From e6d02a085533134ee6bf1bca862b03968a83b591 Mon Sep 17 00:00:00 2001
From: Thomas Voss <mail@thomasvoss.com>
Date: Sun, 25 Aug 2024 00:24:53 +0200
Subject: Rename isascii-avx to simd-isascii

---
 c/simd-isascii/.gitignore |  1 +
 c/simd-isascii/Makefile   |  8 ++++
 c/simd-isascii/isascii.c  | 99 +++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 108 insertions(+)
 create mode 100644 c/simd-isascii/.gitignore
 create mode 100644 c/simd-isascii/Makefile
 create mode 100644 c/simd-isascii/isascii.c

(limited to 'c/simd-isascii')

diff --git a/c/simd-isascii/.gitignore b/c/simd-isascii/.gitignore
new file mode 100644
index 0000000..4328e16
--- /dev/null
+++ b/c/simd-isascii/.gitignore
@@ -0,0 +1 @@
+isascii
diff --git a/c/simd-isascii/Makefile b/c/simd-isascii/Makefile
new file mode 100644
index 0000000..3c49e43
--- /dev/null
+++ b/c/simd-isascii/Makefile
@@ -0,0 +1,8 @@
+include ../base.mk
+
+all: isascii
+isascii: isascii.c
+	$(CC) $(CFLAGS) -lbsd -o $@ $<
+
+clean:
+	rm -f isascii
diff --git a/c/simd-isascii/isascii.c b/c/simd-isascii/isascii.c
new file mode 100644
index 0000000..612aa49
--- /dev/null
+++ b/c/simd-isascii/isascii.c
@@ -0,0 +1,99 @@
+#include <sys/mman.h>
+#include <sys/stat.h>
+
+#include <err.h>
+#include <fcntl.h>
+#include <immintrin.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <time.h>
+#include <unistd.h>
+
+#include <bsd/stdlib.h>
+
+#define VECWDTH (512 / 8)
+
+static const unsigned char *readfile(const char *, size_t *);
+
+static const uint8_t charmsk[64] = {
+	0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
+	0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
+	0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
+	0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
+	0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
+	0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
+	0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
+	0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80,
+};
+
+bool
+strisascii(const unsigned char *s, size_t n)
+{
+	__m512i msk = _mm512_loadu_epi8(charmsk);
+	while (n >= VECWDTH) {
+		if (_mm512_test_epi8_mask(_mm512_loadu_epi8(s), msk) != 0)
+			return false;
+		s += VECWDTH;
+		n -= VECWDTH;
+	}
+	for (size_t i = 0; i < n; i++) {
+		if (s[i] > 0x7F)
+			return false;
+	}
+	return true;
+}
+
+bool
+strisascii_dumb(const unsigned char *s, size_t n)
+{
+	for (size_t i = 0; i < n; i++) {
+		if (s[i] > 0x7F)
+			return false;
+	}
+	return true;
+}
+
+int
+main(int argc, char **argv)
+{
+	if (argc != 2) {
+		fprintf(stderr, "Usage: %s file\n", getprogname());
+		exit(EXIT_FAILURE);
+	}
+
+	size_t len;
+	int rv = EXIT_SUCCESS;
+	const unsigned char *beg = readfile(argv[1], &len);
+
+	clock_t tmbeg = clock();
+	if (!strisascii(beg, len))
+		puts("Non-ASCII");
+	printf("Elapsed time: %.3fs\n", (double)(clock() - tmbeg) / CLOCKS_PER_SEC);
+
+	tmbeg = clock();
+	if (!strisascii_dumb((const unsigned char *)beg, len))
+		puts("Non-ASCII");
+	printf("Elapsed time: %.3fs\n", (double)(clock() - tmbeg) / CLOCKS_PER_SEC);
+
+	munmap((void *)beg, len);
+	return rv;
+}
+
+const unsigned char *
+readfile(const char *filename, size_t *n)
+{
+	int fd = open(filename, O_RDONLY);
+	if (fd == -1)
+		err(1, "open: %s", filename);
+
+	struct stat sb;
+	if (fstat(fd, &sb) == -1)
+		err(1, "fstat: %s", filename);
+
+	*n = sb.st_size;
+	const unsigned char *p = mmap(NULL, *n, PROT_READ, MAP_PRIVATE, fd, 0);
+	if (p == MAP_FAILED)
+		err(1, "mmap: %s", filename);
+	return p;
+}
-- 
cgit v1.2.3