# Simo & David
## DAY 1: 29 Oct - Jacobe optimization Serial
David and I have started working on the analysis of the parallelization of the Jacobi algorithm.
Our goal is to understand how to optimize this iterative method to solve systems of linear equations, exploiting the potential of parallel computing.

1) We have analyzed the problem.
2) Initialized the memory on the Host and Device
3) Created our first Kernel.

We are a bit confused having as never used CUDA and having as a background only worked with C with pthread libraries in the Operating Systems course

## DAY 2: 30 Oct - Jacobe optimization Parallel
### What we Achived:
- Finished the Jacobe Parallelization Problem
- Benchmark test with Leonardo
- An x11 Speedup with the GPU

## Our Conclusion:
We have noticed and understood that it's possible that the CPU can be faster than the GPU (only for small cases)

### Profilation of the Jacobe Parallel Code
Even though Jacobe it's an easy problem we decided to try to do it with NVIDIA Nsight System

### Little Ottimization
Using 2 streams instead of 1 we noticed an improvement, here the details:

## DAY 3: 31 Oct
Goal to Achieve:
- Plot the difference between **BEFORE & AFTER** the optimization ( did this morning in the car out here in the parking lot )

## Here the differences between the two codes:

- Searching for new Optimizations
- Resolve another Exercise