benchmarks/classy.py [9:23]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@torch.inference_mode()
def benchmark_torch_function(iters, f, *args, **kwargs):
    f(*args, **kwargs)
    if torch.cuda.is_available():
        torch.cuda.synchronize()
        start_event = torch.cuda.Event(enable_timing=True)
        end_event = torch.cuda.Event(enable_timing=True)
        start_event.record()
    else:
        t0 = time.time()
    for _ in range(iters):
        f(*args, **kwargs)
    if torch.cuda.is_available():
        end_event.record()
        torch.cuda.synchronize()
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



benchmarks/mha_cuda.py [6:20]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
@torch.inference_mode()
def benchmark_torch_function(iters, f, *args, **kwargs):
    f(*args, **kwargs)
    if torch.cuda.is_available():
        torch.cuda.synchronize()
        start_event = torch.cuda.Event(enable_timing=True)
        end_event = torch.cuda.Event(enable_timing=True)
        start_event.record()
    else:
        t0 = time.time()
    for _ in range(iters):
        f(*args, **kwargs)
    if torch.cuda.is_available():
        end_event.record()
        torch.cuda.synchronize()
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -