00:27 <MY123> rjo: sb0 is offline.

01:00 <rjo> MY123: i can feel his presence though. he usually reads the log.

04:37 <sb0> "To select other pins, consult the Implementation chapter of UG___, or rerun the wizard" - xilinx appnote

05:27 <rjo> sb0: (repost) https://github.com/nist-ionstorage/pdq2 if you want to add another project.

06:22 <sb0> great, thanks!

06:25 <sb0> btw DDR3 on the monster-board is mostly working (with help from Florent)

06:55 <rjo> sb0: i tried to follow that progress. is it more like DDR3-best or DDR3-minimal?

06:56 <sb0> we should have everything at 1Gbps/pin, ie 64Gbps of bandwidth. it's also monster-DRAM.

06:57 <sb0> there isn't really an alternative for the high data rate, as the DLL inside the DRAM chip has a narrow frequency range, and disabling it makes the SDRAM really slow.

06:57 <sb0> it's all or nothing

06:58 <sb0> we're just sorting out a few details now and implementing read/write leveling properly

12:12 <rjo> hmm. impressive.

15:09 <ysionneau> rjo: I just read http://tf.boulder.nist.gov/general/pdf/2668.pdf , it seems you already had some working system for the AWG, in VHDL. If I understood correctly you just rewrote the thing but in Migen this time?

15:11 <ysionneau> ( I'm talking about pdq2 )

15:12 <sb0> yay! broke python again

15:13 <sb0> this time trying to add special __operator__ methods at runtime using closures

15:13 <sb0> (the first time was trying to use classes with static methods as context managers)

15:25 <sb0> hmm, maybe not a bug, but weird behavior for sure

15:30 <sb0> http://pastebin.com/axcbZQv0

15:31 <sb0> though actually, what happens is that x is a global variable in the 1st one. so not a bug.

15:31 <ysionneau> ah ok

15:33 <ysionneau> easy to get it wrong

15:58 <GitHub41> [ARTIQ] sbourdeauducq pushed 1 new commit to master: http://git.io/mvNiMw

15:58 <GitHub41> ARTIQ/master 2e94104 Sebastien Bourdeauducq: language/core: add int64

16:01 <larsc> aren't you just overwriting the definition of printme in each loop iteration in the first example?

16:02 <sb0> larsc, no, but printme does a LOAD_GLOBAL when you run it to access x

16:02 <sb0> so if you run it after the loop has ended, you get the last value of x

16:02 <larsc> makes sense

16:04 <sb0> yeah. so does if(x = 0) in C ;)

16:05 <sb0> idk why Python doesn't make the loop iteration variables local

16:06 <MY123> sb0: I think that C is a better scripting language.

16:07 <MY123> http://bellard.org/tcc/

16:07 <ysionneau> ahah

16:08 <sb0> http://stackoverflow.com/questions/3611760/scoping-in-python-for-loops

16:21 <MY123> ysionneau: Using a JIT C compiler may be better than Python.

16:24 <ysionneau> sure but C language has much less syntaxic sugar than python

19:12 <sukanto> hey guys

19:12 <sukanto> I wanted to find out a bit more about the mlab project

19:12 <sukanto> as i want to do some research on it..

19:20 <MY123> sukanto: There is a small desc in the topic.

20:06 <rjo> ysionneau: re pdq2: yes. this is a complete rewrite with vastly more features and fewer bugs. if you know what kind of code first-year physics grad students who have never coded before come up with, you can imagine what that VHDL code looked like ;)

20:22 <ysionneau> wow ok :)

20:23 <ysionneau> pretty cool to see some serious project like this switching from vhdl to Migen :)

20:24 <ysionneau> next paper about it will mention Migen :D

20:24 <ysionneau> however I tried to click on some links on the article, I was asked to fire my credit card :/

20:37 <rjo> ysionneau: yeah. the usual applies if you can't proxy through your university of choice: our papers you can always get form http://www.tf.nist.gov/timefreq/general/publications.htm for free. most of the other papers have been posted as preprint to arxiv.org at some point in time.

20:40 <ysionneau> oh ok

20:40 <ysionneau> so with a bit of searching I could eventually get them

20:44 <rjo> ysionneau: most of the time in the "hard sciences" (phys, math) you can get a preprint that is very close to the actual published paper. http://adsabs.harvard.edu/index.html is pretty good at linking paper and preprint.

20:45 <ysionneau> awesome!

20:45 <ysionneau> thanks

20:45 <ysionneau> usually how do you handle the difference between those two (preprint and official paper)?

20:45 <ysionneau> what's the real difference?

20:48 <rjo> ysionneau: well. usually you submit the same thing to the journal and the preprint server at the same time. most people then even remember to keep the preprint updated with the peer review revisions. if they don't, that is usually rather embarrassing ;)

20:49 <ysionneau> so except if you forget to update the preprint after peer review, they end up identical ? :o

20:51 <rjo> ysionneau: yes. you can usually follow nicely how peer review forces authors to water down their claims, like in the case of http://arxiv.org/abs/1403.3985

20:51 <ysionneau> ah yes, submission history

20:52 <ysionneau> quantum cosmology :o

20:52 <ysionneau> I thought cosmology was usually studied using general relativity and not quantum mechanics

20:52 <rjo> ysionneau: or better yet, completely reversed: http://arxiv.org/abs/1109.4897

20:54 <rjo> ysionneau: QM and GR need to merge at some point.

20:54 <ysionneau> they need to

20:54 <ysionneau> but as far as I know they still totally don't, do they?

20:54 <ysionneau> that's the issue to understand what happens before the "planck wall", right?

20:54 <ysionneau> happened*

20:59 <rjo> yes. they don't.

20:59 <ysionneau> thanks for the intel about the publication system !

20:59 <ysionneau> see you!