[Vm-dev] CogVM Execution Flow

[Levente Uzonyi]

That would make sense if the six entries were the among the most commonly 
used ones.
We had a discussion recently about adaptive PICs, which would reorder the 
entries based on some statistics, and the conclusion was that the overhead 
would be way higher than the potential benefit.
I did some measurements at that time about what the ideal size of a PIC 
would be. On my machine 12 had the same worst(?) case lookup time as the 
open PIC. Don't ask me how I measured it. :)

Levente

On Mon, 13 Jun 2016, tim Rowledge wrote:

>
>
>> On 13-06-2016, at 7:21 PM, Eliot Miranda <eliot.miranda at gmail.com> wrote:
>>  If ever it misses, the send site will get bound to a "closed" PIC, a little jump table created specific to that send site, that can hold up to 6 class index comparison, jump pairs (it can dispatch up to 6 classes of receiver) and if there are more than 6, will get rewritten to an "open" PIC specific to the selector, that probes the first level method lookup cache.
>
> Whilst reading that a strange thought leaped up from somewhere and landed in my head; might it be useful to just replace the final pic-miss jump with a jump instead to the open lookup but keep the other six previously carefully built cases? It’s not like they’ve gone away in any sense. Might need to change the use of the ClassReg there too? After all, the open lookup is really just a pic-miss that doesn’t build a new cpic entry...
>
>
> tim
> --
> tim Rowledge; tim at rowledge.org; http://www.rowledge.org/tim
> Useful random insult:- Thinks E=MC^2 is a rap star.
>
>
>