In my institute, I run machine learning, deep learning, image processing, data mining and other algorithms of the same stature. The datasets are also quite large, say ranging from a few hundred thousand to a few hundred billion.
For reference of people who don't know, consider a excel table. Now for every row in the table, the columns have several values, and a bunch of mathematical operations need to be done on the column values for that row to get a result. Now this needs to be done multiple times for each row. Here each row is called a data point.
I am confused between which processor I should get. As for the GPU I'm getting a 1080. I will be coding things up in Python/MATLAB/R and running and analysing for datasets. No time time for gaming or overclocking. So which processor should I get -
Any other suggestions are also welcome.
For reference of people who don't know, consider a excel table. Now for every row in the table, the columns have several values, and a bunch of mathematical operations need to be done on the column values for that row to get a result. Now this needs to be done multiple times for each row. Here each row is called a data point.
I am confused between which processor I should get. As for the GPU I'm getting a 1080. I will be coding things up in Python/MATLAB/R and running and analysing for datasets. No time time for gaming or overclocking. So which processor should I get -
1700 - more cores/threads than intel, lower frequency than intel, good cooler inbox
1900x - highest cores/threads than intel/amd, very close frequency to intel, no cooler
7820x - same price as 1900x, but as per amd 1900x is better
7700 - less core/threads than amd, higher frequency than amd, low level cooler inbox
7700k - less core/threads than amd, higher frequency than amd or 7700, no cooler inbox
Any other suggestions are also welcome.