Computational power needed!

[StevenBoon]

Hi guys, I'm new to the LTT forum.

For my thesis I have been writing a C++ program that calculates mathematical sequences.

The problem is that my laptop is a bit on the slow side.. it's not terrible, but a Core i7-6820HQ just isn't going to get me where I want.

I am looking for a powerful computer that can run a program for an extended period of time (couple of hours or up to a weekend, if possible).

Can someone help me, or give an idea for where I can apply for free computational power?

In the attachments is a working version of the script, but if someone wants to help me, I suggest we check for a necessary adjustment based on RAM / runtime / # of threads

(Note! Before randomly running the program, 8 GB of free RAM is required, and the thread_count should be set to a power of 2 if the CPU has 6 or 12 cores! Otherwise this will not work as it is)

Thanks in advance!!!

Sequences-Backtracking.cpp

[Levent]

Free? nowhere. You can rent a VPS on digitalocean and get charged hourly.

[soccer1902]

I think you can get Microsoft azure with 200 dollar credit to start which might be enough for you

[Rybo]

AWS CPU instances are pretty cheap.  They have several 16 core/32 gigs of RAM instance types at around the 15 cents per hour price point.  Your script will have to be resilient to forced shutdowns because the pricing is for spot requests -- that is, "if there's space available, your instance will be turned on.  When that space is no longer available, your instance will be shut off."  (If you want a more permanent option, on demand requests are running 60 cents per hour).  

https://aws.amazon.com/ec2/instance-types/ (this lists the instance types; pricing on them is harder to get to.  You have to log into the EC2 console, go to the Spot Requests menu item, click on Pricing History, and then select the instance type you want to see the pricing on)

[Lii]

11 hours ago, Levent said:

Free? nowhere. You can rent a VPS on digitalocean and get charged hourly.

for free there is google colab, but that's for gpu stuff