Add a very basic SVG render tool to vello_toy
#840
Conversation
| #[arg(long)] | ||
| pub path: String, | ||
| /// The target runtime for calculating how many iterations to run, in milliseconds. | ||
| #[arg(long, default_value_t = 2000)] |
There was a problem hiding this comment.
I'd prefer for this to default to running a single iteration, i.e.:
| #[arg(long, default_value_t = 2000)] | |
| /// For simple benchmarking, our current recommended value is 2000 (i.e. 2 seconds). | |
| #[arg(long, default_value_t = 0)] |
I do agree that the benchmarking is a useful tool, but I don't think it should be the default.
(I can also see arguments for a benchmark flag, but that's probably too much complication for now)
| let avg_runtime = (runtime.as_millis() as f32) / (num_iters as f32); | ||
|
|
||
| eprintln!( | ||
| "ran {} iterations, with an average runtime of {}ms to render {}.", |
There was a problem hiding this comment.
Given that we run until the current item finishes, I think the priority should be on the average_runtime, rather than on the number of iterations.
|
|
||
| eprintln!( | ||
| "ran {} iterations, with an average runtime of {}ms to render {}.", | ||
| num_iters, avg_runtime, file_name |
There was a problem hiding this comment.
Would it also make sense to display the standard deviation? I think that's harder to calculate; we'd need to save all the results?
So probably not worth doing, then
|
I'll let this PR rest for a while until we have merged the other bigger, pending PRs. |
|
Is that referring to #831 (adding Vello Hybrid, which has now landed)? |
|
Yep! |
Example:
cargo run --bin svg --release -- --path "C:\Users\lstam\Documents\Programming\vello\examples\assets\Ghostscript_Tiger.svg" --scale 5Obviously far from feature complete, but my goal for now is to only implement very basic stuff.