adding initial repo commit

Author: andrewCodeDev

.gitignore

@@ -0,0 +1,3 @@
+zig-out/
+zig-cache/
+.zig-cache/

build.zig

@@ -0,0 +1,30 @@
+const std = @import("std");
+
+pub fn build(b: *std.Build) void {
+    const target = b.standardTargetOptions(.{});
+    const optimize = b.standardOptimizeOption(.{
+        .preferred_optimize_mode = .ReleaseFast
+    });
+
+    const exe = b.addExecutable(.{
+        .name = "bench",
+        .root_source_file = b.path("src/main.zig"),
+        .target = target,
+        .optimize = optimize,
+    });
+
+    exe.addCSourceFile(.{
+        .file = b.path("src/hw_counter.c"),
+        .flags = &.{ "-O3" }
+    });
+
+    exe.linkLibC();
+
+    b.installArtifact(exe);
+
+    const run_cmd = b.addRunArtifact(exe);
+
+    const run_step = b.step("run", "run stats test");
+
+    run_step.dependOn(&run_cmd.step);
+}

build.zig.zon

@@ -0,0 +1,72 @@
+.{
+    // This is the default name used by packages depending on this one. For
+    // example, when a user runs `zig fetch --save <url>`, this field is used
+    // as the key in the `dependencies` table. Although the user can choose a
+    // different name, most users will stick with this provided value.
+    //
+    // It is redundant to include "zig" in this name because it is already
+    // within the Zig package namespace.
+    .name = "Stats",
+
+    // This is a [Semantic Version](https://semver.org/).
+    // In a future version of Zig it will be used for package deduplication.
+    .version = "0.0.0",
+
+    // This field is optional.
+    // This is currently advisory only; Zig does not yet do anything
+    // with this value.
+    //.minimum_zig_version = "0.11.0",
+
+    // This field is optional.
+    // Each dependency must either provide a `url` and `hash`, or a `path`.
+    // `zig build --fetch` can be used to fetch all dependencies of a package, recursively.
+    // Once all dependencies are fetched, `zig build` no longer requires
+    // internet connectivity.
+    .dependencies = .{
+        // See `zig fetch --save <url>` for a command-line interface for adding dependencies.
+        //.example = .{
+        //    // When updating this field to a new URL, be sure to delete the corresponding
+        //    // `hash`, otherwise you are communicating that you expect to find the old hash at
+        //    // the new URL.
+        //    .url = "https://example.com/foo.tar.gz",
+        //
+        //    // This is computed from the file contents of the directory of files that is
+        //    // obtained after fetching `url` and applying the inclusion rules given by
+        //    // `paths`.
+        //    //
+        //    // This field is the source of truth; packages do not come from a `url`; they
+        //    // come from a `hash`. `url` is just one of many possible mirrors for how to
+        //    // obtain a package matching this `hash`.
+        //    //
+        //    // Uses the [multihash](https://multiformats.io/multihash/) format.
+        //    .hash = "...",
+        //
+        //    // When this is provided, the package is found in a directory relative to the
+        //    // build root. In this case the package's hash is irrelevant and therefore not
+        //    // computed. This field and `url` are mutually exclusive.
+        //    .path = "foo",
+
+        //    // When this is set to `true`, a package is declared to be lazily
+        //    // fetched. This makes the dependency only get fetched if it is
+        //    // actually used.
+        //    .lazy = false,
+        //},
+    },
+
+    // Specifies the set of files and directories that are included in this package.
+    // Only files and directories listed here are included in the `hash` that
+    // is computed for this package. Only files listed here will remain on disk
+    // when using the zig package manager. As a rule of thumb, one should list
+    // files required for compilation plus any license(s).
+    // Paths are relative to the build root. Use the empty string (`""`) to refer to
+    // the build root itself.
+    // A directory listed here means that all files within, recursively, are included.
+    .paths = .{
+        "build.zig",
+        "build.zig.zon",
+        "src",
+        // For example...
+        //"LICENSE",
+        //"README.md",
+    },
+}

results/test.csv

@@ -0,0 +1,3 @@
+name,clock_mean,clock_median,clock_variance,clock_stddev,clock_curtosis,clock_skew,instructions,branches,branch_misses,cache_misses
+indexOfScalar,156.85800,150.00000,518.64542,22.77379,2.53409,0.90341,2027.91800,356.55000,25.65300,5.69500
+indexOfPosLinear,4747.56000,2945.00000,13095785.38326,3618.80994,2.72652,1.49433,81907.91100,18953.29100,24.29100,3.51400

src/CODE_DUMP

@@ -0,0 +1,40 @@
+    const data = try std.heap.page_allocator.alloc(f64, 1_000);
+        defer std.heap.page_allocator.free(data);
+
+    var timer = std.time.Timer.start() catch unreachable;
+    
+    //// start by calculating our noise floor so we can adjust other
+    //// samples to account for any overhead accrued by measuring
+    noise_floor = calculateNoiseFloor(data);
+    std.debug.print("noise floor: {}\n", .{ noise_floor });
+        
+    for (data[0..]) |*sample| {
+
+        const needle = rand.int(u8);
+
+        timer.reset();
+
+        forceCall(foo, .{ haystack, needle });
+
+        // TODO: same here - subtract from overall times
+        sample.* = @as(f64, @floatFromInt(timer.read())) - noise_floor;
+    }
+
+    const foo_stats = BaseStats.init("foo", data);
+
+    for (data[0..]) |*sample| {
+
+        const needle = rand.int(u8);
+
+        timer.reset();
+
+        forceCall(bar, .{ haystack, needle });
+
+        // TODO: same here - subtract from overall times
+        sample.* = @as(f64, @floatFromInt(timer.read())) - noise_floor;
+    }
+
+    const bar_stats = BaseStats.init("bar", data);
+
+    std.debug.print("{}", .{ foo_stats });
+    std.debug.print("{}", .{ bar_stats });

src/hw_counter.c

@@ -0,0 +1,87 @@
+
+#include <asm/unistd.h>
+#include <linux/perf_event.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/ioctl.h>
+#include <unistd.h>
+
+#include <inttypes.h>
+#include <sys/types.h>
+
+#include "hw_counter.h"
+
+static long perf_event_open(
+    struct perf_event_attr *hw_event,
+    pid_t pid,
+    int cpu, 
+    int group_fd, 
+    unsigned long flags
+) {
+    return syscall(__NR_perf_event_open, hw_event, pid, cpu, group_fd, flags);
+}
+
+long counter_start(unsigned long long COUNT_TYPE)
+{
+    struct perf_event_attr pe;
+    memset(&pe, 0, sizeof(struct perf_event_attr));
+
+    pe.type = PERF_TYPE_HARDWARE;
+    pe.size = sizeof(struct perf_event_attr);
+    pe.config = COUNT_TYPE;
+    pe.disabled = 1;
+    pe.exclude_kernel = 1;
+    pe.exclude_hv = 1;
+
+    long fd = perf_event_open(&pe, 0, -1, -1, 0);
+
+    if (fd == -1) {
+        fprintf(stderr, "Error opening leader %llx\n", pe.config);
+        exit(EXIT_FAILURE);
+    }
+
+    ioctl(fd, PERF_EVENT_IOC_RESET, 0);
+    ioctl(fd, PERF_EVENT_IOC_ENABLE, 0);
+
+    return fd;
+}
+
+long long count_read(long fd) {
+    long long count;
+    ioctl(fd, PERF_EVENT_IOC_DISABLE, 0);
+    read(fd, &count, sizeof(long long));
+    close(fd);
+    return count;
+}
+
+long instruction_start() {
+    return counter_start(PERF_COUNT_HW_INSTRUCTIONS);
+}
+long branch_start() {
+    return counter_start(PERF_COUNT_HW_BRANCH_INSTRUCTIONS);
+}
+long branch_miss_start() {
+    return counter_start(PERF_COUNT_HW_BRANCH_MISSES);
+}
+long cache_miss_start() {
+    return counter_start(PERF_COUNT_HW_CACHE_MISSES);
+}
+
+Descriptors all_start() {
+    Descriptors desc;
+    desc.fds[0] = instruction_start();
+    desc.fds[1] = branch_start();
+    desc.fds[2] = branch_miss_start();
+    desc.fds[3] = cache_miss_start();
+    return desc;
+}
+
+Counts all_read(Descriptors desc) {
+    Counts counts;
+    counts.data[0] = count_read(desc.fds[0]);
+    counts.data[1] = count_read(desc.fds[1]);
+    counts.data[2] = count_read(desc.fds[2]);
+    counts.data[3] = count_read(desc.fds[3]);
+    return counts;
+}

src/hw_counter.h

@@ -0,0 +1,25 @@
+typedef struct {
+    long fds[4];
+} Descriptors;
+
+typedef struct {
+    long long data[4];
+} Counts;
+
+// These functions return a file-descriptor that
+// can be passed to the count_read function
+long instruction_start();
+long branch_start();
+long branch_miss_start();
+long cache_miss_start();
+
+// expects a file descriptor that was created by
+// one of the "foo_start" functions above
+long long count_read(long);
+
+// open each metric and return an array of descriptors
+Descriptors all_start();
+
+// read from descriptors returned by all_start
+Counts all_read(Descriptors);
+

src/main.zig

@@ -0,0 +1,93 @@
+const std = @import("std");
+
+const stats = @import("stats.zig");
+
+pub fn main() !void {
+
+    ///////////////////////////////////////////////
+    
+    var prng = std.Random.DefaultPrng.init(42);
+    const rand = prng.random();
+
+    var gpa = std.heap.GeneralPurposeAllocator(.{}){};
+    const allocator = gpa.allocator();
+
+    // default timer - call reset before each measurement
+    var timer = std.time.Timer.start() catch @panic("Clock not found.");
+
+    // create csv for testing purposes
+    var csv = try std.fs.cwd().createFile("results/test.csv", .{ });
+    const writer = csv.writer();
+
+    // add the header row to the csv
+    try writer.writeAll(stats.CSV_COLUMN_NAMES);
+
+    /////////////////////////////////////////////
+
+    var data = try stats.DataBlock.init(allocator, 1000);
+        defer data.deinit(allocator);
+
+    const haystack = try allocator.alloc(u8, 100_000);
+        defer allocator.free(haystack);
+
+    rand.bytes(haystack);
+
+    ////////////////////////////////////////////
+
+    data.reset();
+
+    for (0..data.samples()) |_| {
+
+        const needle = rand.int(u8);
+
+        const desc = stats.hw.all_start();
+
+        timer.reset();
+    
+        stats.forceCall(foo, .{ haystack, needle });
+
+        data.append(timer.read(), desc);
+    }
+
+    const foo_stats = data.stats("indexOfScalar");
+
+    ////////////////////////////////////////////
+
+    data.reset();
+
+    for (0..data.samples()) |_| {
+
+        const needle = rand.int(u8);
+
+        const desc = stats.hw.all_start();
+
+        timer.reset();
+    
+        stats.forceCall(bar, .{ haystack, needle });
+
+        data.append(timer.read(), desc);
+    }
+
+    const bar_stats = data.stats("indexOfPosLinear");
+
+    ////////////////////////////////////////////
+
+    // append stats as a new row in the csv
+    try foo_stats.write(writer);
+    try bar_stats.write(writer);
+
+    std.debug.print("{}", .{ foo_stats });
+    std.debug.print("{}", .{ bar_stats });
+
+    ////////////////////////////////////////////
+}
+
+// helpers to normalize the interface - not necessary
+pub fn foo(haystack: []const u8, needle: u8) ?usize {
+    return std.mem.indexOfScalar(u8, haystack, needle);
+}
+
+pub fn bar(haystack: []const u8, needle: u8) ?usize {
+    return std.mem.indexOfPosLinear(u8, haystack, 0, &.{ needle });
+}
+

src/root.zig

@@ -0,0 +1,10 @@
+const std = @import("std");
+const testing = std.testing;
+
+export fn add(a: i32, b: i32) i32 {
+    return a + b;
+}
+
+test "basic add functionality" {
+    try testing.expect(add(3, 7) == 10);
+}