Ask HN: Are any computer architecture out there to compete with VonNeumann's?

VonNewmann's architecture has been around for quite a while - and as a software engineer I am very grateful for that, don't get me wrong. On the other hand, new programming languages keep popping up every year - okay, I'm being a bit dramatic here.

Is there any interesting project or study in the computer architecture space you would recommend? Particularly one where it could eventually compete with VonNeumann's for the mainstream or more specifically to compete for personal computing in large scale?

My thoughts on this come from this paper - https://dl.acm.org/doi/10.1145/359576.359579 and this video from Bell Labs back in the day - https://youtu.be/tc4ROCJYbm0

Neither Harvard nor Von Neumann are described in a scalable way. Maybe a contentious opinion, but I don’t think either copes with multiple CPU and virtual memory.

https://en.wikipedia.org/wiki/Harvard_architecture

Security features are slowly transforming some CPUs away from Von Neumann and into Modified Harvard architecture

Take the x64 feature W^X . When enabled, you have 2 kinds of memory pages. Some contain code and some have program-created data. OS mechanisms allow for a controlled way to change this data into code by modifying page table flags. That’s a modified harvard architecture.

https://en.m.wikipedia.org/wiki/Modified_Harvard_architectur...

It has already been done, e.g. windows DEP. Main blocker is that software which generates code needs to be adapted to tell the OS where it put that code.

Attackers adapted by e.g.inventing Return Oriented Programming(ROP): As they could not inject code into processes directly, they started overwriting the stack with adresses to interesting functions. When an attacked function returns, it does not resume the caller, but executes code in the executable chosen by attackers, with args dictated by the attacker.