                        ___                    ___          ___     
                       /\__\                  /\  \        /\  \   
                      /:/  /       ___        \:\  \       \:\  \   
                     /:/  /       /\__\        \:\  \       \:\  \  
                    /:/  /  ___  /:/__/    _____\:\  \  _____\:\  \ 
                   /:/__/  /\__\/::\  \   /::::::::\__\/::::::::\__\
                   \:\  \ /:/  /\/\:\  \__\:\~~\~~\/__/\:\~~\~~\/__/
                    \:\  /:/  /    \:\/\__\\:\  \       \:\  \      
                     \:\/:/  /      \::/  / \:\  \       \:\  \     
                      \::/  /       /:/  /   \:\__\       \:\__\   
                       \/__/        \/__/     \/__/        \/__/

CINN : a Compiler Infrastructure for Neural Networks

Install | Roadmap

The project CINN is a machine learning compiler and executor for multiple hardwares. It is designed to provide multiple layers of APIs to make DNN computation graph easier to define, faster to execute, and more convenient to extend with hardware backends. Currently, it targets X86 CPUs and NVidia GPUs.

This project is under active development.

Example

Let's take C++ APIs as an example, the corresponding Python APIs are available and just differ little.

Load a PaddlePaddle model and execute

#include "cinn/frontend/interpreter.h"
using cinn::hlir::framework;

Interpreter inter({"input0"}/*list of inputs' names*/, 
                  {{1, 30}}/*list of inputs' shapes*/);
inter.LoadPaddleModel(/*string of model directory*/);
auto input_handle = inter.GetTensor("input0");
auto output_handle = inter.GetTensor("output0");
// set data to input_handle
inter.Run();
// get output content from output_handle

Note that for api LoadPaddleModel, the params_combined param is set to be false by default.

Use CINN lower level DSL to define some computation and execute

The following is a naive matrix-multiplication implementation using the CINN DSL

#include "cinn/cinn.h"
using namespace cinn;

// Declare constants
Expr M(10), N(20), K(30);

// Declare the inputs 
auto A = Placeholder<float>("A", {M, K});
auto B = Placeholder<float>("B", {K, N});

auto k1 = Var(K.as_int32(), "k1");
auto C  = Compute(
    {M, N}, [&](Var i, Var j) { return ReduceSum(A(i, k1) * B(k1, j), {k1}); }, "C");

Target target = common::DefaultHostTarget();

int block_size = 32;

// The stages holds all the schedules for each tensors.
auto stages = CreateStages({C});

// Blocking optimization by loop tiling stragety.
auto [i_outer, i_inner, j_outer, j_inner] = stages[C]->Tile(0, 1, bn, bn);
auto [k_outer, k_inner]                   = stages[C]->Split("k0", 4);
stages[C]->Reorder({i_outer, j_outer, k_outer, k_inner, i_inner, j_inner});

// Generate C source code:
Module::Builder builder("module_block", target);
auto func = Lower("matmul_block", stages, {A, B, C});
builder.AddFunction(func);

CodeGenCX86 compiler(target, CodeGenCX86::Feature::AVX512);
Outputs outputs;
outputs = outputs.c_header("./test02_matmul_block.h").c_source("./test02_matmul_block.cc");
compiler.Compile(builder.Build(), outputs);

This can generate the optimized C source code like

void matmul_block(void* _args, int32_t num_args)
{
  const cinn_buffer_t* _A = cinn_pod_value_to_buffer_p(&(((cinn_pod_value_t*)(_args))[0]));
  const cinn_buffer_t* _B = cinn_pod_value_to_buffer_p(&(((cinn_pod_value_t*)(_args))[1]));
  cinn_buffer_t* _C = cinn_pod_value_to_buffer_p(&(((cinn_pod_value_t*)(_args))[2]));
  cinn_buffer_malloc((void*)(0), _C);
  const float* A = ((const float*)(_A->memory));
  const float* B = ((const float*)(_B->memory));
  float* C = ((float*)(_C->memory));
  float* C__reduce_init = ((float*)(_C->memory));
  for (int32_t i = 0; i < 1024; i += 1) {
    for (int32_t j = 0; j < 1024; j += 1) {
      C__reduce_init[((1024 * i) + j)] = 0;
    };
  };
  for (int32_t i_outer = 0; i_outer < 32; i_outer += 1) {
    for (int32_t j_outer = 0; j_outer < 32; j_outer += 1) {
      for (int32_t k0_outer = 0; k0_outer < 256; k0_outer += 1) {
        for (int32_t k0_inner = 0; k0_inner < 4; k0_inner += 1) {
          for (int32_t i_inner = 0; i_inner < 32; i_inner += 1) {
            for (int32_t j_inner = 0; j_inner < 32; j_inner += 1) {
              C[((1024 * i_inner) + ((32768 * i_outer) + ((32 * j_outer) + j_inner)))] = (C[((1024 * i_inner) + ((32768 * i_outer) + ((32 * j_outer) + j_inner)))] + (A[((1024 * i_inner) + ((32768 * i_outer) + ((4 * k0_outer) + k0_inner)))] * B[((32 * j_outer) + ((1024 * k0_inner) + ((4096 * k0_outer) + j_inner)))]));
            };
          };
        };
      };
    };
  };
  cinn_buffer_free((void*)(0), _C);
}

Change the CodeGenCX86 usage to CodeGenLLVM, it will produce a LLVM JIT-compiled function instead which can invoke realtime.

How it works

The CINN lowers a traditional DNN model into a two-level intermediate representation(IR), the high-level IR(HLIR) and CINN IR.

The HLIR helps to define some domain-specific computation and perform some overall optimization on the IR-graph; the CINN IR helps to represent some computation semantic and finally lower to a hardware backend.

Both levels of IR have the similar SSA graph, analysis and optimization facilities.

CINN is based on the polyhedral compilation thus it is easy to extend with more loop optimizations. The schedule transform is applied between the lowering from HLIR to CINN IR.

The overall architecture is as follows

Getting Started

Compile and execute the code

To compile the CINN's code, one need to build the docker image first

cd tools/docker
ln -s Dockerfile.cpu Dockerfile
docker build . -t cinn-dev

Then start a docker container, and compile the code inside it

# inside the docker container

# compile and install isl
sh tools/ci_build.sh

# compile the tests and python library with X86 backends
./build.sh ci

# compile the tests and python library with NVGPU(CUDA) backends
./build.sh gpu_on ci

After compilation, you can launch the C++ and python tests

cd build
ctest -V

Concepts

There are two levels of APIs in CINN, the higher level is HLIR and the lower level is CINN IR, both contain some concepts.

In HLIR

Primitive Emitter(PE), encapsulates the computation of different tensor-based algorithms,
frontend::Interpreter, the container to execute a model (of PaddlePaddle),
frontend::Program, the program helps to define a machine learning computation,
hlir::framework::Tensor, multi-dimensional arrays helps to manage a memory buffer.
hlir::framework::Program, the final executable program in runtime. It holds many basic executable elements.
hlir::framework::Graph, the graph that represents the structure of a model. Each node in the graph represents an operator (conv2d, relu, mul, etc.).
hlir::framework::GraphCompiler, the compiler that transforms the graph representation(hlir::framework::Graph) of a model into an executable program(hlir::framework::Program). In CINN IR
Compute, the method to define a computation,
Lower, the method to lower a computation to the corresponding IR,
LoweredFunc, the function defined in CINN IR,
Var, a scalar variable,
Expr, an expression represents any CINN IR node(no specified Statement node),
Stage, holds some schedule details of a tensor,

Reference the API usage

Read the code in the tests

For Python API, reference the code inside python/tests.

The C++ API locates in cinn/*/*_test.cc, the high level API locates in hlir/frontend, the lower level API is in cinn/cinn.h.

License

CINN is licensed under the Apache 2.0 license.

limin2021/CINN