#milkymist IRC log for Wednesday, 2011-12-21

wpwrakhmm. seems that a short through USB can still upset M103:43
wpwrakhow many amps does the official supply have ? 2 A ?03:44
wpwrakhmm, 3.28 V on the 5 V rail03:48
wpwraklet's have it some more current03:48
wpwrakbah. still the same03:49
wpwrakso we're not short-circuit proof. we're probably still in good company with this03:51
wpwrakdoing the same thing into 2 Ohm barely troubles the 5 V rail. very good.03:57
rohhm.. you have some old spre notebook to test how that reacts?04:00
roh ;)04:00
wpwrakonly some that are already dead ;-)04:02
wpwrakhm. it's interesting that the switch permits those 2 Ohm, though.04:05
rohnice fet *g*04:05
rohhey adam... congrats on the 85 rc304:05
roheh. 8604:06
wpwrakif he goes on like this, he'll have a yield above 100% ;-)04:07
wpwrakah, the switch doesn't tolerate the abuse for long. after a while, it starts cutting the voltage04:10
cladamwroh, thanks ! today it's 87 now. :O04:26
wpwrakextrapolating, 90 for christmas. wolfgang will be so happy :)04:27
cladamwi try but not sure. :)04:33
lekernelkristianpaul: what's "satelite bolitha"? did you spell that correctly?10:51
kristianpaulbolinha sorry10:52
kristianpaul;(10:52
Action: kristianpaul got flu again10:52
Fallenouhttp://electronicdesign.com/article/embedded/How-To-Put-OpenCL-Into-An-FPGA.aspx13:18
lekernel"OpenCL kernel code is essentially C code"13:21
lekernelwell ... :)13:21
FallenouC to logic compiler will not be easy to implement I guess13:21
lekernelC is inappropriate for hardware accelerators13:22
lekernelin fact, Altera already has a C-to-HDL translator, but unsurprisingly enough it's a bit lousy13:22
Fallenou=)13:22
lekernelso, most of this article is probably marketing13:23
Fallenou14:25 < lekernel> C is inappropriate for hardware accelerators < yes but they don't have any other choice I guess13:23
Fallenouif they want to stay compatible with existing OpenCL13:23
Fallenouand use the same abstraction13:24
FallenouThey should have chosen something else in the first place13:24
Action: azonenberg actually thinks V*HDL suck too14:44
azonenbergThey dont allow enough programmer-guided optimization14:44
azonenbergfor example "place this next to that"14:44
azonenbergLava looks interesting but too hard to use14:44
wpwrakazonenberg: couldn't placement hints be added to the existing framework ?15:14
azonenbergNo idea, i havent looked into it15:14
azonenbergbut i want to explore EDA tools in more depth15:14
wpwrak(lava) ah, haskell :) the language of choice for people who prefer to be left alone15:16
azonenbergI dont like lava as-is15:16
azonenbergbut the premise is extremely interesting15:16
wpwrakhmm. for designing an synthesis system, i'd go for a languages that has good overall expressivity and then add hints. a bit like __attribute__ in gcc.15:23
azonenbergProbably15:24
GitHub68[migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/7456195775e9477a4eebfb3419f7f3b76fe97bd522:01
GitHub68[migen/master] Consistent names - Sebastien Bourdeauducq22:01
GitHub122[migen] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/migen/commit/26e0b817e8d491d709a8162eaf5276840d3edcc622:05
GitHub122[migen/master] verilog: ignore variable property in combinatorial block - Sebastien Bourdeauducq22:05
GitHub111[migen] sbourdeauducq pushed 2 new commits to master: https://github.com/milkymist/migen/compare/26e0b81...8a394f923:09
GitHub111[migen/master] verilog: break down Convert function - Sebastien Bourdeauducq23:09
GitHub111[migen/master] verilog: comb reset - Sebastien Bourdeauducq23:09
--- Thu Dec 22 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!