Saturday, 2018-05-19

*** tpb has joined #yosys00:00
*** emeb_mac has joined #yosys00:00
*** promach__ has joined #yosys00:25
*** quigonjinn has quit IRC00:25
*** promach__ is now known as promach200:30
*** emeb has quit IRC02:30
*** cemerick_ has joined #yosys02:40
*** xrexeon has quit IRC02:43
*** jkiv has quit IRC04:25
*** leviathan has joined #yosys04:31
*** cemerick_ has quit IRC04:36
*** leviathan has quit IRC04:45
*** leviathan has joined #yosys04:46
*** X-Scale has quit IRC05:51
*** cemerick_ has joined #yosys05:55
*** leviathan has quit IRC06:02
*** leviathan has joined #yosys06:05
*** cemerick_ has quit IRC06:28
*** dys has joined #yosys06:32
*** GuzTech has joined #yosys06:58
*** dys has quit IRC06:58
*** jpo has quit IRC07:02
*** emeb_mac has quit IRC07:06
*** dys has joined #yosys07:15
*** jwhitmore has joined #yosys07:22
*** jwhitmore has quit IRC07:26
*** fouric has quit IRC08:08
*** fouric has joined #yosys08:09
*** quigonjinn has joined #yosys08:16
*** leviathan has quit IRC08:26
*** proteusguy has quit IRC08:37
*** proteusguy has joined #yosys08:40
*** dmin7 has joined #yosys08:43
*** leviathan has joined #yosys09:04
*** proteus-guy has quit IRC09:23
*** proteus-guy has joined #yosys09:24
*** proteusguy has quit IRC09:41
*** fsasm has joined #yosys09:53
*** mazzoo has quit IRC10:20
*** mazzoo has joined #yosys10:22
*** promach2 has quit IRC10:24
*** proteusguy has joined #yosys10:31
*** promach2 has joined #yosys10:37
*** xrexeon has joined #yosys10:38
*** promach2 has quit IRC10:48
*** promach__ has joined #yosys10:49
*** promach__ is now known as promach210:49
*** promach2 has quit IRC10:54
*** fsasm has quit IRC10:56
*** mjoldfield has quit IRC11:11
*** mjoldfield has joined #yosys11:42
*** jwhitmore has joined #yosys11:55
*** jwhitmore has quit IRC12:00
*** dys has quit IRC12:14
*** dmin7 has quit IRC12:39
*** eduardo__ has joined #yosys12:43
*** eduardo_ has quit IRC12:46
*** dys has joined #yosys12:54
*** ssvb has quit IRC14:00
*** emeb has joined #yosys14:01
*** fsasm has joined #yosys14:14
*** mattvenn has quit IRC14:21
*** jayaura has quit IRC14:43
*** cfelton has quit IRC14:45
*** marex-cloud has quit IRC14:45
*** _florent_ has quit IRC14:46
*** cfelton has joined #yosys14:46
*** marex-cloud has joined #yosys14:46
*** mithro has quit IRC14:47
*** _florent_ has joined #yosys14:48
*** mithro has joined #yosys14:50
*** jayaura has joined #yosys14:50
*** jayaura is now known as Guest8271414:50
*** fsasm has quit IRC14:57
emebSkimmed thru this:
tpbTitle: [1805.03648] Parallel Programming for FPGAs (at
*** promach__ has joined #yosys15:00
*** promach__ is now known as promach215:00
emebWondering how much HLS really improves overall design quality / productivity.15:00
promach2For, why am I having this error "ERROR: Left operand of replicate expression is not constant at ../rtl/test_UART.v:192" ?15:01
tpbTitle: test_UART.v ยท GitHub (at
ZipCPUpromach2: Ahm ... because it isn't constant?  The error is pretty self explanatory.15:11
promach2ZipCPU: huh ? I am using genvar, why is it not constant ?15:16
shaprI had a shower thought yesterday about FPGA layout15:16
shaprbut I'm not quite sure how to turn it into words15:16
shaprfirst thought was, place and route on a sphere would be much easier, especially since the edges connect15:17
daveshahpromach2: because you are not using the genvar directly, but as an index to received_data_zeroes which is non-constant15:17
shaprthat turned into thinking about n-dimensional surface layout, I suspect it'd be easy to do, but I'm not sure about bringing it back down from higher dimensions15:18
ZipCPUshapr: My thought was that the first step to P&R should be an aggregation step, words should be placed together for example.15:18
ZipCPUThat would solve any carry chain issues.15:18
shaprZipCPU: elucidate?15:18
qu1j0t3there's a LOT of literature on this problem :)15:18
ZipCPUqu1j0t3: And I haven't read any of it! ;D15:18
shaprI was thinking about how SA gets stuck in local minima, I suspect doing layout in more dimensions and then reducing down to two may help with that15:19
ZipCPUshapr: Ooohh, yea, okay ... fascinating thought.15:20
qu1j0t3there are some other algorithms you could look at -- cuckoo search etc15:20
shaprrelated to cuckoo hashing, probably?15:20
emeb"reducing down to two" <- there's the rub.15:20
shapremeb: yeah, agreed15:20
ZipCPUMy thought was that: reg [15:0] A; always @(posedge i_clk) A <= A + 1; requires a series of logic elements that need to be placed near each other.  The first step should therefore be a block placement algorithm.15:21
shaprfirst thought about that was 'sticking down' part of a spherical layout, then re-running SA in 2D15:21
qu1j0t3the thing is "stuck in the wrong local minimum" is literally the issue, because the thing is NP-complete.15:21
qu1j0t3you can only mitigate it15:21
qu1j0t3& i think there are things you can do to help SA out?15:21
* ZipCPU pulls at his beard15:21
ZipCPUSituational awareness?15:22
qu1j0t3shapr: do you mean a torus?15:22
qu1j0t3ZipCPU: simulated annealing15:22
ZipCPUAhh ... thanks15:22
shaprtorus would be fine too15:22
shaprjust anything that makes the search easier at first15:22
shaprI also thought that a generic place-and-route might start with zero layout constraints on a sphere, just shrinking the sphere down to the perfect size to fit the layout15:23
shaprand then add in constraints15:23
shaprsuch as output pins and that kind of thing15:23
*** emeb has quit IRC15:24
shaprnow I wish I had a BLIF parser in Haskell15:24
* shapr digs into
promach2daveshah: thanks for the tips15:27
shaprqu1j0t3: what's the trade-off between a torus and sphere?15:28
*** cemerick_ has joined #yosys15:30
qu1j0t3if you are working 2 dimensional ordinates you might have some trouble mapping to a sphere?15:31
qu1j0t3there's also particle swarm optimisation. if i were doing this, i'd probably develop metrics and constraints i could use with pluggable algorithms, as there are quite a few possibilities15:31
qu1j0t3but things like cuckoo search and PSO look promising to me if SA isn't cutting it. but you will definitely want to review literature so you get some benefit from the large investments already made15:33
shapryes, I want a place and route experimental toolkit15:34
*** emeb_mac has joined #yosys15:37
*** thoughtpolice has quit IRC15:37
*** elms has quit IRC15:37
*** furan-- has quit IRC15:37
*** lvrp16 has quit IRC15:37
*** thoughtpolice has joined #yosys15:40
*** danieljabailey has quit IRC15:41
*** furan-- has joined #yosys15:41
*** lvrp16 has joined #yosys15:42
*** danieljabailey has joined #yosys15:43
*** elms has joined #yosys15:43
shaprawygle: were you able to find the source code for the GPU star+ place-and-route paper?16:04
shaprI found the primary author's github account, but don't see the source there16:04
*** promach2 has quit IRC16:05
*** jkiv has joined #yosys16:17
*** X-Scale has joined #yosys16:36
shapris there a specification for the blif format?16:48
*** emeb_mac has quit IRC16:51
awygleshapr: all i know of is
awygleand no, i didn't acquire the source. if you pinged that author and they could provide it that would be great17:03
*** m_t has joined #yosys17:33
*** jwhitmore has joined #yosys17:45
*** leviathan has quit IRC17:54
*** jwhitmore has quit IRC18:18
*** dmin7 has joined #yosys18:30
*** jwhitmore has joined #yosys18:43
*** proteusguy has quit IRC18:51
*** cemerick_ has quit IRC18:52
*** dxld has quit IRC19:03
*** proteusguy has joined #yosys19:04
*** dxld has joined #yosys19:14
*** pie__ has joined #yosys19:56
*** pie_ has quit IRC20:00
*** jkiv has quit IRC20:08
*** jkiv has joined #yosys20:18
*** jkiv has quit IRC20:25
*** jkiv has joined #yosys20:54
*** jkiv has joined #yosys20:55
*** jkiv has quit IRC20:55
*** jkiv has joined #yosys20:56
*** jkiv has joined #yosys20:57
*** jkiv has quit IRC20:57
*** jkiv has joined #yosys20:58
*** jkiv has joined #yosys20:59
*** jkiv has quit IRC20:59
*** dxld has quit IRC21:09
*** dxld has joined #yosys21:09
*** m_t has quit IRC21:17
*** cr1901_modern has quit IRC21:33
*** emeb_mac has joined #yosys21:36
*** dys has quit IRC21:43
*** GuzTech has quit IRC21:59
*** jwhitmore has quit IRC22:51
*** dxld has quit IRC23:45
*** dxld has joined #yosys23:51

Generated by 2.13.1 by Marius Gedminas - find it at!