endif
 			}
 		} else if (!(flag&KSW_EZ_RIGHT)) { // gap left-alignment
 			__m128i *pr = p + (size_t)r * n_col_ - st_;
@@ -175,16 +148,9 @@
 				__m128i d, z, a, b, xt1, vt1, ut, tmp;
 				__dp_code_block1;
 				d = _mm_and_si128(_mm_cmpgt_epi8(a, z), flag1_); // d = a > z? 1 : 0
-#ifdef __SSE4_1__
 				z = _mm_max_epi8(z, a);                          // z = z > a? z : a (signed)
 				tmp = _mm_cmpgt_epi8(b, z);
 				d = _mm_blendv_epi8(d, flag2_, tmp);             // d = b > z? 2 : d
-#else // we need to emulate SSE4.1 intrinsics _mm_max_epi8() and _mm_blendv_epi8()
-				z = _mm_and_si128(z, _mm_cmpgt_epi8(z, zero_));  // z = z > 0? z : 0;
-				z = _mm_max_epu8(z, a);                          // z = max(z, a); this works because both are non-negative
-				tmp = _mm_cmpgt_epi8(b, z);
-				d = _mm_or_si128(_mm_andnot_si128(tmp, d), _mm_and_si128(tmp, flag2_)); // d = b > z? 2 : d; emulating blendv
-#endif
 				__dp_code_block2;
 				tmp = _mm_cmpgt_epi8(a, zero_);
 				_mm_store_si128(&x[t], _mm_and_si128(tmp, a));
@@ -201,16 +167,9 @@
 				__m128i d, z, a, b, xt1, vt1, ut, tmp;
 				__dp_code_block1;
 				d = _mm_andnot_si128(_mm_cmpgt_epi8(z, a), flag1_); // d = z > a? 0 : 1
-#ifdef __SSE4_1__
 				z = _mm_max_epi8(z, a);                          // z = z > a? z : a (signed)
 				tmp = _mm_cmpgt_epi8(z, b);
 				d = _mm_blendv_epi8(flag2_, d, tmp);             // d = z > b? d : 2
-#else // we need to emulate SSE4.1 intrinsics _mm_max_epi8() and _mm_blendv_epi8()
-				z = _mm_and_si128(z, _mm_cmpgt_epi8(z, zero_));  // z = z > 0? z : 0;
-				z = _mm_max_epu8(z, a);                          // z = max(z, a); this works because both are non-negative
-				tmp = _mm_cmpgt_epi8(z, b);
-				d = _mm_or_si128(_mm_andnot_si128(tmp, flag2_), _mm_and_si128(tmp, d)); // d = z > b? d : 2; emulating blendv
-#endif
 				__dp_code_block2;
 				tmp = _mm_cmpgt_epi8(zero_, a);
 				_mm_store_si128(&x[t], _mm_andnot_si128(tmp, a));
@@ -241,13 +200,8 @@
 					_mm_storeu_si128((__m128i*)&H[t], H1);
 					t_ = _mm_set1_epi32(t);
 					tmp = _mm_cmpgt_epi32(H1, max_H_);
-#ifdef __SSE4_1__
 					max_H_ = _mm_blendv_epi8(max_H_, H1, tmp);
 					max_t_ = _mm_blendv_epi8(max_t_, t_, tmp);
-#else
-					max_H_ = _mm_or_si128(_mm_and_si128(tmp, H1), _mm_andnot_si128(tmp, max_H_));
-					max_t_ = _mm_or_si128(_mm_and_si128(tmp, t_), _mm_andnot_si128(tmp, max_t_));
-#endif
 				}
 				_mm_storeu_si128((__m128i*)HH, max_H_);
 				_mm_storeu_si128((__m128i*)tt, max_t_);
@@ -302,4 +256,3 @@
 		kfree(km, mem2); kfree(km, off);
 	}
 }
-#endif // __SSE2__
--- minimap2.orig/ksw2_ll_sse.c
+++ minimap2/ksw2_ll_sse.c
@@ -1,7 +1,8 @@
 #include <stdlib.h>
 #include <stdint.h>
 #include <string.h>
-#include <emmintrin.h>
+#define SIMDE_ENABLE_NATIVE_ALIASES
+#include "simde/x86/sse2.h"
 #include "ksw2.h"
 
 #ifdef __GNUC__
--- minimap2.orig/Makefile
+++ minimap2/Makefile
@@ -6,21 +6,17 @@
 PROG_EXTRA=	sdust minimap2-lite
 LIBS=		-lm -lz -lpthread
 
-ifeq ($(arm_neon),) # if arm_neon is not defined
-ifeq ($(sse2only),) # if sse2only is not defined
+OBJS+=ksw2_extz2_sse.o ksw2_extd2_sse.o ksw2_exts2_sse.o
+
+ifneq ($(amd64),)
 	OBJS+=ksw2_extz2_sse41.o ksw2_extd2_sse41.o ksw2_exts2_sse41.o ksw2_extz2_sse2.o ksw2_extd2_sse2.o ksw2_exts2_sse2.o ksw2_dispatch.o
-else                # if sse2only is defined
-	OBJS+=ksw2_extz2_sse.o ksw2_extd2_sse.o ksw2_exts2_sse.o
+else ifneq ($(i386),)
+	OBJS+=ksw2_extz2_sse2.o ksw2_extd2_sse2.o ksw2_exts2_sse2.o ksw2_dispatch.o
 endif
-else				# if arm_neon is defined
-	OBJS+=ksw2_extz2_neon.o ksw2_extd2_neon.o ksw2_exts2_neon.o
-    INCLUDES+=-Isse2neon
-ifeq ($(aarch64),)	#if aarch64 is not defined
-	CFLAGS+=-D_FILE_OFFSET_BITS=64 -mfpu=neon -fsigned-char
-else				#if aarch64 is defined
+
+ifneq ($(aarch64),)	#if aarch64 is defined
 	CFLAGS+=-D_FILE_OFFSET_BITS=64 -fsigned-char
 endif
-endif
 
 .PHONY:all extra clean depend
 .SUFFIXES:.c .o
@@ -46,7 +42,7 @@
 
 # SSE-specific targets on x86/x86_64
 
-ifeq ($(arm_neon),)   # if arm_neon is defined, compile this target with the default setting (i.e. no -msse2)
+ifneq ($(amd64),)   # if amd64 is not defined, compile this target with the default setting (i.e. no -msse2)
 ksw2_ll_sse.o:ksw2_ll_sse.c ksw2.h kalloc.h
 		$(CC) -c $(CFLAGS) -msse2 $(CPPFLAGS) $(INCLUDES) $< -o $@
 endif
@@ -72,17 +68,6 @@
 ksw2_dispatch.o:ksw2_dispatch.c ksw2.h
 		$(CC) -c $(CFLAGS) -msse4.1 $(CPPFLAGS) -DKSW_CPU_DISPATCH $(INCLUDES) $< -o $@
 
-# NEON-specific targets on ARM
-
-ksw2_extz2_neon.o:ksw2_extz2_sse.c ksw2.h kalloc.h
-		$(CC) -c $(CFLAGS) $(CPPFLAGS) -DKSW_SSE2_ONLY -D__SSE2__ $(INCLUDES) $< -o $@
-
-ksw2_extd2_neon.o:ksw2_extd2_sse.c ksw2.h kalloc.h
-		$(CC) -c $(CFLAGS) $(CPPFLAGS) -DKSW_SSE2_ONLY -D__SSE2__ $(INCLUDES) $< -o $@
-
-ksw2_exts2_neon.o:ksw2_exts2_sse.c ksw2.h kalloc.h
-		$(CC) -c $(CFLAGS) $(CPPFLAGS) -DKSW_SSE2_ONLY -D__SSE2__ $(INCLUDES) $< -o $@
-
 # other non-file targets
 
 clean:
