Code cleanup in build_utils.py: Improve backend detection with _Backe… #3185

Job	Run time
gather-models-gpu	12s
gather-models-cpu	13s
torchchat-command-load-test (macos-14)	1m 28s
compile-gguf (macos-14)	4m 0s
runner-aoti (16-core-ubuntu)	4m 36s
runner-et (16-core-ubuntu)	7m 5s
runner-aoti (macos-14-xlarge)	3m 10s
runner-et (macos-14-xlarge)	4m 57s
test-gguf-util (macos-14)	4m 7s
test-mps (macos-m1-stable)	2m 7s
test-mps-dtype (macos-m1-stable)	2m 39s
test-tinystories-executorch (16-core-ubuntu)	16m 50s
test-tinystories-executorch (macos-14-xlarge)	15m 52s
test-torchao-experimental (macos-14-xlarge)	5m 52s
test-build-runner-et-android / linux-job	10m 41s
test-gpu-aoti-bfloat16 (cuda, stories15M) / linux-job	12m 46s
test-gpu-aoti-float16 (cuda, stories15M) / linux-job	11m 41s
test-gpu-aoti-float32 (cuda, stories15M) / linux-job	11m 55s
test-gpu-eval-sanity-check (cuda, stories15M) / linux-job	11m 20s
test-gpu-compile (cuda, stories15M) / linux-job	23m 0s
test-cpu-aoti (x86_64, stories15M)	17m 6s
test-cpu-compile (x86_64, stories15M)	15m 40s
test-cpu-compile (aarch64, stories15M)	9m 5s
test-cpu-aoti (aarch64, stories15M)	14m 18s
test-cpu-eval-sanity-check (x86_64, stories15M)	9m 25s
test-cpu-eval-sanity-check-float16 (x86_64, stories15M)	29m 31s
test-cpu-eval-sanity-check (aarch64, stories15M)	9m 23s
test-cpu-eval-sanity-check-float16 (aarch64, stories15M)	6m 36s
test-cpu-eval-sanity-check-float32 (x86_64, stories15M)	37m 34s
test-cpu-eval-sanity-check-float32 (aarch64, stories15M)	8m 48s
	5h 11m 57s

Provide feedback