build : on Mac OS enable Metal by default (#2901)

* build : on Mac OS enable Metal by default

* make : try to fix build on Linux

* make : move targets back to the top

* make : fix target clean

* llama : enable GPU inference by default with Metal

* llama : fix vocab_only logic when GPU is enabled

* common : better `n_gpu_layers` assignment

* readme : update Metal instructions

* make : fix merge conflict remnants

* gitignore : metal

.gitignore

@@ -31,28 +31,29 @@ tmp/
 models/*
 models-mnt
 
-/main
-/quantize
-/quantize-stats
-/result
-/perplexity
-/embedding
-/train-text-from-scratch
-/convert-llama2c-to-ggml
-/simple
-/benchmark-matmult
-/vdot
-/server
 /Pipfile
+/baby-llama
+/beam-search
+/benchmark-matmult
+/convert-llama2c-to-ggml
 /embd-input-test
+/embedding
 /gguf
 /gguf-llama-simple
 /libllama.so
 /llama-bench
-/baby-llama
-/beam-search
+/main
+/metal
+/perplexity
+/quantize
+/quantize-stats
+/result
 /save-load-state
+/server
+/simple
 /speculative
+/train-text-from-scratch
+/vdot
 build-info.h
 arm_neon.h
 compile_commands.json