diff --git a/build.zig b/build.zig index ce27b8687..22cc6f099 100644 --- a/build.zig +++ b/build.zig @@ -65,8 +65,10 @@ const Maker = struct { if (m.target.result.abi == .gnu) { try m.addFlag("-D_GNU_SOURCE"); } + if (m.target.result.os.tag == .macos) { + try m.addFlag("-D_DARWIN_C_SOURCE"); + } try m.addFlag("-D_XOPEN_SOURCE=600"); - try m.addFlag("-DGGML_USE_VULKAN"); try m.addProjectInclude(&.{}); try m.addProjectInclude(&.{"common"}); @@ -76,7 +78,7 @@ const Maker = struct { fn obj(m: *const Maker, name: []const u8, src: []const u8) *Compile { const o = m.builder.addObject(.{ .name = name, .target = m.target, .optimize = m.optimize }); - if (std.mem.endsWith(u8, src, ".c")) { + if (std.mem.endsWith(u8, src, ".c") or std.mem.endsWith(u8, src, ".m")) { o.addCSourceFiles(.{ .files = &.{src}, .flags = m.cflags.items }); o.linkLibC(); } else { @@ -106,7 +108,6 @@ const Maker = struct { // linkLibCpp already add (libc++ + libunwind + libc) e.linkLibCpp(); } - e.linkSystemLibrary("vulkan"); m.builder.installArtifact(e); e.want_lto = m.enable_lto; return e; @@ -117,6 +118,31 @@ pub fn build(b: *std.Build) !void { var make = try Maker.init(b); make.enable_lto = b.option(bool, "lto", "Enable LTO optimization, (default: false)") orelse false; + const llama_vulkan = b.option(bool, "llama-vulkan", "Enable Vulkan backend for Llama, (default: false)") orelse false; + const llama_metal = b.option(bool, "llama-metal", "Enable Metal backend for Llama, (default: false, true for macos)") orelse (make.target.result.os.tag == .macos); + const llama_no_accelerate = b.option(bool, "llama-no-accelerate", "Disable Accelerate framework for Llama, (default: false)") orelse false; + const llama_accelerate = !llama_no_accelerate and make.target.result.os.tag == .macos; + + if (llama_accelerate) { + try make.addFlag("-DGGML_USE_ACCELERATE"); + try make.addFlag("-DACCELERATE_USE_LAPACK"); + try make.addFlag("-DACCELERATE_LAPACK_ILP64"); + } + + var extra_objs = ArrayList(*Compile).init(b.allocator); + + if (llama_vulkan) { + try make.addFlag("-DGGML_USE_VULKAN"); + const ggml_vulkan = make.obj("ggml-vulkan", "ggml-vulkan.cpp"); + try extra_objs.append(ggml_vulkan); + } + + if (llama_metal) { + try make.addFlag("-DGGML_USE_METAL"); + const ggml_metal = make.obj("ggml-metal", "ggml-metal.m"); + try extra_objs.append(ggml_metal); + } + const ggml = make.obj("ggml", "ggml.c"); const ggml_alloc = make.obj("ggml-alloc", "ggml-alloc.c"); const ggml_backend = make.obj("ggml-backend", "ggml-backend.c"); @@ -127,18 +153,13 @@ pub fn build(b: *std.Build) !void { const console = make.obj("console", "common/console.cpp"); const sampling = make.obj("sampling", "common/sampling.cpp"); const grammar_parser = make.obj("grammar-parser", "common/grammar-parser.cpp"); - // const train = make.obj("train", "common/train.cpp"); const clip = make.obj("clip", "examples/llava/clip.cpp"); - const ggml_vulkan = make.obj("ggml-vulkan", "ggml-vulkan.cpp"); + // const train = make.obj("train", "common/train.cpp"); - _ = make.exe("main", "examples/main/main.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo, sampling, console, grammar_parser, ggml_vulkan }); - // _ = make.exe("quantize", "examples/quantize/quantize.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo }); - // _ = make.exe("perplexity", "examples/perplexity/perplexity.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo }); - // _ = make.exe("embedding", "examples/embedding/embedding.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo }); - // _ = make.exe("finetune", "examples/finetune/finetune.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo, train }); - // _ = make.exe("train-text-from-scratch", "examples/train-text-from-scratch/train-text-from-scratch.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo, train }); + var exes = ArrayList(*Compile).init(b.allocator); - const server = make.exe("server", "examples/server/server.cpp", &.{ + var objs = ArrayList(*Compile).init(b.allocator); + try objs.appendSlice(&[_]*Compile{ ggml, ggml_alloc, ggml_backend, @@ -147,11 +168,40 @@ pub fn build(b: *std.Build) !void { common, buildinfo, sampling, + console, grammar_parser, clip, - ggml_vulkan, }); + try objs.appendSlice(extra_objs.items); + + const main = make.exe("main", "examples/main/main.cpp", objs.items); + try exes.append(main); + + // _ = make.exe("quantize", "examples/quantize/quantize.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo }); + // _ = make.exe("perplexity", "examples/perplexity/perplexity.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo }); + // _ = make.exe("embedding", "examples/embedding/embedding.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo }); + // _ = make.exe("finetune", "examples/finetune/finetune.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo, train }); + // _ = make.exe("train-text-from-scratch", "examples/train-text-from-scratch/train-text-from-scratch.cpp", &.{ ggml, ggml_alloc, ggml_backend, ggml_quants, llama, common, buildinfo, train }); + + const server = make.exe("server", "examples/server/server.cpp", objs.items); if (make.target.result.os.tag == .windows) { server.linkSystemLibrary("ws2_32"); } + try exes.append(server); + + for (exes.items) |e| { + if (llama_vulkan) { + e.linkSystemLibrary("vulkan"); + } + + if (llama_metal) { + e.linkFramework("Foundation"); + e.linkFramework("Metal"); + e.linkFramework("MetalKit"); + } + + if (llama_accelerate) { + e.linkFramework("Accelerate"); + } + } }