This project demonstrates the use of WaitGroups and goroutines to divide and conquer the rendering of a single ray traced frame. By changing the the number of sections in use by the algorithm, we see various performance characteristics as listed here:
| # Sections | CPU Load | #Threads | Time (ms) |
|---|---|---|---|
| 1 | 100% | 9 | ~ 524 |
| 2 | 198% | 9 | ~ 275 |
| 5 | 337% | 9 | ~ 145 |
| 10 | 571% | 9 | ~ 145 |
| 20 | 617% | 9 | ~ 129 |
| 50 | 733% | 9 | ~ 127 |
In the above, each frame is rendered 200 times and the average frame render time is recorded.