03:39 <sb0> this taobao store http://world.taobao.com/item/523122559768.htm?fromSite=main&spm=a312a.7700824.w4002-9239694932.14.3E5k6a sells stuff for depositing superconducting thin films ?!?

03:40 <sb0> that's the only application i find for NdGaO3

03:43 <sb0> whitequark, did you get your helium cryocooler to work?

07:21 <sb0> actually it seems you can make a josephson junction with a film of tin (easy to evaporate and nontoxic), oxidize it, and redeposit another layer of tin. you just need to cool it below 3.7K, but liquid helium under vacuum can do it

07:22 <sb0> lead works as well but is toxic

08:38 <cr1901_modern> Making a Josephson Junction at home? o.0;

10:15 <sb0> taobao also sells DIY STM kits: http://world.taobao.com/item/44355819555.htm

10:28 <cr1901_modern> STM? (Google translate is failing me miserably)

10:28 <cr1901_modern> http://world.taobao.com/item/41305775815.htm?spm=a312a.7728556.1414651174895.5.jfA9vz&scm=1007.10146.12547.0&id=41305775815&pvid=b5c0ea95-7f43-4867-856b-f6510a4a9154 This also looks cute. Too bad I have no idea what the description is saying.

10:32 <sb0> scanning tunnelling microscope

10:33 <cr1901_modern> That was my guess lol

10:33 <whitequark> sb0: my cryocooler only goes down to 20K

10:34 <sb0> should be good for high temperature superconductors, but the films and barriers are probably a lot more painful

10:36 <sb0> you do need that NdGaO3 stuff, combine it with ytterbium in some way (there is little info I can find about that), and for the barrier - I have no idea

10:37 <whitequark> btw I did not yet bring it up, it needs some fittings...

10:37 <whitequark> well, I think this can be a reasonable thing to try and manufacture in HK, they're very simple

13:51 <larsc> has anybody worked with rifaa? The license information is somewhat confusing on the website the state that you need to get a permission for commercial usage, but in the source itself the only license terms ever mentioned seem to be standard BSD

13:52 <larsc> riffa

13:55 <sb0> larsc, have you tried litepcie?

13:55 <larsc> no

13:55 <larsc> I'm in the research phase at the moment

13:56 <larsc> does it have a Windows driver?

14:02 <sb0> I'm not sure...

14:03 <sb0> whitequark, should we keep dependency_links in setup.py?

14:05 <cr1901_modern> larsc: litepcie?

14:05 <larsc> yes

14:06 <cr1901_modern> AFAIK it only has a Linux driver. I asked _florent_ about the feasibility of adding a BSD driver (Idk enough about Windows internals to add a Windows driver)

14:07 <larsc> the windows driver is unfortunately the important component for my project

14:07 <cr1901_modern> I really should get another laptop so I can write Windows drivers without BSODing my mainlaptop

14:09 <sb0> use a vm?

14:10 <sb0> hm, wait, pcie won't work perhaps

14:13 <cr1901_modern> VirtualBox has USB passthrough; I would suspect that PCIe emulation would be minimal; just enough to support, say, a serial port/classic PC-compatible peripherals on the PCIe bus. But I'm not sure.

14:14 <cr1901_modern> (Idk if on-board USB chipsets are typically accessed through a PCIe controller or not)

14:14 <larsc> I think kvm supports PCI(e) passthrough

14:16 <cr1901_modern> That would be good if I were running Linux lol

15:05 <whitequark> sb0: what the hell is dependency_links?

15:09 <sb0> I guess something that lets setup.py download and install dependencies?

15:13 <sb0> ysionneau, why did you add dependency_links?

15:14 <ysionneau> to tell setup.py where to fetch the packages

15:15 <ysionneau> https://pythonhosted.org/setuptools/setuptools.html ctrl+F dependency_links

17:55 <whitequark> ohh, links as in URLs

17:56 <whitequark> cr1901_modern: PC-compatible peripherals aren't on the PCI-e bus

17:56 <whitequark> besides, from the point of view of the OS, PCI and PCI-e look exactly identical

17:56 <whitequark> and no, vbox doesn't have PCI passthrough

17:56 <whitequark> KVM and Xen do

18:55 <larsc> ~.

18:55 <larsc> ignore that obviously

19:16 <cr1901_modern> whitequark: There exist e.g. PCIe Serial port cards; that was what I was thinking about. I don't think they would be addressed the same way as the ones that would be available through the SuperIO chip/LPC bus/whatever it's called.

19:23 <whitequark> virtualbox does not emulate those

21:40 <GitHub8> [migen] enjoy-digital pushed 1 new commit to master: http://git.io/v8jnz

21:40 <GitHub8> migen/master 598638b Florent Kermarrec: fhdl/structure: fix last test in _Value.__bool__ (a instead of b)

22:03 <robtaylor> whitequark: if i recall correctly, that's not entirely true - istr that MSIs are different at the OS level

22:04 <robtaylor> (MSI-X)

22:05 <whitequark> robtaylor: hm right

22:06 <robtaylor> PCIe can bu *used* by PCIe-unaware systems, but you get more goodies if you understing PCIe :)

22:44 <cr1901_modern> How can you use PCIe by PCIe-unaware systems? Wouldn't there need to be a PCI=>PCIe bridge so PCIe cards know how to respond to requests?

23:10 <felix_> from the hosts point of view pcie is a superset of pci; the different physical layer doesn't matter for the logical interface

23:11 <felix_> if the system isn't pcie aware, it can't use the new features though

23:18 <felix_> a pcie serial port card could be addressed the same way as a legacy serial port on the superio when it uses a legacy io bar as interface; the pcie to serial port card i had, used mmio bars though