--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/gen/gendouble.c	Wed Apr 30 18:36:38 2025 -0400
@@ -0,0 +1,275 @@
+/**
+ * vec - a tiny SIMD vector library in C99
+ * 
+ * Copyright (c) 2024-2025 Paper
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ * 
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+**/
+
+#include "genlib.h"
+
+#define ARRAY_SIZE(x) (sizeof(x)/sizeof((x)[0]))
+
+/* XXX: would it be faster to unroll literally everything instead of defining everything,
+ * and then unpacking it all? */
+static const char *header =
+	"/**\n"
+	" * vec - a tiny SIMD vector library in C99\n"
+	" * \n"
+	" * Copyright (c) 2024-2025 Paper\n"
+	" * \n"
+	" * Permission is hereby granted, free of charge, to any person obtaining a copy\n"
+	" * of this software and associated documentation files (the \"Software\"), to deal\n"
+	" * in the Software without restriction, including without limitation the rights\n"
+	" * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell\n"
+	" * copies of the Software, and to permit persons to whom the Software is\n"
+	" * furnished to do so, subject to the following conditions:\n"
+	" * \n"
+	" * The above copyright notice and this permission notice shall be included in all\n"
+	" * copies or substantial portions of the Software.\n"
+	" * \n"
+	" * THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n"
+	" * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n"
+	" * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\n"
+	" * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n"
+	" * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\n"
+	" * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\n"
+	" * SOFTWARE.\n"
+	"**/\n"
+	"\n"
+	"/* This file is automatically generated! Do not edit it directly!\n"
+	" * Edit the code that generates it in utils/gendbl.c  --paper */\n"
+	"\n"
+	"#ifndef VEC_IMPL_DOUBLE_H_\n"
+	"#define VEC_IMPL_DOUBLE_H_\n"
+	"\n"
+	"#define VEC_DOUBLE_SPLAT(sign, bits, size, halfsize) \\\n"
+	"	VEC_FUNC_IMPL v##sign##int##bits##x##size v##sign##int##bits##x##size##_splat(vec_##sign##int##bits x) \\\n"
+	"	{ \\\n"
+	"		v##sign##int##bits##x##size vec; \\\n"
+	"	\\\n"
+	"		vec.dbl[0] = v##sign##int##bits##x##halfsize##_splat(x); \\\n"
+	"		vec.dbl[1] = v##sign##int##bits##x##halfsize##_splat(x); \\\n"
+	"	\\\n"
+	"		return vec; \\\n"
+	"	}\n"
+	"\n"
+	"#define VEC_DOUBLE_LOAD_EX(name, sign, bits, size, halfsize) \\\n"
+	"	VEC_FUNC_IMPL v##sign##int##bits##x##size v##sign##int##bits##x##size##_##name(const vec_##sign##int##bits x[size]) \\\n"
+	"	{ \\\n"
+	"		v##sign##int##bits##x##size vec; \\\n"
+	"	\\\n"
+	"		vec.dbl[0] = v##sign##int##bits##x##halfsize##_##name(x); \\\n"
+	"		vec.dbl[1] = v##sign##int##bits##x##halfsize##_##name(x + halfsize); \\\n"
+	"	\\\n"
+	"		return vec; \\\n"
+	"	}\n"
+	"\n"
+	"#define VEC_DOUBLE_LOAD(sign, bits, size, halfsize) VEC_DOUBLE_LOAD_EX(load, sign, bits, size, halfsize)\n"
+	"#define VEC_DOUBLE_LOAD_ALIGNED(sign, bits, size, halfsize) VEC_DOUBLE_LOAD_EX(load_aligned, sign, bits, size, halfsize)\n"
+	"\n"
+	"#define VEC_DOUBLE_STORE_EX(name, sign, bits, size, halfsize) \\\n"
+	"	VEC_FUNC_IMPL void v##sign##int##bits##x##size##_##name(v##sign##int##bits##x##size vec, vec_##sign##int##bits x[size]) \\\n"
+	"	{ \\\n"
+	"		v##sign##int##bits##x##halfsize##_##name(vec.dbl[0], x); \\\n"
+	"		v##sign##int##bits##x##halfsize##_##name(vec.dbl[1], x + halfsize); \\\n"
+	"	}\n"
+	"\n"
+	"#define VEC_DOUBLE_STORE(sign, bits, size, halfsize) VEC_DOUBLE_STORE_EX(store, sign, bits, size, halfsize)\n"
+	"#define VEC_DOUBLE_STORE_ALIGNED(sign, bits, size, halfsize) VEC_DOUBLE_STORE_EX(store_aligned, sign, bits, size, halfsize)\n"
+	"\n"
+	"#define VEC_DOUBLE_OP(name, sign, bits, size, halfsize, secondsign) \\\n"
+	"	VEC_FUNC_IMPL v##sign##int##bits##x##size v##sign##int##bits##x##size##_##name(v##sign##int##bits##x##size vec1, v##secondsign##int##bits##x##size vec2) \\\n"
+	"	{ \\\n"
+	"		vec1.dbl[0] = v##sign##int##bits##x##halfsize##_##name(vec1.dbl[0], vec2.dbl[0]); \\\n"
+	"		vec1.dbl[1] = v##sign##int##bits##x##halfsize##_##name(vec1.dbl[1], vec2.dbl[1]); \\\n"
+	"	\\\n"
+	"		return vec1; \\\n"
+	"	}\n"
+	"\n"
+	"#define VEC_DOUBLE_ADD(sign, bits, size, halfsize) VEC_DOUBLE_OP(add, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_SUB(sign, bits, size, halfsize) VEC_DOUBLE_OP(sub, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_MUL(sign, bits, size, halfsize) VEC_DOUBLE_OP(mul, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_DIV(sign, bits, size, halfsize) VEC_DOUBLE_OP(div, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_MOD(sign, bits, size, halfsize) VEC_DOUBLE_OP(mod, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_AVG(sign, bits, size, halfsize) VEC_DOUBLE_OP(avg, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_LSHIFT(sign, bits, size, halfsize) VEC_DOUBLE_OP(lshift, sign, bits, size, halfsize, u)\n"
+	"#define VEC_DOUBLE_RSHIFT(sign, bits, size, halfsize) VEC_DOUBLE_OP(rshift, sign, bits, size, halfsize, u)\n"
+	"#define VEC_DOUBLE_LRSHIFT(sign, bits, size, halfsize) VEC_DOUBLE_OP(lrshift, sign, bits, size, halfsize, u)\n"
+	"#define VEC_DOUBLE_AND(sign, bits, size, halfsize) VEC_DOUBLE_OP(and, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_OR(sign, bits, size, halfsize) VEC_DOUBLE_OP(or, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_XOR(sign, bits, size, halfsize) VEC_DOUBLE_OP(xor, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_MIN(sign, bits, size, halfsize) VEC_DOUBLE_OP(min, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_MAX(sign, bits, size, halfsize) VEC_DOUBLE_OP(max, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_CMPLT(sign, bits, size, halfsize) VEC_DOUBLE_OP(cmplt, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_CMPLE(sign, bits, size, halfsize) VEC_DOUBLE_OP(cmple, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_CMPEQ(sign, bits, size, halfsize) VEC_DOUBLE_OP(cmpeq, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_CMPGE(sign, bits, size, halfsize) VEC_DOUBLE_OP(cmpge, sign, bits, size, halfsize, sign)\n"
+	"#define VEC_DOUBLE_CMPGT(sign, bits, size, halfsize) VEC_DOUBLE_OP(cmpgt, sign, bits, size, halfsize, sign)\n"
+	"\n"
+	"#define VEC_DOUBLE_NOT(sign, bits, size, halfsize) \\\n"
+	"	VEC_FUNC_IMPL v##sign##int##bits##x##size v##sign##int##bits##x##size##_not(v##sign##int##bits##x##size vec) \\\n"
+	"	{ \\\n"
+	"		vec.dbl[0] = v##sign##int##bits##x##halfsize##_not(vec.dbl[0]); \\\n"
+	"		vec.dbl[1] = v##sign##int##bits##x##halfsize##_not(vec.dbl[1]); \\\n"
+	"	\\\n"
+	"		return vec; \\\n"
+	"	}\n"
+	"\n"
+	"#endif /* VEC_IMPL_DOUBLE_H_ */ \n"
+	"\n"
+	"/* ------------------------------------------------------------------------ */\n"
+	"/* PREPROCESSOR HELL INCOMING */\n"
+	"";
+
+static const char *footer = 
+	"" /* nothing */;
+
+/* ------------------------------------------------------------------------ */
+
+static void op_print_pp_halfsize(int op, int type, int bits, int size)
+{
+	struct op_info *op_info = gen_op_info(op);
+
+	printf("defined(V%s%dx%d_%s_DEFINED)", type_str[type].u, bits, size / 2, op_info->u);
+}
+
+static void op_print_twoop(int op, int type, int bits, int size)
+{
+	struct op_info *op_info = gen_op_info(op);
+
+	printf("\tvec1.dbl[0] = ");
+	gen_print_vtype(type, bits, size / 2);
+	printf("_%s(vec1.dbl[0], vec2.dbl[0]);\n", op_info->l);
+
+	printf("\tvec1.dbl[1] = ");
+	gen_print_vtype(type, bits, size / 2);
+	printf("_%s(vec1.dbl[1], vec2.dbl[1]);\n", op_info->l);
+
+	printf("\treturn vec1;\n");
+}
+
+static void op_print_unoop(int op, int type, int bits, int size)
+{
+	struct op_info *op_info = gen_op_info(op);
+
+	printf("\tvec.dbl[0] = ");
+	gen_print_vtype(type, bits, size / 2);
+	printf("_%s(vec.dbl[0]);\n", op_info->l);
+
+	printf("\tvec1.dbl[1] = ");
+	gen_print_vtype(type, bits, size / 2);
+	printf("_%s(vec.dbl[1]);\n", op_info->l);
+
+	printf("\treturn vec;\n");
+}
+
+static inline void op_print_load(int op, int type, int bits, int size)
+{
+	struct op_info *op_info = gen_op_info(op);
+
+	printf("\t");
+	gen_print_vtype(type, bits, size);
+	printf(" vec;\n");
+
+	printf("\tvec.dbl[0] = ");
+	gen_print_vtype(type, bits, size / 2);
+	printf("_%s(x);\n", op_info->l);
+
+	printf("\tvec.dbl[1] = ");
+	gen_print_vtype(type, bits, size / 2);
+	printf("_%s(x + %d);\n", op_info->l, size / 2);
+
+	printf("\treturn vec;\n");
+}
+
+static inline void op_print_splat(int op, int type, int bits, int size)
+{
+	struct op_info *op_info = gen_op_info(op);
+
+	printf("\t");
+	gen_print_vtype(type, bits, size);
+	printf(" vec;\n");
+
+	printf("\tvec.dbl[0] = ");
+	gen_print_vtype(type, bits, size / 2);
+	printf("_%s(x);\n", op_info->l);
+
+	printf("\tvec.dbl[1] = ");
+	gen_print_vtype(type, bits, size / 2);
+	printf("_%s(x);\n", op_info->l);
+
+	printf("\treturn vec;\n");
+}
+
+static inline void op_print_store(int op, int type, int bits, int size)
+{
+	struct op_info *op_info = gen_op_info(op);
+
+	printf("\t");
+	gen_print_vtype(type, bits, size / 2);
+	printf("_%s(vec.dbl[0], x);\n", op_info->l);
+
+	printf("\t");
+	gen_print_vtype(type, bits, size / 2);
+	printf("_%s(vec.dbl[1], x + %d);\n", op_info->l, size / 2);
+}
+
+static struct op_impl op_impl[OP_FINAL_] = {
+	[OP_SPLAT] = {NULL, op_print_pp_halfsize, op_print_splat},
+	[OP_LOAD_ALIGNED] = {NULL, op_print_pp_halfsize, op_print_load},
+	[OP_LOAD] = {NULL, op_print_pp_halfsize, op_print_load},
+	[OP_STORE_ALIGNED] = {NULL, op_print_pp_halfsize, op_print_store},
+	[OP_STORE] = {NULL, op_print_pp_halfsize, op_print_store},
+
+	/* arithmetic */
+	[OP_ADD] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_SUB] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_MUL] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_DIV] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_MOD] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_AVG] = {NULL, op_print_pp_halfsize, op_print_twoop},
+
+	/* bitwise */
+	[OP_AND] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_OR] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_XOR] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_NOT] = {NULL, op_print_pp_halfsize, op_print_unoop},
+
+	/* min/max */
+	[OP_MIN] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_MAX] = {NULL, op_print_pp_halfsize, op_print_twoop},
+
+	/* bitshift */
+	[OP_LSHIFT] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_LRSHIFT] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_RSHIFT] = {NULL, op_print_pp_halfsize, op_print_twoop},
+
+	/* comparison */
+	[OP_CMPLT] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_CMPLE] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_CMPEQ] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_CMPGE] = {NULL, op_print_pp_halfsize, op_print_twoop},
+	[OP_CMPGT] = {NULL, op_print_pp_halfsize, op_print_twoop},
+};
+
+int main(void)
+{
+	gen(op_impl, "double");
+}