Luke Gorrie

@pygy I would be very interested in any citations in this area. I am hoping that the performance hit for Spectre will ultimately only be paid when switching between multiple...

My take is that interpreter performance is almost completely irrelevant. It's like swap memory on a Linux machine. Sure, it's nice that it exists as a fallback, but you would...

(Sorry to rant @pygy :-) I am just really looking forward to putting the complex interpreter code behind me and optimizing it for maintainability instead.)

I must confess that I am not tracking the Spectre and Meltdown issues in detail and so I could be missing something important. Just now my mental model is that...

Thanks for sharing @pygy. Sorry to fly off the handle whenever a discussion leads towards spending more time thinking about the interpreter code :). In the Snabb context if we...

Thank you for using a Pull Request to raise this issue. It is very helpful to have the exact code you are working with connected with the bug report. That...

@raj2569 Could you please try merging #1242 and post a dump? I hope this will make the JIT trace easier to read by breaking the traces up into smaller parts...

I am using this problem as a test case for improved profiler functions. Have updated the profiler backend in raptorjit/raptorjit#124 with information that should point to a solution, and now...

I wonder what the best solution is? Maybe to print a better error message and provide a way to select a more widely available counter e.g. `task-clock` for scheduled CPU...

If you do decide you want to share huge pages between processes then check out the trick in `memory.lua` that does this for DMA memory. The basic technique is to...