Jump to content

dev7117

Member
  • Posts

    6
  • Joined

  • Last visited

Awards

This user doesn't have any awards

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

dev7117's Achievements

  1. Alright man, sorry I couldn't help. I hadn't seen a cloud option or anything about 24/7 / data sized mentioned so thought I would see if another solution might help. Best of luck with your work
  2. I know this might be little late, but any reason you are trying to go your own hardware route? Are you looking for a solution for your company or personal work? Security concerns maybe? Going off the little hardware knowledge part I would suggest going with a cloud based option. You can use s3 (or any competitors options) for cheap data storage of large sets and use the prebuilt non continuous running tools. If you are using tensor flow for your AI workflows then maybe go the GCP route and use something like Dataproc for data prep / pipelines and the re persist in blob storage and only pay for CPU and GPU time when you need it. Seems like a much cheaper / less money sucking option if you are not really prepared to handle to nitty gritty part of managing the hardware and your largest bottleneck (the network). I mainly work with Hadoop on the day to day but assuming your workload for data science / ai is going fluctuate between high cpu/mem options onto gp workloads for ai. That is likely to be the cheapest and most powerful option. Im not trying to push you any one direction but ai workloads / data science process while easier to deploy on a single box are not scalable or meant for a single box. Getting a distributed system is really going to save money and time in the long run. Have you considering syncing some time into continuous integration to help manage the deployment process? It will take time upfront but might help with those headaches down the road without having to sit on a single box.
  3. I recently purchased this machine with the same ideas you have stated here. I will say that the 1050 performance has been less then great. I cant run wow passed a 4/10 on the graphical settings and many games are running far below what i would have expected. I opted for the top of the line model but with the 1080p screen for better battery life. Other then this, I would consider the machine to be the perfect laptop for myself personally. But since the 1050 isnt running games well for me, I am sitting in the spot of needed the eGpu earlier then I expected. How did you plan to accomplish this? razer core with a bios update to allow it?
×