#milkymist IRC log for Thursday, 2011-03-03

lekernelhmm, if this isn't all bullshit, lm32 should run at 166MHz in kintex00:00
lekernelit's faster than v6, so that's rather surprising00:00
Fallenouwith the newer ISE you could try to synthetize it00:02
lekernelyeah, that's where the 166MHz figure comes from00:03
Fallenouoh !00:06
Fallenounice nice00:06
kristianpaulmwalle: woah, nice scope02:02
Action: kristianpaul waiting for xilinx/avnet introduce dirty cheap kintex kit02:02
kristianpaulfunny interactive banner sebastien put on the top right place of milkymist.org ;-)02:47
kristianpaul"Milkymist One can run software written in Ruby and Lua." ah nice ;-)02:48
wpwrakkristianpaul: a better one would be simply "Milkymist can". or maybe "Milkymist can do things your Arduino doesn't even dare to dream of" ;-)04:13
GurtyHappy Birthday Fallenou <3 *\o/*07:23
scrts`congratz :)07:26
xiangfuNot Found08:05
xiangfuThe requested URL /doc/system.pdf was not found on this server.08:05
xiangfureport a issue . :)08:05
xiangfumaybe it should point to : http://www.milkymist.org/socdoc/bios.pdf08:12
kristianpaulFallenou: Feliz Cumpleaños :-)10:43
Fallenouoh :)11:00
Fallenouthanks !11:00
scrts`Fallenou: how much?11:34
Fallenouscrts`: 23 :)11:55
GurtyFallenou :D12:01
FallenouGurty: hey :) thanks for the text message !12:03
scrts`same for me tomorrow12:05
GurtyIt's always a pleasure Fallenou ;)12:06
Fallenouoh 23 too tomorrow ?12:06
GurtyFallenou, A happy birthday fkosmala (kostek) wishes you12:12
Fallenouthanks !12:13
methril_worklekernel, what is missing from the driver side?18:20
methril_workinto RTEMS18:21
lekernelethernet fixes/rewrites18:21
Fallenoufile manager18:21
lekernelthat's not rtems18:21
Fallenouoh ok18:21
lekernelbut yes, gui filemanager18:22
methril_workand in the linux side?18:24
methril_worki know there is no much interest in the Linux side18:24
methril_worki think that Linux word attract developers18:25
lekernelabout every driver is missing i'd say18:27
lekernelor at least requires serious cleanup18:27
methril_workif it`s a long term goal, could i start to hack it in qemu?18:28
lekernelsure, go ahead18:30
lekernelis there anything that works on spartan6?18:52
lekernelthe funny thing is we don't seem to have run into this problem18:52
mwalleFallenou: happy birthday :)19:28
Fallenouthanks mwalle :)19:29
methril_workFallenou, today is your birthday? Happy Birthday then!!19:29
methril_workmine is 13 days left19:29
Fallenouhéhé thank you all :)19:30
Fallenoua lot of people are born in march i guess :p19:31
lekernelFallenou: happy bday :)19:32
lekernelhey, with ise 13.1 we run at 90MHz now19:32
Fallenouhey nice !19:32
lekernelshould check the stability though :)19:33
Fallenouthat's the birthday gift :)19:33
lekernelwell, let's hope it's not poisoned with heisenbugs19:33
Fallenouit was running at 88 MHz before, right ? and then you switched to 8019:34
Fallenouor something like that19:34
lekernel83 and then 8019:34
lekernelnow 9019:34
Fallenouwhy the drop from 83 to 80 ?19:35
Fallenouto meet timings ?19:35
lekernelbecause it wasn't possible anymore to meet timing at 8319:35
Fallenouoh ok19:35
Fallenouyou just gave it a try blindly or ise told you "you can run at 90 MHz" ?19:36
lekernelwell I saw that it was easily meeting timing at 80 (it was a total mess before), so I pushed it a bit19:36
Fallenouok :)19:36
Fallenouat least they improve Xst a little bit each time19:36
Fallenouthey don't just change the GUI19:36
lekernelseems too good to be true imo19:36
lekerneli expect a few surprises19:37
Fallenouhehe we'll see19:37
lekernelhmm... there are some slight "dancing pixels" on the screen19:37
lekernelnot sure where this comes from19:38
mwallelekernel: is this with debug enabled? (except the breakpoints?)19:53
lekernelmwalle: no, no debug at all20:01
kristianpaulha thats probably what larsc pointed time ago in qi http://www.xilinx.com/technology/roadmap/zynq7000/features.htm20:14
scrtsI wonder what artix and kintex devices will be :)20:28
lekernelprobably even buggier than s620:46
lekernelit's amazing how crippled the s6 is... silicon bugs with BRAM asynchronous clocking, silicon bugs with BRAM initialization, and others, and they said they will NOT fix them20:47
lekernelgiven how rushed up the 7 series seems to be, i expect the worst20:48
lekerneli'd much prefer to see a "s6-stable" device than a freaky "k7"20:48
kristianpaulbram > buffer > fixo rx overflow  !! (That hapen just when doing pld load of the milkymist soc bitstream)20:49
lekerneloh, and no characterization of the IODELAY timing too... "let's shift the 45nm process problems to our customers"... that's why I had lots of intermittent DRAM problems on the M120:50
lekernelterpstra: don't Altera multiplier blocks have internal pipeline registers?21:10
lekernelit seems weird to me that you have to split the operation manually21:10
methril_worklekernel, when you work in the USB part, do you thinks some usuall mistakes  that i could translate to CAN Bus?21:14
lekernelhmm as expected it's unstable at 90M22:32
lekerneli'd have been surprised otherwise22:33
lekernelstupid xilinx crap22:35
CIA-94milkymist: Sebastien Bourdeauducq master * rf731094 / (5 files in 4 dirs): LM32 GSI patches - http://bit.ly/dEEqtc22:41
CIA-94milkymist: Sebastien Bourdeauducq master * r5fc9102 / boards/milkymist-one/synthesis/common.ucf : ISE 13.1 support - http://bit.ly/ffruCr22:41
CIA-94milkymist: Sebastien Bourdeauducq master * r9679c6d / build_bitstream.sh : Build host tools for bitstream (needs byteswap) - http://bit.ly/fFb81x22:41
CIA-94milkymist: Sebastien Bourdeauducq master * r4c314c5 / (2 files in 2 dirs): Bump SoC version number - http://bit.ly/hnY3Iv22:41
CIA-94milkymist: Sebastien Bourdeauducq master * r5366fec / software/bios/main.c : Use timer for boot timeout - http://bit.ly/houF3722:41
lekernelseems they can't characterize a 45nm chip (IODELAY data mess + bugs in the timing analyzer) so those 28nm ones will be a disaster22:42
Fallenouis it so difficult to get the IO timing ?22:43
lekernelthe I/O thing I'm talking about is:22:44
lekernelin S6 they introduced a new low-cost way to delay all incoming signals on every I/O pin22:45
lekernelon previous fpgas, it was done using a delay chain with a variable tap, and it included some calibration circuitry that worked great22:45
lekernelit was only available in virtex though22:46
lekernelin S6, they replaced that with a multi-GHz ring oscillator that drives programmable counters that generate the delayed signal when they overflow22:46
Fallenouhum ok22:47
Fallenousounds great too22:47
lekernelthe problem is the frequency of said ring oscillator is totally unstable22:47
lekerneland your delays vary by a ridiculous margin, some 1000%  (can't remember the exact figure from the datasheet, but it's a total mess)22:47
lekerneland you can't easily calibrate them22:48
lekernelthey removed the previously available calibration logic that compared the delays against a reference clock22:50
lekerneland allowed you to have precise delays without much of a mess22:50
Fallenouthe  | IODELAY thing ?22:50
Fallenouin the ucf22:51
lekerneloh, funny, it has yet more silicon bugs too: http://www.xilinx.com/support/answers/38408.htm22:51
lekernelI think my next FPGA design will have an Altera chip22:52
lekernelFallenou: it's not in the UCF, IODELAY2 is a logic primitive you instantiate in the signal path to (hopefully) implement the programmable delay22:56
Fallenouoh ok22:56
FallenouI remembered something about "delay" in the ucf22:56
Fallenouthat you can put on the pins of the fpga22:57
Fallenouoh ok it's the SLEW23:00
jackgassetthello, I was just checking out the Softusb project and was wondering what was used for synthesis.23:48
jackgassettI tried with Xilinx ISE and it gives some errors.23:48
lekernelphew, LZMA decompression is slow as hell... boot time went from 9s to 43s23:53
lekerneljackgassett: well right now there aren't that many options. we use ISE too23:54
lekernelsoftusb cannot be used stand alone23:54
lekernelI don't know what you tried to do...23:54
jackgassettok, so it probably just needs some love then.23:54
jackgassettI just wanted to see if it should work with ISE.23:54
lekernelif you just take the contents of the softusb/rtl folder and drop that into an ISE project, there's no chance it would work23:55
jackgassettok, so it looks like memory needs to be added through wishbone bus then?23:55
lekernelmemory is integrated23:56
jackgassettand then peripherals can be added with your csr bus?23:56
lekernelwell, have a look at the ports on the top level23:56
jackgassettok, right I see that memory is included, Data and Program.23:57
jackgassettI commented out the sie and hostif modules because I just wanted to check out navre and wishbone controller for now.23:57
CIA-94milkymist: Sebastien Bourdeauducq master * r5b47999 / (5 files in 2 dirs): Flash boot image LZMA compression support - http://bit.ly/f1MNkB23:57
jackgassettshould the softusb core work for that purpose without much modification?23:59
--- Fri Mar 4 201100:00

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