Dynamic Kernel Stacks Proposed For Linux With Big Memory Savings

[Brane215]

There's much less copying of graphics buffers between devices on Chromebooks with Intel GPUs, which means you don't need to reserve memory in both the source and the destination, so you wind up needing half the memory compared to Linux on the same hardware: https://medium.com/joone/using-the-c...s-38f56f2ed6e4

I don't get it. One of main points of AMD's iGPUs for quite some time was unified memory space that both /i/d/GPU and PCU can access.
So, what makes Intel uniquely better here ?🙄

[jokeyrhyme]

I don't get it. One of main points of AMD's iGPUs for quite some time was unified memory space that both /i/d/GPU and PCU can access.
So, what makes Intel uniquely better here ?🙄

AMD Chromebooks are much younger than Intel ones, so perhaps it is indeed something that is physically possible and just needs some optimisation work? Maybe Intel has donated engineer time, or was prioritised because it impacts so many more users than any AMD work? Or perhaps the memory saving are already active on AMD Chromebooks and I've just not heard about it?

[LtdJorge]

This feature seem to benefit those programming languages that lack support for lightweight (a.k.a. green) threads (as supported in Haskell,Erlang), but have to rely instead on kernel (heavyweight) threads (e.g. Java). (EDIT: Java has support now for lightweight threads, see Amaranth 's reply)

Erlang's processes are a league above any other green thread implementation (which are coroutines), because they are preemptive. In Erlang you write blocking code and the scheduler gives every process a fair slice of time, no matter for how long the process is blocked.