cache system file deduplication

[andrewrk]

closes #16149

Contains some bonus enhancements to array hash maps.

Performance

Good news, this appears to make cache hits significantly faster.

Data point: cache hit building hello world with static musl libc

Benchmark 1 (61 runs): master/zig build-exe hello.c -target native-native-musl -lc
  measurement          mean ± σ            min … max           outliers         delta
  wall_time          81.4ms ± 1.76ms    77.7ms … 87.1ms          1 ( 2%)        0%
  peak_rss           64.6MB ± 77.7KB    64.4MB … 64.7MB          0 ( 0%)        0%
  cpu_cycles         97.2M  ± 1.04M     95.1M  …  101M           1 ( 2%)        0%
  instructions        153M  ± 11.1K      152M  …  153M           0 ( 0%)        0%
  cache_references   2.21M  ± 97.1K     2.05M  … 2.54M           2 ( 3%)        0%
  cache_misses        529K  ± 24.4K      486K  …  600K           4 ( 7%)        0%
  branch_misses       409K  ± 6.45K      397K  …  437K           1 ( 2%)        0%
Benchmark 2 (189 runs): cache-dedup/zig build-exe hello.c -target native-native-musl -lc
  measurement          mean ± σ            min … max           outliers         delta
  wall_time          25.8ms ± 1.26ms    23.9ms … 30.7ms         11 ( 6%)        ⚡- 68.4% ±  0.5%
  peak_rss           65.2MB ± 61.8KB    65.1MB … 65.4MB          2 ( 1%)        💩+  1.0% ±  0.0%
  cpu_cycles         41.2M  ±  608K     40.1M  … 46.3M           4 ( 2%)        ⚡- 57.6% ±  0.2%
  instructions       64.3M  ± 12.6K     64.3M  … 64.4M           2 ( 1%)        ⚡- 57.8% ±  0.0%
  cache_references   1.28M  ± 34.5K     1.21M  … 1.35M           0 ( 0%)        ⚡- 41.9% ±  0.7%
  cache_misses        348K  ± 18.6K      297K  …  396K           0 ( 0%)        ⚡- 34.2% ±  1.1%
  branch_misses       199K  ± 1.34K      197K  …  206K           6 ( 3%)        ⚡- 51.2% ±  0.2%

[squeek502]

Hitting a segfault on Windows after this, will try to investigate more and file an issue (EDIT: Issue filed here: #19408), but here's the reproduction:

> zig init

> zig build

then modify anything in build.zig (delete a comment, etc), then:

> zig build
thread 19376 panic: Segmentation fault at address 0xffffffffffffffff
C:\Users\Ryan\Programming\Zig\zig\lib\std\Build\Cache.zig:65:46: 0xeeb63c in hash (zig.exe.obj)
        return @truncate(std.hash.Wyhash.hash(pp.prefix, pp.sub_path));
                                             ^
C:\Users\Ryan\Programming\Zig\zig\lib\std\Build\Cache.zig:323:43: 0xa0117b in hash (zig.exe.obj)
            return file.prefixed_path.hash();
                                          ^
C:\Users\Ryan\Programming\Zig\zig\lib\std\array_hash_map.zig:1849:28: 0x6f66ee in checkedHash__anon_79003 (zig.exe.obj)
            return ctx.hash(key);
                           ^
C:\Users\Ryan\Programming\Zig\zig\lib\std\array_hash_map.zig:1634:87: 0x1391d4b in getSlotByIndex__anon_115743 (zig.exe.obj)
            const h = if (store_hash) slice.items(.hash)[entry_index] else checkedHash(ctx, slice.items(.key)[entry_index]);
                                                                                      ^
C:\Users\Ryan\Programming\Zig\zig\lib\std\array_hash_map.zig:1599:45: 0xeebc8f in removeFromIndexByIndexGeneric__anon_105092 (zig.exe.obj)
            const slot = self.getSlotByIndex(entry_index, ctx, header, I, indexes);
                                            ^
C:\Users\Ryan\Programming\Zig\zig\lib\std\array_hash_map.zig:1593:58: 0xa017c1 in removeFromIndexByIndex (zig.exe.obj)
                .u8 => self.removeFromIndexByIndexGeneric(entry_index, ctx, header, u8, header.indexes(u8)),
                                                         ^
C:\Users\Ryan\Programming\Zig\zig\lib\std\array_hash_map.zig:1379:48: 0x6f8707 in shrinkRetainingCapacityContext (zig.exe.obj)
                    self.removeFromIndexByIndex(i, if (store_hash) {} else ctx, header);
                                               ^
C:\Users\Ryan\Programming\Zig\zig\lib\std\array_hash_map.zig:1367:55: 0x428a2f in shrinkRetainingCapacity (zig.exe.obj)
            return self.shrinkRetainingCapacityContext(new_len, undefined);
                                                      ^
C:\Users\Ryan\Programming\Zig\zig\lib\std\Build\Cache.zig:639:43: 0x2511c2 in unhit (zig.exe.obj)
        self.files.shrinkRetainingCapacity(input_file_count);
                                          ^
C:\Users\Ryan\Programming\Zig\zig\lib\std\Build\Cache.zig:605:27: 0x24f935 in hit (zig.exe.obj)
                self.unhit(bin_digest, input_file_count);
                          ^
C:\Users\Ryan\Programming\Zig\zig\src\Compilation.zig:2003:35: 0x2b09ec in update (zig.exe.obj)
            const is_hit = man.hit() catch |err| {
                                  ^
C:\Users\Ryan\Programming\Zig\zig\src\main.zig:4507:24: 0x2e3212 in updateModule (zig.exe.obj)
        try comp.update(main_progress_node);
                       ^
C:\Users\Ryan\Programming\Zig\zig\src\main.zig:5250:25: 0x36b428 in cmdBuild (zig.exe.obj)
            updateModule(comp, color) catch |err| switch (err) {
                        ^
C:\Users\Ryan\Programming\Zig\zig\src\main.zig:276:24: 0x18433b in mainArgs (zig.exe.obj)
        return cmdBuild(gpa, arena, cmd_args);
                       ^
C:\Users\Ryan\Programming\Zig\zig\src\main.zig:206:20: 0x18151e in main (zig.exe.obj)
    return mainArgs(gpa, arena, args);
                   ^
C:\Users\Ryan\Programming\Zig\zig\lib\std\start.zig:484:5: 0x18124a in main (zig.exe.obj)
    return callMainWithArgs(@as(usize, @intCast(c_argc)), @as([*][*:0]u8, @ptrCast(c_argv)), envp);
    ^
C:\Users\Ryan\Programming\Zig\zig\lib\libc\mingw\crt\crtexe.c:267:0: 0x294b7c4 in __tmainCRTStartup (crt2.obj)
    mainret = _tmain (argc, argv, envp);

C:\Users\Ryan\Programming\Zig\zig\lib\libc\mingw\crt\crtexe.c:188:0: 0x294b81b in mainCRTStartup (crt2.obj)
  ret = __tmainCRTStartup ();

???:?:?: 0x7ffe569f7343 in ??? (KERNEL32.DLL)
???:?:?: 0x7ffe571426b0 in ??? (ntdll.dll)

%errorlevel% -2147483645