Remove all C code and build system support for C. Enable bugprone-signed-char-misuse clang-tidy check

.clang-tidy

@@ -43,7 +43,6 @@ Checks: "*,
         -bugprone-macro-parentheses,
         -bugprone-narrowing-conversions,
         -bugprone-reserved-identifier,
-        -bugprone-signed-char-misuse,
         -bugprone-string-integer-assignment,
         -cert-dcl37-c,
         -cert-dcl50-cpp,

.github/workflows/build_and_lint.yml

@@ -88,9 +88,6 @@ jobs:
     - name: Display tidy_results_cpp.log
       if: ${{ (matrix.container.name == 'CUDA') && (always()) }}
       run: cat tidy_results_cpp.log
-    - name: Display tidy_results_c.log
-      if: ${{ (matrix.container.name == 'CUDA') && (always()) }}
-      run: cat tidy_results_c.log
     - name: Display tidy_results_gpu.log
       if: ${{ (matrix.container.name == 'CUDA') && (always()) }}
       run: cat tidy_results_gpu.log

Makefile

@@ -17,13 +17,11 @@ DIRS     := src src/analysis src/chemistry_gpu src/cooling src/cooling_grackle s
 
 SUFFIX ?= .$(TYPE).$(MACHINE)
 
-CFILES   := $(foreach DIR,$(DIRS),$(wildcard $(DIR)/*.c))
 CPPFILES := $(foreach DIR,$(DIRS),$(wildcard $(DIR)/*.cpp))
 GPUFILES := $(foreach DIR,$(DIRS),$(wildcard $(DIR)/*.cu))
 
 # Build a list of all potential object files so cleaning works properly
-CLEAN_OBJS := $(subst .c,.o,$(CFILES)) \
-              $(subst .cpp,.o,$(CPPFILES)) \
+CLEAN_OBJS := $(subst .cpp,.o,$(CPPFILES)) \
               $(subst .cu,.o,$(GPUFILES))
 
 # Check if it should include testing flags
@@ -49,30 +47,24 @@ ifeq ($(ADD_TEST_FLAGS), yes)
   SUFFIX    := $(strip $(SUFFIX)).tests
   LIBS      += -L$(GOOGLETEST_ROOT)/lib64 -pthread -lgtest -lhdf5_cpp
   TEST_FLAGS = -I$(GOOGLETEST_ROOT)/include
-  CFLAGS   += $(TEST_FLAGS)
   CXXFLAGS += $(TEST_FLAGS)
   GPUFLAGS += $(TEST_FLAGS)
 else
   # This isn't a test build so clear out testing related files
-  CFILES   := $(filter-out src/system_tests/% %_tests.c,$(CFILES))
   CPPFILES := $(filter-out src/system_tests/% %_tests.cpp,$(CPPFILES))
   CPPFILES := $(filter-out src/utils/testing_utilities.cpp,$(CPPFILES))
   GPUFILES := $(filter-out src/system_tests/% %_tests.cu,$(GPUFILES))
 endif
 
-OBJS     := $(subst .c,.o,$(CFILES)) \
-            $(subst .cpp,.o,$(CPPFILES)) \
+OBJS     := $(subst .cpp,.o,$(CPPFILES)) \
             $(subst .cu,.o,$(GPUFILES))
 
 #-- Set default compilers and flags
-CC                ?= cc
 CXX               ?= CC
 
-CFLAGS_OPTIMIZE   ?= -g -Ofast
 CXXFLAGS_OPTIMIZE ?= -g -Ofast -std=c++17
 GPUFLAGS_OPTIMIZE ?= -g -O3 -std=c++17
 
-CFLAGS_DEBUG      ?= -g -O0
 CXXFLAGS_DEBUG    ?= -g -O0 -std=c++17
 ifdef HIPCONFIG
   GPUFLAGS_DEBUG    ?= -g -O0 -std=c++17
@@ -82,13 +74,11 @@ endif
 
 BUILD             ?= OPTIMIZE
 
-CFLAGS            += $(CFLAGS_$(BUILD))
 CXXFLAGS          += $(CXXFLAGS_$(BUILD))
 GPUFLAGS          += $(GPUFLAGS_$(BUILD))
 
 #-- Add flags and libraries as needed
 
-CFLAGS   += $(DFLAGS) -Isrc
 CXXFLAGS += $(DFLAGS) -Isrc
 GPUFLAGS += $(DFLAGS) -Isrc
 
@@ -170,7 +160,7 @@ ifeq ($(findstring -DCHEMISTRY_GPU,$(DFLAGS)),-DCHEMISTRY_GPU)
   DFLAGS += -DSCALAR
 endif
 
-.SUFFIXES: .c .cpp .cu .o
+.SUFFIXES: .cpp .cu .o
 
 EXEC := bin/cholla$(SUFFIX)
 
@@ -184,27 +174,21 @@ DFLAGS      += $(MACRO_FLAGS)
 LIBS_CLANG_TIDY     := $(subst -I/, -isystem /,$(LIBS))
 LIBS_CLANG_TIDY     += -isystem $(MPI_ROOT)/include
 CXXFLAGS_CLANG_TIDY := $(subst -I/, -isystem /,$(LDFLAGS))
-CFLAGS_CLANG_TIDY   := $(subst -I/, -isystem /,$(CFLAGS))
 GPUFLAGS_CLANG_TIDY := $(subst -I/, -isystem /,$(GPUFLAGS))
 GPUFLAGS_CLANG_TIDY := $(filter-out -ccbin=mpicxx -fmad=false --expt-extended-lambda,$(GPUFLAGS))
 GPUFLAGS_CLANG_TIDY += --cuda-host-only --cuda-path=$(CUDA_ROOT) -isystem /clang/includes
 CPPFILES_TIDY := $(CPPFILES)
-CFILES_TIDY   := $(CFILES)
 GPUFILES_TIDY := $(GPUFILES)
 
 ifdef TIDY_FILES
   CPPFILES_TIDY := $(filter $(TIDY_FILES), $(CPPFILES_TIDY))
-  CFILES_TIDY   := $(filter $(TIDY_FILES), $(CFILES_TIDY))
   GPUFILES_TIDY := $(filter $(TIDY_FILES), $(GPUFILES_TIDY))
 endif
 
 $(EXEC): prereq-build $(OBJS)
 	mkdir -p bin/ && $(LD) $(LDFLAGS) $(OBJS) -o $(EXEC) $(LIBS)
 	eval $(EXTRA_COMMANDS)
 
-%.o: %.c
-	$(CC) $(CFLAGS) -c $< -o $@
-
 %.o: %.cpp
 	$(CXX) $(CXXFLAGS) -c $< -o $@
 
@@ -221,9 +205,8 @@ tidy:
 # - --warnings-as-errors=<string> Upgrade all warnings to error, good for CI
 	clang-tidy --verify-config
 	(time clang-tidy $(CLANG_TIDY_ARGS) $(CPPFILES_TIDY) -- $(DFLAGS) $(CXXFLAGS_CLANG_TIDY) $(LIBS_CLANG_TIDY)) > tidy_results_cpp.log 2>&1 & \
-	(time clang-tidy $(CLANG_TIDY_ARGS) $(CFILES_TIDY)   -- $(DFLAGS) $(CFLAGS_CLANG_TIDY)   $(LIBS_CLANG_TIDY)) > tidy_results_c.log   2>&1 & \
 	(time clang-tidy $(CLANG_TIDY_ARGS) $(GPUFILES_TIDY) -- $(DFLAGS) $(GPUFLAGS_CLANG_TIDY) $(LIBS_CLANG_TIDY)) > tidy_results_gpu.log 2>&1 & \
-	for i in 1 2 3; do wait -n; done
+	for i in 1 2; do wait -n; done
 
 clean:
 	rm -f $(CLEAN_OBJS)

builds/make.host.c3po

@@ -1,8 +1,5 @@
 #-- Compiler and flags for different build type
-CC                = mpicc
 CXX               = mpicxx
-CFLAGS_DEBUG      = -g -O0
-CFLAGS_OPTIMIZE   = -g -O2
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17 ${F_OFFLOAD}
 CXXFLAGS_OPTIMIZE = -g -Ofast -std=c++17 ${F_OFFLOAD}
 GPUFLAGS_DEBUG    = -g -G -cudart shared -O0 -std=c++17 -ccbin=mpicxx -Xcompiler -rdynamic

builds/make.host.crc

@@ -1,9 +1,6 @@
 
 #-- Compiler and flags for different build type
-CC                = mpicc
 CXX               = mpicxx
-CFLAGS_DEBUG      = -g -O0
-CFLAGS_OPTIMIZE   = -Ofast
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17
 CXXFLAGS_OPTIMIZE = -Ofast -std=c++17
 GPUFLAGS_OPTIMIZE = -g -O3 -std=c++17

builds/make.host.frontier

@@ -1,13 +1,9 @@
 #-- make.host for Frontier at the OLCF with
 #-- Compiler and flags for different build type
-CC                = cc
 CXX               = CC
 #GPUCXX           ?= CC -x hip
 GPUCXX           ?= hipcc
 
-CFLAGS_DEBUG      = -g -O0
-CFLAGS_OPTIMIZE   = -g -O2
-
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17
 CXXFLAGS_OPTIMIZE = -g -Ofast -std=c++17 -Wno-unused-result
 

builds/make.host.github

@@ -1,8 +1,5 @@
 #-- Compiler and flags for different build type
-CC                = mpicc
 CXX               = mpicxx
-CFLAGS_DEBUG      = -g -O0
-CFLAGS_OPTIMIZE   = -g -O2
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17 ${F_OFFLOAD}
 CXXFLAGS_OPTIMIZE = -Ofast -std=c++17 ${F_OFFLOAD}
 GPUFLAGS_DEBUG    = -g -G -cudart shared -O0 -std=c++17
@@ -27,8 +24,6 @@ GOOGLETEST_ROOT := ${GOOGLETEST_ROOT}
 
 ifdef HIPCONFIG
 	MPI_ROOT := ${MPI_ROOT}
-	CFLAGS_DEBUG      += -fPIE
-	CFLAGS_OPTIMIZE   += -fPIE
 	CXXFLAGS_DEBUG    += -fPIE
 	CXXFLAGS_OPTIMIZE += -fPIE
 	GPUFLAGS_DEBUG    += -fPIE

builds/make.host.lux

@@ -1,10 +1,7 @@
 #-- make.inc for the Shamrock Server
 
 #-- Compiler and flags for different build type
-CC                = mpicc
 CXX               = mpicxx
-CFLAGS_DEBUG      = -g -O0
-CFLAGS_OPTIMIZE   = -g -O2
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17
 CXXFLAGS_OPTIMIZE = -Ofast -std=c++17
 GPUFLAGS         = -std=c++17

builds/make.host.poplar

@@ -1,10 +1,7 @@
 #-- make.inc for Poplar, COE cluster at HPE
 
 #-- Compiler and flags for different build type
-CC                = cc
 CXX               = CC
-CFLAGS_DEBUG      = -g -O0 ${F_OFFLOAD}
-CFLAGS_OPTIMIZE   = -Ofast ${F_OFFLOAD}
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17 ${F_OFFLOAD}
 CXXFLAGS_OPTIMIZE = -Ofast -std=c++17 ${F_OFFLOAD}
 

builds/make.host.poplar.aomp

@@ -1,11 +1,8 @@
 #-- make.inc for Poplar, COE cluster at HPE
 
 #-- Compiler and flags for different build type
-CC                = mpicc
 CXX               = mpicxx
 HIPCONFIG         = $(shell hipconfig -C)
-CFLAGS_DEBUG      = -g -O0
-CFLAGS_OPTIMIZE   = -Ofast
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17
 CXXFLAGS_OPTIMIZE = -Ofast -std=c++17
 

builds/make.host.poplar.cce+hip

@@ -1,11 +1,8 @@
 #-- make.inc for Poplar, HPE COE cluster
 
 #-- Compiler and flags for different build type
-CC                = cc
 CXX               = CC
 HIPCONFIG         = $(shell hipconfig -C)
-CFLAGS_DEBUG      = -g -O0
-CFLAGS_OPTIMIZE   = -Ofast
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17
 CXXFLAGS_OPTIMIZE = -Ofast -std=c++17
 

builds/make.host.shamrock

@@ -1,12 +1,9 @@
 #-- make.inc for the Lux Cluster
 
 #-- Compiler and flags for different build type
-CC                = mpicc
 CXX               = mpicxx
 #CC                = gcc
 #CXX               = g++
-CFLAGS_DEBUG      = -g -O0
-CFLAGS_OPTIMIZE   = -g -O2
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17
 CXXFLAGS_OPTIMIZE = -Ofast -std=c++17
 

builds/make.host.spock

@@ -1,12 +1,8 @@
 #-- make.inc for Spock EAS at the OLCF with
 
 #-- Compiler and flags for different build type
-CC                = cc
 CXX               = CC
 
-CFLAGS_DEBUG      = -g -O0
-CFLAGS_OPTIMIZE   = -g -O2
-
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17
 CXXFLAGS_OPTIMIZE = -g -Ofast -std=c++17
 

builds/make.host.summit

@@ -2,10 +2,7 @@
 #   https://www.olcf.ornl.gov/summit/
 
 #-- Compiler and flags for different build type
-CC                = mpicc
 CXX               = mpicxx
-CFLAGS_DEBUG      = -g -O0
-CFLAGS_OPTIMIZE   = -g -O2
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17 ${F_OFFLOAD}
 CXXFLAGS_OPTIMIZE = -Ofast -std=c++17 ${F_OFFLOAD}
 GPUFLAGS_DEBUG    = -g -O0 -std=c++17 -ccbin=mpicxx -G -cudart shared

builds/make.host.tornado

@@ -1,10 +1,7 @@
 #-- make.inc for the Lux Cluster
 
 #-- Compiler and flags for different build type
-CC                = gcc
 CXX               = g++
-CFLAGS_DEBUG      = -g -O0
-CFLAGS_OPTIMIZE   = -g -O2
 CXXFLAGS_DEBUG    = -g -O0 -std=c++17
 CXXFLAGS_OPTIMIZE = -Ofast -std=c++17
 

builds/setup.poplar.aomp.sh

@@ -1,12 +1,11 @@
 #!/bin/bash
 
 module purge
-module load craype-x86-naples craype-network-infiniband 
+module load craype-x86-naples craype-network-infiniband
 module load shared slurm
 module use /home/users/twhite/share/modulefiles
 module load ompi/4.0.4-rocm-3.9 hdf5
 
-export OMPI_CC=$(which clang)
 export OMPI_CXX=$(which clang)
 
 export CHOLLA_MACHINE=poplar.aomp

src/mpi/mpi_routines.cpp

@@ -8,7 +8,6 @@
 
   #include "../global/global.h"
   #include "../io/io.h"
-  #include "../mpi/MPI_Comm_node.h"
   #include "../mpi/cuda_mpi_routines.h"
   #include "../utils/error_handling.h"
 
@@ -21,7 +20,6 @@ int procID_node; /*process rank on node*/
 int nproc_node;  /*number of MPI processes on node*/
 
 MPI_Comm world; /*global communicator*/
-MPI_Comm node;  /*global communicator*/
 
 MPI_Datatype MPI_CHREAL; /*set equal to MPI_FLOAT or MPI_DOUBLE*/
 
@@ -216,8 +214,6 @@ void InitializeChollaMPI(int *pargc, char **pargv[])
   }
   #endif
 
-  /*set up node communicator*/
-  node = MPI_Comm_node(&procID_node, &nproc_node);
   // #ifdef ONLY_PARTICLES
   // chprintf("ONLY_PARTICLES: Initializing without CUDA support.\n");
   // #else