| --- Log opened Sat Jun 17 00:00:33 2017 | ||
| promach | What is the purpose of the busy signal here at http://i.imgur.com/SaIqMI2.png ? this is about reorder buffer mechanism | 04:42 |
|---|---|---|
| promach | Entry 1 -> 2: BUSY = no | 04:42 |
| promach | Entry 3 ->10: BUSY = yes | 04:42 |
| promach | Why ? | 04:42 |
| -!- wbx_ is now known as wbx | 07:25 | |
| shorne | promach: I am not sure, I havent really looked at fpu before | 20:08 |
| shorne | bandvig I think is the expert, it seems he is away now | 20:09 |
| shorne | promach: oh... this is not from openrisc spec , this is from "Computer Architecture: A Quantitative Approach" | 20:11 |
| shorne | there is a long explaination of this reorder buffer here... | 21:18 |
| shorne | https://cseweb.ucsd.edu/classes/fa14/cse240A-a/pdf/07/CSE240A-MBT-L13-ReorderBuffer.ppt.pdf | 21:18 |
| shorne | Interestingly they dont mention the busy status | 21:18 |
| shorne | promach: I think here busy=yes means instruction in the reorder buffer slot is not executed | 21:23 |
| shorne | or not committed | 21:23 |
| shorne | Down below in the pseudo code it shows as 'ROB[b].Ready <- No' | 21:25 |
| shorne | Also, the comment says first 2 instructions are committed | 21:27 |
| shorne | (but shown for completeness) meaning those reorder buffer slots are actually empty | 21:27 |
| --- Log closed Sun Jun 18 00:00:35 2017 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!
