slow perf? about glamor-jss

[dan-lee]

Oh that does not look good indeed. I figured that all my previous performance tests were cached quite heavily (by not providing a incrementing counter in the style declarations).
Looks like I need to find the bottle necks and some improvements. I hope JSS isn't the cause.

[jeremy-coleman]

What counts as an op/s though? Like glamor and jss both use caching so if they do a comparison and use the cached version, will that show as an op? 100 vs 7k seems like maybe the bench isnt comparing apples to apples

[kof]

I have strong feeling that its the hashing. Also 6k vs 13k is not crazy slow, it is only 2x slower on a micro benchmark, which could be already a result of a single small suboptimal call.

[jeremy-coleman]

140 vs 7k is though, but i also read an isse about instantiating jss , but im not sure how that accounts for glamor as well

[kof]

if glamor scores 100 vs 7k emotion on the same benchmark, something is wrong with the benchmark, it does something naively wrong that deoptimizes everything.

[dan-lee]

I removed the performance benchmarks now, as @kof said I think they were flawed and not well thought through. I might take another stab at them