keyxuliang / chgemm

symmetric int8 gemm

chgemm

---

chgemm is an symmetric int8 project, which is slightly different from BLAS sgemm:

1. when you input an int8_t type of matrix [-127,+127], you will get an int32_t one. PS: pay attention to the overflow;
2. considering the application scene of the deeep learning, the packAB interface is open and can be adjusted;
3. the common design plan is alpha*A*B+beta*C=C, but mine is C=A*B, because they have no utility in deep learning inference;
4. row major;
5. the speed of this project is quicker than any other projects'.

chgemm 是一个 int8 gemm 工程，与 BLAS gemm 不完全相同：

1. 输入为 [-127, +127] 范围内的 int8_t 类型矩阵，输出为 int32_t 矩阵。需注意溢出；
2. 更多地为深度学习应用场景考虑，packAB 接口暴露出来可以调整；
3. 实现为 C = A * B。alpha 和 beta 在深度学习推理中无实用意义；
4. 行主序实现，放弃远古 fortran 时代的列主序；
5. 不低于其他项目的 symmint8 gemm 速度。

---

test result

Compiled on RK3399 with -O3 flag. The current peek can be 18.6 gflops, and the orange line is the single-core fp32 limit(14.3 gflops).

速度

-O3 编译，目前在 rk3399 单核结果。目前极限可以到 18.6 gflops，橙线是 rk3399 单核 fp32 极限。

---

使用方式

参照 MMult_4x8_21.c 调用矩阵乘法，将代码嵌入到自己的项目中。可根据推理库的实现做相应修改。

application with chgemm inside

chgemm is pleased to support ncnn available, check gemm_symm_int8.h.