IRC logs for #openrisc Tuesday, 2017-01-03

--- Log opened Tue Jan 03 00:00:56 2017
promachwallento: for lisnoc tutorial 1, why is the 'READY' signal high only after t=19ns as in http://i.imgur.com/jjLcfXO.jpg ?00:46
wallentopromach, hi06:54
wallentolet me check06:54
wallentoits hardcoded06:57
wallentothe sink doesn't transition states until the reset is deasserted, then it waits six cycles and the combinational logic uses the negative edge, that brings us to 19ns06:58
wallentohttps://github.com/TUM-LIS/lisnoc/blob/master/tutorials/tutorial_01/tutorial_01.v#L156 and https://github.com/TUM-LIS/lisnoc/blob/master/tutorials/tutorial_01/tutorial_01.v#L18906:58
wallentonot the most awesome coding style, but I wanted to keep it simple at that time06:58
wallentoand great people use it. Motivates my for lesson 3 :)06:58
promachwallento: I am a bit confused on the line      localparam FLIT_WIDTH = FLIT_DATA_WIDTH + FLIT_TYPE_WIDTH;07:17
wallentoit defines the total flit width, which is data plus the control signals07:18
wallentoso the actual flit has the width FLIT_DATA_WIDTH07:18
wallentoand there are at least two extra bits with each flit: first and last07:18
-!- Netsplit *.net <-> *.split quits: LoneTech07:19
wallentoso a worm looks like (1,0,D0) (0,0,D1) (0,1,D2)07:19
-!- Netsplit over, joins: LoneTech07:19
promachwallento: I do not get the purpose of 'worm'07:22
wallentoIn a NoC a worm is an indivisible unit07:23
wallentocomposed of multiple flits07:23
wallentoit is necessary that worm don't interleave on a link because otherwise your data gets corrupted07:23
wallentoso while there can be multiple worms in the NoC, they can only enter and leave the NoC serialized07:23
-!- Netsplit *.net <-> *.split quits: jonmasters, mithro, Neuron1k14:35
-!- Netsplit over, joins: jonmasters14:36
wbxshorne: hmm. is gdb or gdbserver for openrisc linux possible to cross-compile?18:57
--- Log closed Wed Jan 04 00:00:58 2017

Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!