GitLab

**Description**
Add rocBLAS building logic in third_party.

**Major Revision**
- Add rocm_rocblas target in third_party/Makefile.
- Add rocblas building logic

**Description**
Support rocm in third_party/makefile and add rccl-tests as a submodule with building logic.

**Major Revision**
- Support rocm in third_party/makefile
- Add rccl-tests as a submodule 
- Add build logic in third_party/Makefile for rccl-tests 

**Description**
Support rocm in third_party/makefile.

**Major Revision**
- Split rocm and cuda target in makefile
- Add target in dockerfile

**Description**
Add FIO benchmark tool into third-party dependency.

**Major Revision**
- Add FIO submodule into third-party directory and modify Makefile to enable it.

Add perftest as a submodule and add build logic

Benchmarks: Build Pipeline - Add nccl-tests as a submodule and add build logic.

Benchmarks: Fix bug - fix bug of third_party/cuda-samples git checkout issue when building docker (#126)

* fix bug in docker build of third_party/cuda-samples

Add microbenchmark, example, test, config for cuda memory performance and Add cuda-samples(tag with cuda version) as git submodule and update related makefile

* add cutlass as submodule.
* add build script for cutlass.
* only support compute capability 7.0(V100) and 8.0(A100)