Yea the Speed-record is what i am looking for actually.
I want to address alot of pins inside of my assembler part and the extended assembly just takes 10 commands and i only have 12 registers.
So i could only get the SODR/CODR/pinValues for about 3 pins in total like this.
And thats not enough for what i am planning to do.
EDIT:
I also read about parallel pin axcess somewhere, but is that only "pseudo parallel" like a one core cpu is kind of multitasking or is it really parallel?