Nccl telemetry feature

[dmitry-monakhov]

Work in par with https://github.com/poolsideai/nccl-telemetry-py

from nccl_telemetry import NCCLProfiler, Phase

profiler = NCCLProfiler()

for step in range(num_steps):
    with profiler.phase(step, Phase.FORWARD):
        output = model(input)
        loss = criterion(output, target)

    with profiler.phase(step, Phase.BACKWARD):
        loss.backward()

    with profiler.phase(step, Phase.OPTIMIZER):
        optimizer.step()

• titan integration: TBD