vec_kern_unr_pref.h File Reference

Kernels to be called for array operations Unrolled and with prefetching. More...

#include "tbci/basics.h"

Include dependency graph for vec_kern_unr_pref.h:

Go to the source code of this file.

Macros
#define	ADD3(r, v1, v2, f1, f2)   r = v1 + v2
#define	SUB3(r, v1, v2, f1, f2)   r = v1 - v2
#define	MUL3(r, v1, v2, f1, f2)   r = v1 * v2
#define	CMUL3(r, v1, v2, f1, f2)   r = CPLX__ conj(v1) * v2
#define	DIV3(r, v1, v2, f1, f2)   r = v1 / v2
#define	CDIV3(r, v1, v2, f1, f2)   r = CPLX__ conj(v1) / v2
#define	ADD2(r, v1, f1, f2)   r += v1
#define	SUB2(r, v1, f1, f2)   r -= v1
#define	SUB2I(r, v1, f1, f2)   r = v1 - r
#define	MUL2(r, v1, f1, f2)   r *= v1
#define	CMUL2(r, v1, f1, f2)   r = CPLX__ conj(r) * v1
#define	CMUL2I(r, v1, f1, f2)   r *= CPLX__ conj(v1)
#define	DIV2(r, v1, f1, f2)   r /= v1
#define	DIV2I(r, v1, f1, f2)   r = v1 / r
#define	CDIV2(r, v1, f1, f2)   r = CPLX__ conj(r) / v1
#define	CDIV2I(r, v1, f1, f2)   r = CPLX__ conj(v1) / r
#define	ADD2NV(r, v1, f1, f2)   r = v1 + f2
#define	SUB2NV(r, v1, f1, f2)   r = v1 - f2
#define	MUL2NV(r, v1, f1, f2)   r = v1 * f2
#define	ADD2RV(r, v1, f1, f2)   r = f2 + v1
#define	SUB2RV(r, v1, f1, f2)   r = f2 - v1
#define	MUL2RV(r, v1, f1, f2)   r = f2 * v1
#define	DIV2RV(r, v1, f1, f2)   r = f2 / v1
#define	ADD1NV(r, f1, f2)   r += f2
#define	SUB1NV(r, f1, f2)   r -= f2
#define	SUB1RV(r, f1, f2)   r = f2 - r
#define	MUL1NV(r, f1, f2)   r *= f2
#define	DIV1NV(r, f1, f2)   r /= f2
#define	DIV1RV(r, f1, f2)   r = f2 / r
#define	ADD1RV(r, f1, f2)   r = f2 + r
#define	MUL1RV(r, f1, f2)   r = f2 * r;
#define	ADD2NS(r, v1, f1, f2)   r += f2*v1
#define	SUB2NS(r, v1, f1, f2)   r -= f2*v1
#define	SUB2RS(r, v1, f1, f2)   r = f2*v1 - r
#define	ADD3NS(r, v1, v2, f1, f2)   r = v1 + f2*v2
#define	SUB3NS(r, v1, v2, f1, f2)   r = v1 - f2*v2
#define	ADD3SN(r, v1, v2, f1, f2)   r = f2*v1 + v2
#define	SUB3SN(r, v1, v2, f1, f2)   r = f2*v1 - v2
#define	ADD3SS(r, v1, v2, f1, f2)   r = f1*v1 + f2*v2
#define	SUB3SS(r, v1, v2, f1, f2)   r = f1*v1 - f2*v2
#define	ADD2SN(r, v1, f1, f2)   r = f2*r + v1
#define	SUB2SN(r, v1, f1, f2)   r = f2*r - v1
#define	ADD2SS(r, v1, f1, f2)   r = f1*r + f2*v1
#define	SUB2SS(r, v1, f1, f2)   r = f1*r - f2*v1
#define	ADD2SV(r, v1, f1, f2)   r = f1*v1 + f2
#define	SUB2SV(r, v1, f1, f2)   r = f1*v1 - f2
#define	ADD1SV(r, f1, f2)   r = f1*r + f2
#define	SUB1SV(r, f1, f2)   r = f1*r - f2
#define	ADD2VS(r, v1, f1, f2)   r = f1 + f2*v1
#define	SUB2VS(r, v1, f1, f2)   r = f1 - f2*v1
#define	DIV2VS(r, v1, f1, f2)   r = f1 / (f2*v1)
#define	NEG2(r, v1, f1, f2)   r = -v1
#define	NEG1(r, f1, f2)   r = -r
#define	DOT2(r, v1, f1, f2)   f2 += CPLX__ conj(r) * v1
#define	XDOT2(r, v1, f1, f2)   { T y = CPLX__ conj(r) * v1; T t = f2+y; f1 += (t-f2)-y; f2 = t; }
#define	MULT2(r, v1, f1, f2)   f2 += r * v1
#define	XMULT2(r, v1, f1, f2)   { T y = r * v1; T t = f2+y; f1 += (t-f2)-y; f2 = t; }
#define	FABS1(r, f1, f2)   f2 += fabssqr(r)
#define	XFABS1(r, f1, f2)   { double y = fabssqr(r); double t = f2+y; f1 += (t-f2)-y; f2 = t; }
#define	SQR1(r, f1, f2)   f2 += r*r
#define	XSQR1(r, f1, f2)   { T y = r*r; T t = f2+y; f1 += (t-f2)-y; f2 = t; }
#define	SUM1(r, f1, f2)   f2 += r
#define	XSUM1(r, f1, f2)   { T t = f2+r; f1 += (t-f2)-r; f2 = t; }
#define	SUMMULT3(r, v1, v2, f1, f2)   r += v1*v2
#define	SUMCMULT3(r, v1, v2, f1, f2)   r += CPLX__ conj(v1)*v2

Functions
	VKERN_TEMPL_3V (do_vec_vec_add, ADD3)
	vec = vec + vec; More...
	VKERN_TEMPL_3V (do_vec_vec_sub, SUB3)
	vec = vec - vec; More...
	VKERN_TEMPL_3V (do_vec_vec_mul, MUL3)
	vec = emul (vec, vec); More...
	VKERN_TEMPL_3V (do_vec_vec_cmul, CMUL3)
	vec = cemul(vec, vec); More...
	VKERN_TEMPL_3V (do_vec_vec_div, DIV3)
	vec = ediv (vec, vec); More...
	VKERN_TEMPL_3V (do_vec_vec_cdiv, CDIV3)
	vec = cediv(vec, vec); More...
	VKERN_TEMPL_2V (do_vec_add_vec, ADD2)
	vec += vec; More...
	VKERN_TEMPL_2V (do_vec_sub_vec, SUB2)
	vec -= vec; More...
	VKERN_TEMPL_2V (do_vec_sub_vec_inv, SUB2I)
	vec -= vec; vec = -vec; More...
	VKERN_TEMPL_2V (do_vec_mul_vec, MUL2)
	vec *= vec; (emul) More...
	VKERN_TEMPL_2V (do_vec_cmul_vec, CMUL2)
	vec *= conj(vec); (cemul) More...
	VKERN_TEMPL_2V (do_vec_cmul_vec_inv, CMUL2I)
	vec *= conj(vec); (cemul) More...
	VKERN_TEMPL_2V (do_vec_div_vec, DIV2)
	vec /= vec; (ediv) More...
	VKERN_TEMPL_2V (do_vec_div_vec_inv, DIV2I)
	vec /= vec; (ediv) More...
	VKERN_TEMPL_2V (do_vec_cdiv_vec, CDIV2)
	vec /= conj(vec); (cediv) More...
	VKERN_TEMPL_2V (do_vec_cdiv_vec_inv, CDIV2I)
	vec /= conj(vec); (cediv) More...
	VKERN_TEMPL_2V_C (do_vec_val_add, ADD2NV)
	vec = vec + val More...
	VKERN_TEMPL_2V_C (do_vec_val_sub, SUB2NV)
	vec = vec - val More...
	VKERN_TEMPL_2V_C (do_vec_val_mul, MUL2NV)
	vec = vec * val; More...
	VKERN_TEMPL_2V_C (do_val_vec_add, ADD2RV)
	vec = val + vec; More...
	VKERN_TEMPL_2V_C (do_val_vec_sub, SUB2RV)
	vec = val - vec; More...
	VKERN_TEMPL_2V_C (do_val_vec_mul, MUL2RV)
	vec = val * vec; More...
	VKERN_TEMPL_2V_C (do_val_vec_div, DIV2RV)
	vec = val / vec; More...
	VKERN_TEMPL_1V_C (do_vec_add_val, ADD1NV)
	vec += val; More...
	VKERN_TEMPL_1V_C (do_vec_sub_val, SUB1NV)
	vec -= val; More...
	VKERN_TEMPL_1V_C (do_val_sub_vec, SUB1RV)
	vec -= val; vec = -vec; More...
	VKERN_TEMPL_1V_C (do_vec_mul_val, MUL1NV)
	vec *= val; More...
	VKERN_TEMPL_1V_C (do_vec_div_val, DIV1NV)
	vec /= val; More...
	VKERN_TEMPL_1V_C (do_val_div_vec, DIV1RV)
	vec = val / vec; More...
	VKERN_TEMPL_1V_C (do_val_add_vec, ADD1RV)
	vec += val More...
	VKERN_TEMPL_2V_C (do_vec_add_svc, ADD2NS)
	vec += val*vec More...
	VKERN_TEMPL_2V_C (do_vec_sub_svc, SUB2NS)
	vec -= val*vec More...
	VKERN_TEMPL_2V_C (do_vec_sub_svc_inv, SUB2RS)
	vec -= val*vec More...
	VKERN_TEMPL_3V_C (do_vec_svc_add, ADD3NS)
	vec = vec + s*vec; More...
	VKERN_TEMPL_3V_C (do_vec_svc_sub, SUB3NS)
	vec = vec - s*vec; More...
	VKERN_TEMPL_3V_C (do_svc_vec_add, ADD3SN)
	vec = s*vec + vec; More...
	VKERN_TEMPL_3V_C (do_svc_vec_sub, SUB3SN)
	vec = s*vec - vec; More...
	VKERN_TEMPL_3V_CC (do_svc_svc_add, ADD3SS)
	vec = s*vec + s*vec; More...
	VKERN_TEMPL_3V_CC (do_svc_svc_sub, SUB3SS)
	vec = s*vec - s*vec; More...
	VKERN_TEMPL_2V_C (do_svc_add_vec, ADD2SN)
	vec *= s; vec += vec; More...
	VKERN_TEMPL_2V_C (do_svc_sub_vec, SUB2SN)
	vec *= s; vec -= vec; More...
	VKERN_TEMPL_2V_CC (do_svc_add_svc, ADD2SS)
	vec *= s; vec += s*vec; More...
	VKERN_TEMPL_2V_CC (do_svc_sub_svc, SUB2SS)
	vec *= s; vec -= s*vec; More...
	VKERN_TEMPL_2V_CC (do_svc_val_add, ADD2SV)
	vec = s*vec + val More...
	VKERN_TEMPL_2V_CC (do_svc_val_sub, SUB2SV)
	vec = s*vec - val More...
	VKERN_TEMPL_1V_CC (do_svc_add_val, ADD1SV)
	vec = s*vec; vec += val; More...
	VKERN_TEMPL_1V_CC (do_svc_sub_val, SUB1SV)
	vec = s*vec; vec -= val; More...
	VKERN_TEMPL_2V_CC (do_val_svc_add, ADD2VS)
	vec = val + s*vec; More...
	VKERN_TEMPL_2V_CC (do_val_svc_sub, SUB2VS)
	vec = val - s*vec; More...
	VKERN_TEMPL_2V_CC (do_val_svc_div, DIV2VS)
	vec = val / s*vec; More...
	VKERN_TEMPL_2V (do_vec_neg_vec, NEG2)
	vec = -vec More...
	VKERN_TEMPL_1V (do_vec_neg, NEG1)
	vec = -vec More...
	VKERN_TEMPL_2V_T (do_vec_dot_quick, DOT2, T)
	val = SUM vec * ~vec; More...
	VKERN_TEMPL_2V_T (do_vec_dot_exact, XDOT2, T)
	val = SUM vec * ~vec; More...
	VKERN_TEMPL_2V_T (do_vec_mult_quick, MULT2, T)
	val = SUM vec * vec; More...
	VKERN_TEMPL_2V_T_STRIDE (do_vec_mult_stride_quick, MULT2, T)
	VKERN_TEMPL_2V_T (do_vec_mult_exact, XMULT2, T)
	val = SUM vec * vec; More...
	VKERN_TEMPL_2V_T_STRIDE (do_vec_mult_stride_exact, XMULT2, T)
	VKERN_TEMPL_1V_T_LD (do_vec_fabssqr_quick, FABS1, double)
	val = SUM fabssqr vec ; More...
	VKERN_TEMPL_1V_T (do_vec_fabssqr_exact, XFABS1, double)
	val = SUM fabssqr vec ; More...
	VKERN_TEMPL_1V_T (do_vec_sumsqr_quick, SQR1, T)
	val = SUM sqr vec ; More...
	VKERN_TEMPL_1V_T (do_vec_sumsqr_exact, XSQR1, T)
	val = SUM sqr vec ; More...
	VKERN_TEMPL_1V_T (do_vec_sum_quick, SUM1, T)
	val = SUM vec ; More...
	VKERN_TEMPL_1V_T (do_vec_sum_exact, XSUM1, T)
	val = exact SUM vec More...
	VKERN_TEMPL_3V (do_add_vec_vec_mul, SUMMULT3)
	VKERN_TEMPL_3V (do_add_vec_vec_cmul, SUMCMULT3)

Detailed Description

Kernels to be called for array operations Unrolled and with prefetching.

Kurt Garloff kurt@.nosp@m.garl.nosp@m.off.d.nosp@m.e, 07/2002, GNU LGPL

Id:

vec_kern_unr_pref.h,v 1.1.2.39 2019/05/28 11:13:02 garloff Exp

Definition in file vec_kern_unr_pref.h.

Macro Definition Documentation

#define ADD1NV	(		r,
			f1,
			f2
	)		r += f2

Definition at line 114 of file vec_kern_unr_pref.h.

#define ADD1RV	(		r,
			f1,
			f2
	)		r = f2 + r

Definition at line 138 of file vec_kern_unr_pref.h.

#define ADD1SV	(		r,
			f1,
			f2
	)		r = f1*r + f2

Definition at line 215 of file vec_kern_unr_pref.h.

#define ADD2	(		r,
			v1,
			f1,
			f2
	)		r += v1

Definition at line 42 of file vec_kern_unr_pref.h.

#define ADD2NS	(		r,
			v1,
			f1,
			f2
	)		r += f2*v1

Definition at line 148 of file vec_kern_unr_pref.h.

#define ADD2NV	(		r,
			v1,
			f1,
			f2
	)		r = v1 + f2

Definition at line 84 of file vec_kern_unr_pref.h.

#define ADD2RV	(		r,
			v1,
			f1,
			f2
	)		r = f2 + v1

Definition at line 97 of file vec_kern_unr_pref.h.

#define ADD2SN	(		r,
			v1,
			f1,
			f2
	)		r = f2*r + v1

Definition at line 188 of file vec_kern_unr_pref.h.

#define ADD2SS	(		r,
			v1,
			f1,
			f2
	)		r = f1*r + f2*v1

Definition at line 197 of file vec_kern_unr_pref.h.

#define ADD2SV	(		r,
			v1,
			f1,
			f2
	)		r = f1*v1 + f2

Definition at line 206 of file vec_kern_unr_pref.h.

#define ADD2VS	(		r,
			v1,
			f1,
			f2
	)		r = f1 + f2*v1

Definition at line 224 of file vec_kern_unr_pref.h.

#define ADD3	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = v1 + v2

Definition at line 17 of file vec_kern_unr_pref.h.

#define ADD3NS	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = v1 + f2*v2

Definition at line 161 of file vec_kern_unr_pref.h.

#define ADD3SN	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = f2*v1 + v2

Definition at line 170 of file vec_kern_unr_pref.h.

#define ADD3SS	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = f1*v1 + f2*v2

Definition at line 179 of file vec_kern_unr_pref.h.

#define CDIV2	(		r,
			v1,
			f1,
			f2
	)		r = CPLX__ conj(r) / v1

Definition at line 74 of file vec_kern_unr_pref.h.

#define CDIV2I	(		r,
			v1,
			f1,
			f2
	)		r = CPLX__ conj(v1) / r

Definition at line 78 of file vec_kern_unr_pref.h.

#define CDIV3	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = CPLX__ conj(v1) / v2

Definition at line 37 of file vec_kern_unr_pref.h.

#define CMUL2	(		r,
			v1,
			f1,
			f2
	)		r = CPLX__ conj(r) * v1

Definition at line 58 of file vec_kern_unr_pref.h.

#define CMUL2I	(		r,
			v1,
			f1,
			f2
	)		r *= CPLX__ conj(v1)

Definition at line 62 of file vec_kern_unr_pref.h.

#define CMUL3	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = CPLX__ conj(v1) * v2

Definition at line 29 of file vec_kern_unr_pref.h.

#define DIV1NV	(		r,
			f1,
			f2
	)		r /= f2

Definition at line 130 of file vec_kern_unr_pref.h.

#define DIV1RV	(		r,
			f1,
			f2
	)		r = f2 / r

Definition at line 134 of file vec_kern_unr_pref.h.

#define DIV2	(		r,
			v1,
			f1,
			f2
	)		r /= v1

Definition at line 66 of file vec_kern_unr_pref.h.

#define DIV2I	(		r,
			v1,
			f1,
			f2
	)		r = v1 / r

Definition at line 70 of file vec_kern_unr_pref.h.

#define DIV2RV	(		r,
			v1,
			f1,
			f2
	)		r = f2 / v1

Definition at line 109 of file vec_kern_unr_pref.h.

#define DIV2VS	(		r,
			v1,
			f1,
			f2
	)		r = f1 / (f2*v1)

Definition at line 232 of file vec_kern_unr_pref.h.

#define DIV3	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = v1 / v2

Definition at line 33 of file vec_kern_unr_pref.h.

#define DOT2	(		r,
			v1,
			f1,
			f2
	)		f2 += CPLX__ conj(r) * v1

Definition at line 245 of file vec_kern_unr_pref.h.

#define FABS1	(		r,
			f1,
			f2
	)		f2 += fabssqr(r)

Definition at line 263 of file vec_kern_unr_pref.h.

#define MUL1NV	(		r,
			f1,
			f2
	)		r *= f2

Definition at line 126 of file vec_kern_unr_pref.h.

#define MUL1RV	(		r,
			f1,
			f2
	)		r = f2 * r;

Definition at line 143 of file vec_kern_unr_pref.h.

#define MUL2	(		r,
			v1,
			f1,
			f2
	)		r *= v1

Definition at line 54 of file vec_kern_unr_pref.h.

#define MUL2NV	(		r,
			v1,
			f1,
			f2
	)		r = v1 * f2

Definition at line 92 of file vec_kern_unr_pref.h.

#define MUL2RV	(		r,
			v1,
			f1,
			f2
	)		r = f2 * v1

Definition at line 105 of file vec_kern_unr_pref.h.

#define MUL3	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = v1 * v2

Definition at line 25 of file vec_kern_unr_pref.h.

#define MULT2	(		r,
			v1,
			f1,
			f2
	)		f2 += r * v1

Definition at line 253 of file vec_kern_unr_pref.h.

#define NEG1	(		r,
			f1,
			f2
	)		r = -r

Definition at line 241 of file vec_kern_unr_pref.h.

#define NEG2	(		r,
			v1,
			f1,
			f2
	)		r = -v1

Definition at line 237 of file vec_kern_unr_pref.h.

#define SQR1	(		r,
			f1,
			f2
	)		f2 += r*r

Definition at line 271 of file vec_kern_unr_pref.h.

#define SUB1NV	(		r,
			f1,
			f2
	)		r -= f2

Definition at line 118 of file vec_kern_unr_pref.h.

#define SUB1RV	(		r,
			f1,
			f2
	)		r = f2 - r

Definition at line 122 of file vec_kern_unr_pref.h.

#define SUB1SV	(		r,
			f1,
			f2
	)		r = f1*r - f2

Definition at line 219 of file vec_kern_unr_pref.h.

#define SUB2	(		r,
			v1,
			f1,
			f2
	)		r -= v1

Definition at line 46 of file vec_kern_unr_pref.h.

#define SUB2I	(		r,
			v1,
			f1,
			f2
	)		r = v1 - r

Definition at line 50 of file vec_kern_unr_pref.h.

#define SUB2NS	(		r,
			v1,
			f1,
			f2
	)		r -= f2*v1

Definition at line 152 of file vec_kern_unr_pref.h.

#define SUB2NV	(		r,
			v1,
			f1,
			f2
	)		r = v1 - f2

Definition at line 88 of file vec_kern_unr_pref.h.

#define SUB2RS	(		r,
			v1,
			f1,
			f2
	)		r = f2*v1 - r

Definition at line 156 of file vec_kern_unr_pref.h.

#define SUB2RV	(		r,
			v1,
			f1,
			f2
	)		r = f2 - v1

Definition at line 101 of file vec_kern_unr_pref.h.

#define SUB2SN	(		r,
			v1,
			f1,
			f2
	)		r = f2*r - v1

Definition at line 192 of file vec_kern_unr_pref.h.

#define SUB2SS	(		r,
			v1,
			f1,
			f2
	)		r = f1*r - f2*v1

Definition at line 201 of file vec_kern_unr_pref.h.

#define SUB2SV	(		r,
			v1,
			f1,
			f2
	)		r = f1*v1 - f2

Definition at line 210 of file vec_kern_unr_pref.h.

#define SUB2VS	(		r,
			v1,
			f1,
			f2
	)		r = f1 - f2*v1

Definition at line 228 of file vec_kern_unr_pref.h.

#define SUB3	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = v1 - v2

Definition at line 21 of file vec_kern_unr_pref.h.

#define SUB3NS	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = v1 - f2*v2

Definition at line 165 of file vec_kern_unr_pref.h.

#define SUB3SN	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = f2*v1 - v2

Definition at line 174 of file vec_kern_unr_pref.h.

#define SUB3SS	(		r,
			v1,
			v2,
			f1,
			f2
	)		r = f1*v1 - f2*v2

Definition at line 183 of file vec_kern_unr_pref.h.

#define SUM1	(		r,
			f1,
			f2
	)		f2 += r

Definition at line 279 of file vec_kern_unr_pref.h.

#define SUMCMULT3	(		r,
			v1,
			v2,
			f1,
			f2
	)		r += CPLX__ conj(v1)*v2

Definition at line 291 of file vec_kern_unr_pref.h.

#define SUMMULT3	(		r,
			v1,
			v2,
			f1,
			f2
	)		r += v1*v2

Definition at line 289 of file vec_kern_unr_pref.h.

#define XDOT2	(		r,
			v1,
			f1,
			f2
	)		{ T y = CPLX__ conj(r) * v1; T t = f2+y; f1 += (t-f2)-y; f2 = t; }

Definition at line 249 of file vec_kern_unr_pref.h.

#define XFABS1	(		r,
			f1,
			f2
	)		{ double y = fabssqr(r); double t = f2+y; f1 += (t-f2)-y; f2 = t; }

Definition at line 267 of file vec_kern_unr_pref.h.

#define XMULT2	(		r,
			v1,
			f1,
			f2
	)		{ T y = r * v1; T t = f2+y; f1 += (t-f2)-y; f2 = t; }

Definition at line 258 of file vec_kern_unr_pref.h.

#define XSQR1	(		r,
			f1,
			f2
	)		{ T y = r*r; T t = f2+y; f1 += (t-f2)-y; f2 = t; }

Definition at line 275 of file vec_kern_unr_pref.h.

#define XSUM1	(		r,
			f1,
			f2
	)		{ T t = f2+r; f1 += (t-f2)-r; f2 = t; }

Definition at line 283 of file vec_kern_unr_pref.h.

Function Documentation

VKERN_TEMPL_1V	(	do_vec_neg	,
		NEG1
	)

vec = -vec

VKERN_TEMPL_1V_C	(	do_vec_add_val	,
		ADD1NV
	)

vec += val;

VKERN_TEMPL_1V_C	(	do_vec_sub_val	,
		SUB1NV
	)

vec -= val;

VKERN_TEMPL_1V_C	(	do_val_sub_vec	,
		SUB1RV
	)

vec -= val; vec = -vec;

VKERN_TEMPL_1V_C	(	do_vec_mul_val	,
		MUL1NV
	)

vec *= val;

VKERN_TEMPL_1V_C	(	do_vec_div_val	,
		DIV1NV
	)

vec /= val;

VKERN_TEMPL_1V_C	(	do_val_div_vec	,
		DIV1RV
	)

vec = val / vec;

VKERN_TEMPL_1V_C	(	do_val_add_vec	,
		ADD1RV
	)

vec += val

VKERN_TEMPL_1V_CC	(	do_svc_add_val	,
		ADD1SV
	)

vec = s*vec; vec += val;

VKERN_TEMPL_1V_CC	(	do_svc_sub_val	,
		SUB1SV
	)

vec = s*vec; vec -= val;

VKERN_TEMPL_1V_T	(	do_vec_fabssqr_exact	,
		XFABS1	,
		double
	)

val = SUM fabssqr vec ;

VKERN_TEMPL_1V_T	(	do_vec_sumsqr_quick	,
		SQR1	,
		T
	)

val = SUM sqr vec ;

VKERN_TEMPL_1V_T	(	do_vec_sumsqr_exact	,
		XSQR1	,
		T
	)

val = SUM sqr vec ;

VKERN_TEMPL_1V_T	(	do_vec_sum_quick	,
		SUM1	,
		T
	)

val = SUM vec ;

VKERN_TEMPL_1V_T	(	do_vec_sum_exact	,
		XSUM1	,
		T
	)

val = exact SUM vec

VKERN_TEMPL_1V_T_LD	(	do_vec_fabssqr_quick	,
		FABS1	,
		double
	)

val = SUM fabssqr vec ;

VKERN_TEMPL_2V	(	do_vec_add_vec	,
		ADD2
	)

vec += vec;

VKERN_TEMPL_2V	(	do_vec_sub_vec	,
		SUB2
	)

vec -= vec;

VKERN_TEMPL_2V	(	do_vec_sub_vec_inv	,
		SUB2I
	)

vec -= vec; vec = -vec;

VKERN_TEMPL_2V	(	do_vec_mul_vec	,
		MUL2
	)

vec *= vec; (emul)

VKERN_TEMPL_2V	(	do_vec_cmul_vec	,
		CMUL2
	)

vec *= conj(vec); (cemul)

VKERN_TEMPL_2V	(	do_vec_cmul_vec_inv	,
		CMUL2I
	)

vec *= conj(vec); (cemul)

VKERN_TEMPL_2V	(	do_vec_div_vec	,
		DIV2
	)

vec /= vec; (ediv)

VKERN_TEMPL_2V	(	do_vec_div_vec_inv	,
		DIV2I
	)

vec /= vec; (ediv)

VKERN_TEMPL_2V	(	do_vec_cdiv_vec	,
		CDIV2
	)

vec /= conj(vec); (cediv)

VKERN_TEMPL_2V	(	do_vec_cdiv_vec_inv	,
		CDIV2I
	)

vec /= conj(vec); (cediv)

VKERN_TEMPL_2V	(	do_vec_neg_vec	,
		NEG2
	)

vec = -vec

VKERN_TEMPL_2V_C	(	do_vec_val_add	,
		ADD2NV
	)

vec = vec + val

VKERN_TEMPL_2V_C	(	do_vec_val_sub	,
		SUB2NV
	)

vec = vec - val

VKERN_TEMPL_2V_C	(	do_vec_val_mul	,
		MUL2NV
	)

vec = vec * val;

VKERN_TEMPL_2V_C	(	do_val_vec_add	,
		ADD2RV
	)

vec = val + vec;

VKERN_TEMPL_2V_C	(	do_val_vec_sub	,
		SUB2RV
	)

vec = val - vec;

VKERN_TEMPL_2V_C	(	do_val_vec_mul	,
		MUL2RV
	)

vec = val * vec;

VKERN_TEMPL_2V_C	(	do_val_vec_div	,
		DIV2RV
	)

vec = val / vec;

VKERN_TEMPL_2V_C	(	do_vec_add_svc	,
		ADD2NS
	)

vec += val*vec

VKERN_TEMPL_2V_C	(	do_vec_sub_svc	,
		SUB2NS
	)

vec -= val*vec

VKERN_TEMPL_2V_C	(	do_vec_sub_svc_inv	,
		SUB2RS
	)

vec -= val*vec

VKERN_TEMPL_2V_C	(	do_svc_add_vec	,
		ADD2SN
	)

vec *= s; vec += vec;

VKERN_TEMPL_2V_C	(	do_svc_sub_vec	,
		SUB2SN
	)

vec *= s; vec -= vec;

VKERN_TEMPL_2V_CC	(	do_svc_add_svc	,
		ADD2SS
	)

vec *= s; vec += s*vec;

VKERN_TEMPL_2V_CC	(	do_svc_sub_svc	,
		SUB2SS
	)

vec *= s; vec -= s*vec;

VKERN_TEMPL_2V_CC	(	do_svc_val_add	,
		ADD2SV
	)

vec = s*vec + val

VKERN_TEMPL_2V_CC	(	do_svc_val_sub	,
		SUB2SV
	)

vec = s*vec - val

VKERN_TEMPL_2V_CC	(	do_val_svc_add	,
		ADD2VS
	)

vec = val + s*vec;

VKERN_TEMPL_2V_CC	(	do_val_svc_sub	,
		SUB2VS
	)

vec = val - s*vec;

VKERN_TEMPL_2V_CC	(	do_val_svc_div	,
		DIV2VS
	)

vec = val / s*vec;

VKERN_TEMPL_2V_T	(	do_vec_dot_quick	,
		DOT2	,
		T
	)

val = SUM vec * ~vec;

VKERN_TEMPL_2V_T	(	do_vec_dot_exact	,
		XDOT2	,
		T
	)

val = SUM vec * ~vec;

VKERN_TEMPL_2V_T	(	do_vec_mult_quick	,
		MULT2	,
		T
	)

val = SUM vec * vec;

VKERN_TEMPL_2V_T	(	do_vec_mult_exact	,
		XMULT2	,
		T
	)

val = SUM vec * vec;

VKERN_TEMPL_2V_T_STRIDE	(	do_vec_mult_stride_quick	,
		MULT2	,
		T
	)

VKERN_TEMPL_2V_T_STRIDE	(	do_vec_mult_stride_exact	,
		XMULT2	,
		T
	)

VKERN_TEMPL_3V	(	do_vec_vec_add	,
		ADD3
	)

vec = vec + vec;

VKERN_TEMPL_3V	(	do_vec_vec_sub	,
		SUB3
	)

vec = vec - vec;

VKERN_TEMPL_3V	(	do_vec_vec_mul	,
		MUL3
	)

vec = emul (vec, vec);

VKERN_TEMPL_3V	(	do_vec_vec_cmul	,
		CMUL3
	)

vec = cemul(vec, vec);

VKERN_TEMPL_3V	(	do_vec_vec_div	,
		DIV3
	)

vec = ediv (vec, vec);

VKERN_TEMPL_3V	(	do_vec_vec_cdiv	,
		CDIV3
	)

vec = cediv(vec, vec);

VKERN_TEMPL_3V	(	do_add_vec_vec_mul	,
		SUMMULT3
	)

VKERN_TEMPL_3V	(	do_add_vec_vec_cmul	,
		SUMCMULT3
	)

VKERN_TEMPL_3V_C	(	do_vec_svc_add	,
		ADD3NS
	)

vec = vec + s*vec;

VKERN_TEMPL_3V_C	(	do_vec_svc_sub	,
		SUB3NS
	)

vec = vec - s*vec;

VKERN_TEMPL_3V_C	(	do_svc_vec_add	,
		ADD3SN
	)

vec = s*vec + vec;

VKERN_TEMPL_3V_C	(	do_svc_vec_sub	,
		SUB3SN
	)

vec = s*vec - vec;

VKERN_TEMPL_3V_CC	(	do_svc_svc_add	,
		ADD3SS
	)

vec = s*vec + s*vec;

VKERN_TEMPL_3V_CC	(	do_svc_svc_sub	,
		SUB3SS
	)

vec = s*vec - s*vec;