Closed monajalal closed 4 years ago
Hi @monajalal ,
RAM: ? HDD/SSD:?
I am not sure what would be ideal for our case. I think it will only become clear after we look at the usage profile on the AWS console after a week or two.
Currently trying to decide if would be okay to downgrade from p3.2xlarge to p2.xlarge would be feasible for us or not?
Like I said in the email, I think running the web server on a GPU instance is too expensive. I think we should focus on making the system asynchronous(#301 ) so that the user doesnt have to keep their tab open.
If necessary, in the future, there's also the possibility of modifying the code such that we spin up a separate GPU instance for limited slots of time(which means we get charged GPU prices for limited slots of time) to handle the background jobs. The web server itself would remain on a CPU instance. But I think that's a stretch goal, one we can reach for, in the future, after we deal with more pertinent issues.
Closing since we reached a decision yesterday.
RAM: ? HDD/SSD:?
Currently trying to decide if would be okay to downgrade from p3.2xlarge to p2.xlarge would be feasible for us or not?
(p3.2xlarge) would cost us 32460= $4320 for duration of two months (p2.xlarge) would cost us 0.92460=$1296 for duration of two months
Currently, running on a p3.2xlarge EC2 instance: