Hacker News Reader

Comment by amenod | Hacker News Reader

> Genetic algorithm running on 56-core machine

Wow, just... wow. I wonder why they didn't utilize (multiple) GPUs instead? I would guess it would be far more efficient in all aspects. Especially now that there is TensorFlow & co.

View on HN

At the time the algorithm was running on one of the servers in the data centre and until recently we didn't have anything in the DC kitted out with GPUs (see http://blog.marcgravell.com/2016/05/how-i-found-cuda-or-rewr... for the reason behind that).

I'm not sure if we're planning to use GPUs for future iterations of algorithm testing; the GPUs are installed in hardware serving production traffic so it's unlikely to be used for experimental stuff like this. That said, a bunch of us just got Dell XPS-15s with the NVIDIA GTX 1050 so we can probably have a go on our local machines to get a feel for performance characteristics... watch this space, I guess!

Because they built it in C# by hand and had already spent too long on the wrong method, when R can do it 70 times faster?

It's not a neural network, so everything would be hand coded TF operations. It's unclear how that would be quicker than gradient decent.

I'd be more interested to know why they didn't try the learning-to-rank implementations in XGB or LightGBM?

$150k on GPGPUs seems a bit steep if you're going to throw that approach away anyway...

Why $150k? Ordinary GPU (<$1k) would probably be equivalent to 56 CPU cores, if not faster.

I imagine the reason for using CPU was that they would throw it out anyway, so there was no need to make an ideal solution - and they already had those 56 CPU cores.

That's precisely why, indeed. This 56-core machine is what was at hand at that moment. Genetic algorithms is something we tried, but we didn't have enough confidence in it to invest too much money and effort. Which was a good call, since we soon realized there was a much better approach.

That was just the price of some nVidia kit I was looking at, roughly.

I had no idea they have this, but it would probably be an overkill for SO. An ordinary (nVidia) GPU which starts at $150 is roughly >10 times faster at NN problems than 4 core CPU.

This: https://www.nvidia.com/en-us/data-center/dgx-1/

Or you can spin an instance on AWS of your choosing?

Yes, of course. I seem to remember that their data is still in their own datacenter though, so they might not be comfortable with using Amazon for this.

Reply on news.ycombinator.com