Sunday, 2021-08-08

*** tpb <[email protected]> has joined #yosys00:00
*** strobo_ <[email protected]> has joined #yosys01:17
*** strobo <[email protected]> has quit IRC (Ping timeout: 256 seconds)01:20
*** kristianpaul <kristianpaul!~paul@user/kristianpaul> has quit IRC (Quit: WeeChat 2.3)02:19
*** kristianpaul <[email protected]> has joined #yosys02:19
*** sm2n_ is now known as sm2n04:17
*** emeb_mac <[email protected]> has quit IRC (Ping timeout: 272 seconds)05:38
*** emeb_mac <[email protected]> has joined #yosys05:42
*** Sarayan <Sarayan!~galibert@2a01:e34:ec57:ec20:3e8c:f8ff:fe21:d13> has quit IRC (Quit: vacayyyyyyy)06:13
*** emeb_mac <[email protected]> has quit IRC (Quit: Leaving.)06:18
*** jeanthom <[email protected]> has joined #yosys08:29
*** gatecat <[email protected]> has quit IRC (Ping timeout: 250 seconds)09:51
*** lkcl <[email protected]> has quit IRC (Ping timeout: 245 seconds)09:51
*** gatecat <[email protected]> has joined #yosys09:51
*** Max-P <Max-P!~Max-P@thelounge/maintainer/Max-P> has quit IRC (Ping timeout: 245 seconds)09:51
*** anticw <[email protected]> has quit IRC (Ping timeout: 250 seconds)09:55
*** tnt <tnt!~tnt@osmocom/tnt> has quit IRC (Ping timeout: 252 seconds)09:59
*** e_ter <[email protected]> has quit IRC (Ping timeout: 245 seconds)10:03
*** gmc_ <[email protected]> has quit IRC (Ping timeout: 252 seconds)10:04
*** gmc_ <[email protected]> has joined #yosys10:05
*** eater <[email protected]> has joined #yosys10:11
*** tnt <tnt!~tnt@osmocom/tnt> has joined #yosys10:17
*** anticw <[email protected]> has joined #yosys10:26
*** vidbina_ <[email protected]> has joined #yosys10:39
*** adjtm <[email protected]> has quit IRC (Read error: Connection reset by peer)12:24
*** adjtm <[email protected]> has joined #yosys12:26
*** strobo_ <[email protected]> has quit IRC (Read error: Connection reset by peer)12:26
*** strobo <[email protected]> has joined #yosys12:26
cr1901>(9:04:36 AM) sorear: trouble finding hardware?12:35
cr1901No, in ABC's case, C (C++?) doesn't make it exceptionally easy to do endian-independent "store/load word from smaller parts", and ABC, well, bakes in little-endian assumptions12:35
cr1901And when a yosys bug was filed that triggered the bug on big endian systems, it became clear that it's too much work to either add #ifdef BIG_ENDIAN or alter all the store/loads into endian-independent ones12:36
sorear<lofty> mwk: Trying to find big-endian hardware for these things is hard, but at least I discovered it >.>12:49
soreari was asking what that meant12:49
mwk... I swear if we keep finding abc endianness / pointer-size issues, I'm putting it into a wasm32 jail where it belongs and shipping yosys with wasi interpreter13:04
whitequarkmwk: don't need an interpreter13:08
whitequarkuse wasm2c13:08
whitequarki'm excited for yowasp to be shipping two nested layers of wasm runtimes13:09
cr1901sorear: Hmmm... what big endian CPUs still have hardware made for them, new in, say, 2019 (before parts shortage)?13:11
mwkSystem Z, POWER13:11
sorearcr1901: I'm told you can run big endian VMs on all armv8 hardware, haven't gotten around to trying it yet13:12
mwkplus, well, technically ARM is bi, but noone really uses it that way13:12
cr1901RISCV allows big-endian as nonstandard, but yea no one is gonna use that either13:12
cr1901sorear: I don't understand what you mean. You can run big-endian emulators on x86 too. It's just not fast :P.13:14
sorearcr1901: aarch64 kvm supports guest endian != host endian13:14
cr1901Ahhh, that sounds stressful to implement, but good for ARM13:15
sorearthat's why it's kvm and not general user applications - much smaller ABI surface to translate13:17
cr1901I can just imagine me trying to implement, say, a 68k JIT, and screwing up which part of a word I load/store when generate host code for insns that access memory13:18
cr1901(Sorry, 68k is still my go-to example for big endian :P)13:18
*** jeanthom <[email protected]> has quit IRC (Ping timeout: 252 seconds)13:19
cr1901mwk: I forgot POWER was big-endian. I have plans to try out MicroWatt running Linux too13:24
mwkPOWER is migrating to little-endian though13:24
mwklike, big-endian is still a thing, but it's becoming legacy13:25
cr1901Shame... they got bought out by the little-endian lobbyists :P13:27
*** emeb <[email protected]> has joined #yosys14:56
*** jeanthom <[email protected]> has joined #yosys16:03
XarkThank goodness PowerPC on PS3 also had a "little endian" load/store opcodes or it would have been a bummer running translated MIPS code.  Perhaps last big-endian system in wide use...16:30
*** strobo <[email protected]> has quit IRC (Quit: leaving)16:48
mwkwhitequark: so you mentioned you had some thoughts about poison, X, and UB in yosys; do you have that written down somewhere?  I feel this is a discussion we need to have soon-ish about our IR17:19
mwk(see also: https://github.com/YosysHQ/yosys/issues/2924 and probably more issues like that in the codebase)17:20
*** vidbina_ <[email protected]> has quit IRC (Ping timeout: 245 seconds)17:25
*** Max-P <Max-P!~Max-P@thelounge/maintainer/Max-P> has joined #yosys17:45
*** indy <[email protected]> has joined #yosys18:09
*** esden <[email protected]> has quit IRC (Ping timeout: 240 seconds)18:48
*** esden <[email protected]> has joined #yosys18:51
*** GenTooMan <GenTooMan!~cyberman@2601:547:437f:e5c6:21f:5bff:fefe:a883> has quit IRC (Ping timeout: 272 seconds)19:40
*** lkcl <[email protected]> has joined #yosys19:52
*** GenTooMan <GenTooMan!~cyberman@2601:547:437f:e5c6:21f:5bff:fefe:a883> has joined #yosys19:53
*** vidbina_ <[email protected]> has joined #yosys20:31
*** emeb_mac <[email protected]> has joined #yosys20:36
*** jeanthom <[email protected]> has quit IRC (Ping timeout: 245 seconds)20:52
*** emeb <[email protected]> has quit IRC (Quit: Leaving.)21:51
*** vidbina_ <[email protected]> has quit IRC (Ping timeout: 256 seconds)22:42

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!