Fix some lint warnings (#3)

* Fix some lint warnings
* Exclude some paths in cpplint
* Fix some tests and formatting

Makefile

 # Copyright (c) Microsoft Corporation.
 # Licensed under the MIT License.
 
-CPPSOURCES := $(shell find $(CURDIR) -regextype posix-extended -regex '.*\.(c|cpp|h|hpp|cc|cxx|cu)')
+CPPSOURCES := $(shell find $(CURDIR) \
+	-path $(CURDIR)/.git -prune -o \
+	-path $(CURDIR)/.venv -prune -o \
+	-path $(CURDIR)/build -prune -o \
+	-path $(CURDIR)/third_party -prune -o \
+	-regextype posix-extended -regex '.*\.(c|cpp|h|hpp|cc|cxx|cu)' -print)
+CLANG_FORMAT ?= clang-format-14
 
 .PHONY: cpplint cppformat cppbuild thirdparty postinstall
 
 cpplint:
-	clang-format --verbose --dry-run --Werror $(CPPSOURCES)
+	$(CLANG_FORMAT) --verbose --dry-run --Werror $(CPPSOURCES)
 
 cppformat:
-	clang-format --verbose -i $(CPPSOURCES)
+	$(CLANG_FORMAT) --verbose -i $(CPPSOURCES)
 
 cppbuild:
 	cd ./superbench/benchmarks/ && bash build.sh

superbench/benchmarks/micro_benchmarks/dtk_gemm_flops_performance.py

@@ -253,7 +253,7 @@ def add_parser_arguments(self):
             help='Set default device to be used for the benchmark process.',
         )
 
-    def _preprocess(self):
+    def _preprocess(self):    # noqa: C901
         """Preprocess/preparation operations before the benchmarking.
 
         Return:

superbench/benchmarks/micro_benchmarks/dtk_hipblaslt_function.py

@@ -126,7 +126,6 @@ def _process_raw_result(self, cmd_idx, raw_output):
 
             batch_count_index = header.index('batch_count')
             m_index = header.index('m')
-            n_index = header.index('n')
             k_index = header.index('k')
             gflops_index = header.index('hipblaslt-Gflops')
 

superbench/benchmarks/micro_benchmarks/gpu_copy_performance/gpu_copy.cu

@@ -1046,9 +1046,9 @@ int RunAllToAllBench(const Opts &opts, int gpu_count, int src_rank, int dst_rank
                     return -1;
                 }
             }
-            SMOneToAllCopyKernel<<<gpu_count * opts.all_to_all_num_thread_blocks_per_rank,
-                                   thread_block_size, 0, streams[rank]>>>(
-                (ulong2 **)dst_buffer_gpu_args[rank], (ulong2 *)src_buffers_gpu[rank], opts.size, rank, dst_rank,
+            SMOneToAllCopyKernel<<<gpu_count * opts.all_to_all_num_thread_blocks_per_rank, thread_block_size, 0,
+                                   streams[rank]>>>((ulong2 **)dst_buffer_gpu_args[rank],
+                                                    (ulong2 *)src_buffers_gpu[rank], opts.size, rank, dst_rank,
                                                     gpu_count);
             if (i == opts.num_warm_up + opts.num_loops - 1) {
                 cuda_err = cudaEventRecord(stop_events[rank], streams[rank]);

superbench/runner/runner.py

@@ -530,7 +530,7 @@ def __merge_monitor_metrics(self, node_path):
 
         return metrics_summary
 
-    def _run_proc(self, benchmark_name, mode, vars):
+    def _run_proc(self, benchmark_name, mode, vars):    # noqa: C901
         """Run the process.
 
         Args:

tests/benchmarks/micro_benchmarks/test_dtk_hipblaslt_function.py

@@ -99,7 +99,8 @@ def test_hipblaslt_gemm_result_parsing(self):
 
 Is supported 1 / Total solutions: 1
 [0]:transA,transB,grouped_gemm,batch_count,m,n,k,alpha,lda,stride_a,beta,ldb,stride_b,ldc,stride_c,ldd,stride_d,a_type,b_type,c_type,d_type,compute_type,scaleA,scaleB,scaleC,scaleD,amaxD,activation_type,bias_vector,bias_type,hipblaslt-Gflops,us
-    N,N,0,1,4096,4096,4096,1,4096,16777216,0,4096,16777216,4096,16777216,4096,16777216,f32_r,f32_r,f32_r,f32_r,f32_r,0,0,0,0,0,none,0,non-supported type,1595.18,86159.1
+    N,N,0,1,4096,4096,4096,1,4096,16777216,0,4096,16777216,4096,16777216,4096,16777216,
+    f32_r,f32_r,f32_r,f32_r,f32_r,0,0,0,0,0,none,0,non-supported type,1595.18,86159.1
 """
         self.assertTrue(benchmark._process_raw_result(0, example_raw_output))
         self.assertEqual(ReturnCode.SUCCESS, benchmark.return_code)

tests/runner/test_runner.py

@@ -188,7 +188,7 @@ def test_get_mode_command(self):
                 },
                 'expected_command': (
                     'mpirun -tag-output -allow-run-as-root -hostfile hostfile -map-by ppr:8:node -bind-to numa '
-                    '-mca coll_hcoll_enable 0 -x SB_MICRO_PATH=/sb -x FOO=BAR -x RANK=2 -x NUM=8 '
+                    '-mca coll_hcoll_enable 0 -x "SB_MICRO_PATH=/sb" -x "FOO=BAR" -x "RANK=2" -x "NUM=8" '
                     f'sb exec --output-dir {self.sb_output_dir} -c sb.config.yaml -C superbench.enable=foo'
                 ),
             },
@@ -213,7 +213,7 @@ def test_get_mode_command(self):
                 },
                 'expected_command': (
                     'mpirun -tag-output -allow-run-as-root -host localhost:8 -bind-to numa '
-                    '-mca coll_hcoll_enable 0 -x SB_MICRO_PATH=/sb -x FOO=BAR -x RANK=2 -x NUM=8 '
+                    '-mca coll_hcoll_enable 0 -x "SB_MICRO_PATH=/sb" -x "FOO=BAR" -x "RANK=2" -x "NUM=8" '
                     f'sb exec --output-dir {self.sb_output_dir} -c sb.config.yaml -C superbench.enable=foo'
                 ),
             },
@@ -225,6 +225,7 @@ def test_get_mode_command(self):
                     'node_num': 1,
                     'proc_num': 4,
                     'proc_rank': 1,
+                    'bind_to': 'numa',
                     'mca': {},
                     'env': {
                         'NCCL_BUFFSIZE': 4194304,
@@ -233,7 +234,7 @@ def test_get_mode_command(self):
                     },
                 },
                 'expected_command': (
-                    "mpirun -tag-output -allow-run-as-root -host localhost:4 -bind-to numa "
+                    'mpirun -tag-output -allow-run-as-root -host localhost:4 -bind-to numa  '
                     '-x "NCCL_BUFFSIZE=4194304" -x "NCCL_RINGS=0 1 2 3|0 3 2 1" -x PATH '
                     f'sb exec --output-dir {self.sb_output_dir} -c sb.config.yaml -C superbench.enable=foo'
                 ),