Almost 50 times slower in parallel processing?

Hey there

I know there’s been discussions about this and why parallel processing isn’t always the fastest, but I feel there might be something wrong here…
In this example it’s doing 50 pretty basic shape intersections.

I’ve also switched recently to AMD 7950x cpu (dunno if that could have anything to do with it).


lazy_parallel.gh (25.1 KB)

AMD Ryzen 9 5950X 16-Core Processor, 3401 MHz

The more AMD cores the slower it gets ! :man_mage:

1 Like

same with Intel Core i9-12900H.
Weird
image

Apple M1 Ultra (20 CPU cores, 3.2GHz)

Screenshot_1

Intel I5-12600K 3700 MHz 10 Cores

(MSInfo says “16 logical processors”, but I have no idea what that means.)

Higher the frequency, lower the single core time (ok)
Higher core count, higher multi core time (wait what?) (except apple, because apple)

Seems like intel > amd for multi core, but both intel have less cores so it might fall into the second statement.

Although one graph to rule them all (sorry apple, had to remove you)
(lower time the better, lower price the better)

image