ggml : add ggml-common.h to deduplicate shared code

[ggerganov]

close #5220

Add ggml-common.h that contains common bits used in the different backends. The goal is to reduce code duplication. For now, it contains just the quantization related constants, but can potentially get extended with more things in the future

TODO:

• Reuse quantum tables in SYCL backend
• Try to move quantum block structs in ggml-common.h and reuse (see ggml : reuse quantum structs across backends #5943)

[ggerganov]

This could cause some trouble potentially, because it looks like the Metal compiler is looking for headers just in the folder where the binary is located and I can't seem to find a way to pass an include path or make it search in the current working dir.

For example, the following command no longer works:

make -j tests && ./tests/test-backend-ops -b Metal
I ccache found, compilation results will be cached. Disable with LLAMA_NO_CCACHE.
I llama.cpp build info: 
I UNAME_S:   Darwin
I UNAME_P:   arm
I UNAME_M:   arm64
I CFLAGS:    -I. -Icommon -D_XOPEN_SOURCE=600 -D_DARWIN_C_SOURCE -DNDEBUG -DGGML_USE_ACCELERATE -DACCELERATE_NEW_LAPACK -DACCELERATE_LAPACK_ILP64 -DGGML_USE_METAL  -std=c11   -fPIC -O3 -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function -Wshadow -Wstrict-prototypes -Wpointer-arith -Wmissing-prototypes -Werror=implicit-int -Werror=implicit-function-declaration -pthread -Wunreachable-code-break -Wunreachable-code-return -Wdouble-promotion 
I CXXFLAGS:  -std=c++11 -fPIC -O3 -Wall -Wextra -Wpedantic -Wcast-qual -Wno-unused-function -Wmissing-declarations -Wmissing-noreturn -pthread   -Wunreachable-code-break -Wunreachable-code-return -Wmissing-prototypes -Wextra-semi -I. -Icommon -D_XOPEN_SOURCE=600 -D_DARWIN_C_SOURCE -DNDEBUG -DGGML_USE_ACCELERATE -DACCELERATE_NEW_LAPACK -DACCELERATE_LAPACK_ILP64 -DGGML_USE_METAL 
I NVCCFLAGS: -std=c++11 -O3 
I LDFLAGS:   -framework Accelerate -framework Foundation -framework Metal -framework MetalKit 
I CC:        Apple clang version 15.0.0 (clang-1500.1.0.2.5)
I CXX:       Apple clang version 15.0.0 (clang-1500.1.0.2.5)

Testing 2 backends

Backend 1/2 (CPU)
  Skipping
Backend 2/2 (Metal)
ggml_metal_init: allocating
ggml_metal_init: found device: Apple M2 Ultra
ggml_metal_init: picking default device: Apple M2 Ultra
ggml_metal_init: default.metallib not found, loading from source
ggml_metal_init: GGML_METAL_PATH_RESOURCES = nil
ggml_metal_init: error: could not use bundle path to find ggml-metal.metal, falling back to trying cwd
ggml_metal_init: loading 'ggml-metal.metal'
ggml_metal_init: error: Error Domain=MTLLibraryErrorDomain Code=3 "program_source:4:10: fatal error: 'ggml-common.h' file not found
#include "ggml-common.h"
         ^~~~~~~~~~~~~~~
" UserInfo={NSLocalizedDescription=program_source:4:10: fatal error: 'ggml-common.h' file not found
#include "ggml-common.h"
         ^~~~~~~~~~~~~~~
}
GGML_ASSERT: tests/test-backend-ops.cpp:2269: backend != NULL
Abort trap: 6

A workaround is to make a symlink:

ln -sfn ../ggml-common.h ./tests/

[slaren]

Yeah that's not very nice, it will also require the header file even with LLAMA_METAL_EMBED_LIBRARY. We could probably do something like cat ggml-common.h ggml-metal.metal > ggml-metal-bundled.metal during the build to avoid the include.