00:59
tangrs has left #m-labs ["Leaving"]
01:01
d_n|a has quit [Ping timeout: 248 seconds]
05:56
<
_florent_ >
rjo: thanks for the cordic, I'll integrate that today
06:14
FabM has joined #m-labs
06:40
<
sb0 >
bb-m-labs_, force build --branch=rtio-sed artiq
06:40
<
bb-m-labs_ >
build forced [ETA 34m35s]
06:40
<
bb-m-labs_ >
I'll give a shout when the build finishes
06:40
<
GitHub199 >
artiq/rtio-sed 5437f0e Sebastien Bourdeauducq: rtio: make sequence errors consistently asychronous
06:51
<
rjo >
sb0: do you want to check the mods from sinara#327?
06:52
<
rjo >
sb0: yes. imho we can forego the RCs. feel free to go ahead and release 3.0
06:52
<
sb0 >
well it didn't work on the board we had in the US either
06:53
<
sb0 >
the last comments could explain why the media converter doesn't find a link with the HK boards though. i'll take a look.
06:55
<
sb0 >
hmm but didn't Greg say this chip was doing rate conversion?
06:55
<
sb0 >
with the SFP side always at 1G
07:07
<
sb0 >
also my instruction for ethernet was "copy the kc705 schematics". if Joe and Greg had followed this, and disconnected the mmc, this mess would not exist.
07:08
<
sb0 >
and I suggested that after Joe said he did not want to pay for more Ethernet work. we should also have him revise that position.
07:22
<
sb0 >
okay, I'll finish some rtio work (report channel number on async error) then release 3.0 with or without #830
07:22
<
rjo >
the kc705 schematics would not have worked with the crate backplane. that's why the chip was changed. iirc you were there when that was discussed.
07:24
<
sb0 >
I wasn't happy about in-crate ethernet either
08:33
<
sb0 >
bb-m-labs_, force build --branch=rtio-sed artiq
08:33
<
bb-m-labs_ >
build forced [ETA 34m35s]
08:33
<
bb-m-labs_ >
I'll give a shout when the build finishes
08:33
<
GitHub16 >
artiq/rtio-sed 6c049ad Sebastien Bourdeauducq: rtio: report channel numbers in asynchronous errors
08:36
<
sb0 >
bb-m-labs_, force build migen
08:36
<
bb-m-labs_ >
build forced [ETA 3m26s]
08:36
<
bb-m-labs_ >
I'll give a shout when the build finishes
08:36
<
sb0 >
bb-m-labs_, force build misoc
08:36
<
bb-m-labs_ >
The build has been queued, I'll give a shout when it starts
08:39
<
bb-m-labs_ >
build forced [ETA 2m51s]
08:39
<
bb-m-labs_ >
I'll give a shout when the build finishes
08:41
<
GitHub37 >
artiq/master 6d72329 Sebastien Bourdeauducq: RELEASE_NOTES: 3.0
08:41
<
GitHub37 >
artiq/master 071439f Sebastien Bourdeauducq: conda: use migen/misoc releases
08:43
<
sb0 >
bb-m-labs_, force build misoc
08:43
<
bb-m-labs_ >
build forced [ETA 2m51s]
08:43
<
bb-m-labs_ >
I'll give a shout when the build finishes
08:53
<
GitHub >
conda-recipes/master 44bda1d Sébastien Bourdeauducq: llvmlite-artiq: remove .dev from version number...
08:53
<
sb0 >
bb-m-labs: force build --props=package=llvmlite-artiq conda-all
08:54
<
sb0 >
rjo, should artiq 3.0 use jesd204b 0.3 or 0.4?
08:59
<
rjo >
sb0: 0.3 is fine.
08:59
<
sb0 >
bb-m-labs, force build --props=package=llvmlite-artiq conda-all
08:59
<
sb0 >
bb-m-labs_, force build --props=package=llvmlite-artiq conda-all
08:59
<
bb-m-labs_ >
build #89 forced
08:59
<
bb-m-labs_ >
I'll give a shout when the build finishes
09:01
<
sb0 >
is there any release that doesn't involve working around a dozen conda bugs?
09:07
<
sb0 >
ah wow, conda on win32 is really fucked, I cannot even run the activate script
09:10
<
sb0 >
bb-m-labs_, force build --props=package=llvmlite-artiq conda-win64
09:10
<
bb-m-labs_ >
build #178 forced
09:10
<
bb-m-labs_ >
I'll give a shout when the build finishes
09:31
<
sb0 >
who is still using win32?
09:32
<
sb0 >
bb-m-labs_, force build --props=package=llvmlite-artiq conda-win64
09:32
<
bb-m-labs_ >
The build has been queued, I'll give a shout when it starts
09:39
<
sb0 >
well basically it seems that python 3.5 with conda is hosed on win32
09:43
<
bb-m-labs_ >
build #179 forced
09:43
<
bb-m-labs_ >
I'll give a shout when the build finishes
09:47
<
rjo >
sb0: that was only the old nist machines due to a compat issue. ping daniel and then stop shipping 32 bit precompiled packages.
10:53
<
rjo >
_florent_: thenks the updates. if you have given it a quick test, could you drill greg and instruct him how to use it? probably good to note that this thing runs at 125 MHz, un-interpolated, and that max amplitude is 0x4db9.
10:54
<
rjo >
_florent_: from the code is looks like there is only two-fold sample repotition, i.e. 250MHz sample rate. is that correct? aren't you using 10 Gbit/s, 1000 MHz sample rate?
10:57
<
rjo >
_florent_: thanks for posting the bitstreams. i'll write that instruction. but what's the actual fabric frequency of the cordic and why does it not repeat each sample 1000 MHz/125 MHz = 8 times?
11:31
attie has quit [Remote host closed the connection]
11:36
attie has joined #m-labs
12:54
d_n|a has joined #m-labs
13:08
<
d_n|a >
whitequark/sb0: What's the holdup with #830?
13:08
<
sb0 >
whitequark, ^
13:10
<
GitHub196 >
artiq/master 74dbb4c Sebastien Bourdeauducq: conda: use non-dev llvmlite-artiq
13:18
<
d_n|a >
(I.e. just generate the obvious code using a conditional for now, and worry about optimizing it beyond what LLVM does later)
13:48
d_n|a has quit [Ping timeout: 240 seconds]
14:44
<
sb0 >
bb-m-labs: force build --props=package=artiq-kc705-nist_qc2 artiq-board
14:47
halfr is now known as halfr_
14:48
halfr_ is now known as halfr1
14:54
<
GitHub13 >
artiq/release-3 ac28b37 Sebastien Bourdeauducq: targets: phaser → kc705_phaser
14:54
<
sb0 >
bb-m-labs: force build --props=package=artiq-kc705-phaser --branch=release-3 artiq-board
14:55
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-nist_qc2 artiq-board
14:55
<
bb-m-labs_ >
build forced [ETA 13m00s]
14:55
<
bb-m-labs_ >
I'll give a shout when the build finishes
14:55
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-phaser --branch=release-3 artiq-board
14:55
<
bb-m-labs_ >
The build has been queued, I'll give a shout when it starts
14:57
<
sb0 >
rjo, do we keep artiq/gateware/targets/kc705_sma_spi.py?
15:03
<
rjo >
sb0: let's keep it. i use it.
15:10
<
sb0 >
what major features should we announce? DMA? PDQ2? DRTIO and SAWG demos?
15:11
<
bb-m-labs_ >
build forced [ETA 14m48s]
15:11
<
bb-m-labs_ >
I'll give a shout when the build finishes
15:14
<
GitHub85 >
artiq/release-3 04a9a0c Sebastien Bourdeauducq: doc: no more win32 packages
15:15
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-nist-qc2 --branch=release-3 artiq-board
15:15
<
bb-m-labs_ >
build forced [ETA 14m48s]
15:15
<
bb-m-labs_ >
I'll give a shout when the build finishes
15:18
<
sb0 >
bb-m-labs_: force build --branch=release-3 artiq
15:18
<
bb-m-labs_ >
build forced [ETA 36m16s]
15:18
<
bb-m-labs_ >
I'll give a shout when the build finishes
15:27
d_n|a has joined #m-labs
15:33
d_n|a has quit [Ping timeout: 258 seconds]
15:34
<
rjo >
sb0: in no particular order: DMA, DRTIO, SAWG, out-of-tree PDQ, rust, smoltcp
15:45
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-nist-qc2 --branch=release-3 artiq-board
15:45
<
bb-m-labs_ >
build forced [ETA 13m38s]
15:45
<
bb-m-labs_ >
I'll give a shout when the build finishes
15:45
d_n|a has joined #m-labs
15:45
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-phaser --branch=release-3 artiq-board
15:45
<
bb-m-labs_ >
build forced [ETA 13m38s]
15:45
<
bb-m-labs_ >
I'll give a shout when the build finishes
15:46
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-nist_qc2 --branch=release-3 artiq-board
15:46
<
bb-m-labs_ >
The build has been queued, I'll give a shout when it starts
16:06
<
bb-m-labs_ >
build forced [ETA 16m23s]
16:06
<
bb-m-labs_ >
I'll give a shout when the build finishes
16:08
gric has quit [Ping timeout: 260 seconds]
16:08
gric has joined #m-labs
16:14
<
whitequark >
sb0: no holdup. I tried to understand it yesterday properly but couldn't quite, let me just copy the code
16:25
<
sb0 >
bb-m-labs_: force build --branch=release-3 artiq
16:25
<
bb-m-labs_ >
build forced [ETA 43m50s]
16:25
<
bb-m-labs_ >
I'll give a shout when the build finishes
16:29
d_n|a has quit [Ping timeout: 248 seconds]
16:31
<
sb0 >
bb-m-labs_: force build --branch=release-3 artiq
16:31
<
bb-m-labs_ >
build forced [ETA 43m50s]
16:31
<
bb-m-labs_ >
I'll give a shout when the build finishes
16:42
<
rjo >
whitequark, sb0: should we hook up the machine here in berlin as a buildslave?
16:43
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-nist_qc2 --branch=release-3 artiq-board
16:43
<
bb-m-labs_ >
The build has been queued, I'll give a shout when it starts
16:44
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-phaser --branch=release-3 artiq-board
16:44
<
bb-m-labs_ >
The build has been queued, I'll give a shout when it starts
16:45
<
GitHub103 >
artiq/4.0.dev e7594b5 Sebastien Bourdeauducq: RELEASE_NOTES: add 4.0 section
16:45
<
GitHub196 >
artiq/master e7594b5 Sebastien Bourdeauducq: RELEASE_NOTES: add 4.0 section
16:45
<
GitHub196 >
artiq/master c7cdd20 Sebastien Bourdeauducq: doc: no more win32 packages
16:45
<
sb0 >
rjo, to the main buildbot?
16:48
<
whitequark >
rjo: what's the benefit?
16:48
<
whitequark >
or rather, the purpose?
16:48
<
rjo >
dunno. that's why i'm asking. do we need more cpu/mem/kc705s?
16:51
<
sb0 >
mem/kc705 no, cpu not sure
16:55
<
bb-m-labs_ >
build forced [ETA 14m35s]
16:55
<
bb-m-labs_ >
I'll give a shout when the build finishes
17:02
<
sb0 >
conda/builbot is again a trash fire ...
17:04
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-nist_clock --branch=release-3 artiq-board
17:04
<
bb-m-labs_ >
The build has been queued, I'll give a shout when it starts
17:07
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-nist_qc2 --branch=release-3 artiq-board
17:07
<
bb-m-labs_ >
The build has been queued, I'll give a shout when it starts
17:10
<
GitHub47 >
artiq/master 4bafdec Sebastien Bourdeauducq: RELEASE_NOTES: formatting
17:10
<
GitHub31 >
artiq/release-3 d500e61 Sebastien Bourdeauducq: RELEASE_NOTES: formatting
17:11
<
GitHub23 >
artiq/master b4c52c3 Sebastien Bourdeauducq: Merge branch 'sinara'
17:15
<
whitequark >
sb0: more ram is always good as it lets us use vms more widely
17:15
<
sb0 >
i'll merge rtio-sed later, since it has sawg/sayma breakage potential
17:15
<
whitequark >
e.g. we could move all builds to checkpointed vms
17:16
<
whitequark >
that's 1-2 days of fixing buildbot config and it gives us completely repeatable builds sans intrinsic things like embedded timestamps
17:17
<
sb0 >
we have 16GB RAM now. is that not enough for doing that already?
17:17
<
whitequark >
barely
17:17
<
whitequark >
vivado needs how much?
17:18
<
whitequark >
well, we could limit parallel vivado builds, and use ballooning
17:18
<
bb-m-labs_ >
build forced [ETA 18m30s]
17:18
<
bb-m-labs_ >
I'll give a shout when the build finishes
17:20
<
_florent_ >
rjo: indeed there is something wrong, I should have fix converter datas, but cordic is running at 250Mhz, can you give the new csr parameters if differents?
17:25
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-nist_qc2 --branch=release-3 artiq-board
17:25
<
bb-m-labs_ >
The build has been queued, I'll give a shout when it starts
17:27
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-nist_qc2 --revision=04a9a0ce95947185a8fb38f886e515ab239e544d artiq-board
17:27
<
bb-m-labs_ >
The build has been queued, I'll give a shout when it starts
17:32
<
GitHub88 >
artiq/release-3 a49bb2b Sebastien Bourdeauducq: conda: fix llvmlite-artiq dependency
17:32
<
GitHub152 >
artiq/master 0f4bc54 Sebastien Bourdeauducq: conda: fix llvmlite-artiq dependency
17:34
<
rjo >
_florent_: but if the cordic runs at 250MHz, then you still need to repeat every sample 4 times (not 2).
17:34
<
bb-m-labs_ >
build forced [ETA 15m34s]
17:34
<
bb-m-labs_ >
I'll give a shout when the build finishes
17:34
<
_florent_ >
rjo: yes i changed that
17:35
<
rjo >
if the cordic runs at 250MHz (assuming it meets timing) then the frequency is just twice the value, i.e. 20 MHz.
17:37
<
_florent_ >
rjo: ok thanks, i uploaded the new bitstreams
17:37
d_n|a has joined #m-labs
17:38
<
bb-m-labs_ >
build forced [ETA 15m34s]
17:38
<
bb-m-labs_ >
I'll give a shout when the build finishes
17:40
<
sb0 >
bb-m-labs_: force build --props=package=artiq-kc705-nist_clock --revision=04a9a0ce95947185a8fb38f886e515ab239e544d artiq-board
17:40
<
bb-m-labs_ >
The build has been queued, I'll give a shout when it starts
17:47
<
rjo >
_florent_: i updated the comment.
17:47
<
rjo >
_florent_: thanks for integrating that!
17:48
<
rjo >
_florent_: would you do me a favor with your sayma? if you have a linux machine (a bit easier with linux than with windows but it should also work just fine on windows), could you try our openocd branch and the proxy bitstream on your sayma and see whether the second flash chip works?
17:54
<
_florent_ >
rjo: i'll try to test that next week
17:56
<
bb-m-labs_ >
build forced [ETA 16m17s]
17:56
<
bb-m-labs_ >
I'll give a shout when the build finishes
17:56
<
rjo >
_florent_: merci!
17:57
FabM has quit [Quit: ChatZilla 0.9.93 [Firefox 52.3.0/20170811091919]]
17:59
<
_florent_ >
rjo: can you just give me the exact intructions to execute? (to avoid loosing time)
18:27
d_n|a has quit [Ping timeout: 240 seconds]
18:46
rohitksingh has joined #m-labs
19:02
<
GitHub61 >
artiq/master 2821f50 Robert Jordens: conda: jesd204b 0.4
19:02
<
GitHub61 >
artiq/master 2604806 Robert Jordens: sayma_rtm: make build dir
19:02
<
GitHub61 >
artiq/master 5e3cc83 Robert Jordens: sayma_amc: SAWG (untested)
20:22
rohitksingh has quit [Quit: Leaving.]
23:12
d_n|a has joined #m-labs