| --- Log opened Tue Aug 23 00:00:38 2016 | ||
| SMDhome1 | Can someone clarify what does "branch stall" means from openrisc arch document, page 309? Is it brn miss count or clocks needed to flush pipeline because of brn miss? | 11:50 |
|---|---|---|
| ZipCPU | Is there an implementation that can be checked to know? If the document is the ultimate definition, I would argue the latter. | 12:31 |
| ZipCPU | That is, the number of clocks that do not complete instructions as a result of a branch. | 12:31 |
| ZipCPU | I.e. number of branch misses * clocks needed to flush pipeline due to a miss. | 12:31 |
| ZipCPU | But ... that's only if the document is the controlling definition. | 12:32 |
| SMDhome1 | ZipCPU I think there's only that document. I've done w/ perfomance counter unit and now I have to wire all the events that need to be counted. | 12:38 |
| olofk | SMDhome1: Some of these things that has never been implemented could be a bit under specified | 17:40 |
| olofk | I'd say you just do what you think is right, and then we can discuss it | 17:40 |
| ZipCPU | olofk: I'm trying to bite the bullet today and actually download OR1K to build it. <Gulp> | 17:47 |
| ZipCPU | So far, I've been pulling project pieces from github.com/openrisc ... | 17:47 |
| ZipCPU | binutils-gdb, or1k-gcc, mor1kx, ... I'll try to grab newlib next. | 17:48 |
| Finde | https://techxplore.com/news/2016-08-microchip-efficiency-scalable.html | 17:50 |
| Finde | we're finally able to show off some die pics of the Piton chip | 17:50 |
| Finde | :D | 17:50 |
| kc5tja | Finde: Congradulations. :) | 20:35 |
| ZipCPU | In case you missed it, the opencores.com domain has been bought and sold to someone else. | 20:46 |
| kc5tja | Or the DNS record has elapsed, and is now officially DNS squatter property. | 23:47 |
| --- Log closed Wed Aug 24 00:00:39 2016 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!