Add CanvasItem drawing benchmarks

[OverloadedOrama]

Adds the following benchmarks from #36:

• 🟥CPU🟥 [CanvasItem]: Rendering: Draw different shapes (images, circles, etc) using the CanvasItem 2D drawing API. Draw 5000 elements. Measure performance.
• 🟥CPU🟥 [CanvasItem]: Re-Rendering: Same as above, but every frame call queue_update() so it redraws. Measure performance.

I am quite puzzled as to why drawing the elements once sometimes results in more CPU & GPU time, than re-rendering them every frame. I added benchmarks for 10000 and 20000 elements in case 5000 is too low and it's being affected by an error margin, but I still get weird results for these benchmarks as well. Not sure if I'm doing something wrong, the logic seems fine to me, but let me know if I should keep or remove the benchmarks for 10000 & 20000 elements.

Results on my PC

{
"benchmarks": [
{
"category": "Rendering > Canvas Item",
"name": "Draw 10 000 Shapes Every Frame",
"results": {
"render_cpu": 2.542,
"render_gpu": 4.311,
"time": 0.029
}
},
{
"category": "Rendering > Canvas Item",
"name": "Draw 10 000 Shapes Once",
"results": {
"render_cpu": 2.563,
"render_gpu": 4.435,
"time": 0.035
}
},
{
"category": "Rendering > Canvas Item",
"name": "Draw 20 000 Shapes Every Frame",
"results": {
"render_cpu": 5.243,
"render_gpu": 8.6,
"time": 0.023
}
},
{
"category": "Rendering > Canvas Item",
"name": "Draw 20 000 Shapes Once",
"results": {
"render_cpu": 5.02,
"render_gpu": 8.739,
"time": 0.049
}
},
{
"category": "Rendering > Canvas Item",
"name": "Draw 5000 Shapes Every Frame",
"results": {
"render_cpu": 1.216,
"render_gpu": 2.162,
"time": 0.041
}
},
{
"category": "Rendering > Canvas Item",
"name": "Draw 5000 Shapes Once",
"results": {
"render_cpu": 1.336,
"render_gpu": 2.21,
"time": 0.1
}
}
],
"engine": {
"version": "v4.3.beta1.official",
"version_hash": "a4f2ea91a1bd18f70a43ff4c1377db49b56bc3f0"
},
"system": {
"cpu_architecture": "x86_64",
"cpu_count": 12,
"cpu_name": "AMD Ryzen 5 1600 Six-Core Processor",
"os": "Linux"
}
}

[Calinou]

I am quite puzzled as to why drawing the elements once sometimes results in more CPU & GPU time, than re-rendering them every frame. I added benchmarks for 10000 and 20000 elements in case 5000 is too low and it's being affected by an error margin, but I still get weird results for these benchmarks as well. Not sure if I'm doing something wrong, the logic seems fine to me, but let me know if I should keep or remove the benchmarks for 10000 & 20000 elements.

The results I get are logical on my end:

Running benchmark 1 of 6: rendering/canvas_item/draw_10_000_shapes_every_frame
Result: {"render_cpu":0.7615,"render_gpu":1.54,"time":0.012}

Running benchmark 2 of 6: rendering/canvas_item/draw_10_000_shapes_once
Result: {"render_cpu":0.6177,"render_gpu":1.302,"time":0.013}

Running benchmark 3 of 6: rendering/canvas_item/draw_20_000_shapes_every_frame
Result: {"render_cpu":2.107,"render_gpu":3.457,"time":0.014}

Running benchmark 4 of 6: rendering/canvas_item/draw_20_000_shapes_once
Result: {"render_cpu":1.688,"render_gpu":2.677,"time":0.03}

Running benchmark 5 of 6: rendering/canvas_item/draw_5000_shapes_every_frame
Result: {"render_cpu":0.3111,"render_gpu":0.8509,"time":0.016}

Running benchmark 6 of 6: rendering/canvas_item/draw_5000_shapes_once
Result: {"render_cpu":0.2862,"render_gpu":0.6509,"time":0.014}

PC specifications

• CPU: Intel Core i9-13900K
• GPU: NVIDIA GeForce RTX 4090
• RAM: 64 GB (2×32 GB DDR5-5800 C30)
• SSD: Solidigm P44 Pro 2 TB
• OS: Linux (Fedora 39)

[Calinou]

Thanks!