| openpowerbot_ | [mattermost] <toshywoshy> Cesar, get well soon | 00:56 |
|---|---|---|
| awilfox_ | lkcl: I don't see any issue with the kernel emulation doing its job, that's why it's there after all :) I don't know if it will especially have v2.06 ISA enabled | 03:02 |
| awilfox_ | my thought was more -march=powerpc64 (which should be baseline ISA, I think v2.0?) with -mno-altivec -mno-vsx | 03:02 |
| awilfox_ | obviously this won't be tuned very well, you could have faster speed compiling with the v3.0 extensions enabled that Libre-SOC does support | 03:02 |
| awilfox_ | but the point is to have a distro that can run on e5500 and Libre-SOC, which are two targets I keep hearing about and people are interested in, but not getting much attention it seems from the bigger distros | 03:03 |
| *** awilfox_ is now known as awilfox | 03:03 | |
| awilfox | I was thinking about e5500 support for quite some time, but it seemed to be a "waste" of resources when each subarch takes somewhere near ~100GB of mirror space, + testing and such | 03:04 |
| awilfox | adding Libre-SOC to the planned "subarch" makes it a lot more attractive as now we can have two targets supported with one spin | 03:05 |
| *** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC | 03:50 | |
| *** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc | 04:07 | |
| *** Ritish <Ritish!~Ritish@2401:4900:1f2d:31dd:494c:873e:38e5:91bc> has joined #libre-soc | 05:02 | |
| *** Ritish <Ritish!~Ritish@2401:4900:1f2d:31dd:494c:873e:38e5:91bc> has quit IRC | 05:22 | |
| *** Ritish <Ritish!~Ritish@2401:4900:1f2d:31dd:2277:6642:1a25:979b> has joined #libre-soc | 05:22 | |
| *** Ritish <Ritish!~Ritish@2401:4900:1f2d:31dd:2277:6642:1a25:979b> has quit IRC | 07:36 | |
| *** Ritish <Ritish!~Ritish@2401:4900:1f2d:31dd:9095:f091:e7a6:cd3e> has joined #libre-soc | 08:37 | |
| markos | awilfox, iirc, there was a debian port for e5500, powerpc-spe? | 09:17 |
| markos | ah no, that's e500 | 09:18 |
| markos | e5500 are the QorIQs iirc, capable or running LE but only have VMX not VSX, iirc | 09:33 |
| *** octavius <octavius!~octavius@92.40.168.143.threembb.co.uk> has joined #libre-soc | 09:37 | |
| sadoon[m] | cesar: get well soon, I got a slight fever on my way back too, too many people from different countries at FOSDEM I suppose | 09:45 |
| markos | cesar, sadoon[m], same her, I tested myself for covid/flu and was negative, but I definitely got something that kept me tired for a couple of days and sneeze a lot :D | 09:56 |
| openpowerbot_ | [mattermost] <cesar_strauss> Thanks for all the well-wishes. I'm taking my medication and feeling increasingly better. Much better than Saturday, in any case. | 09:56 |
| sadoon[m] | Sounds about right :p | 09:57 |
| markos | it's late summer there no? get out in the sun and have some oranges, vitamin C & D :) | 09:57 |
| lkcl | cesar, fantastic - yeah i actually brought along some 4,000 AU vitamin D with me | 09:57 |
| lkcl | awilfox, there's actually three: microwatt as well | 09:58 |
| markos | (always read AU as astronomical unit for some reason which gives it a funny twist) | 09:58 |
| lkcl | the cost of adding bare-minium VSx hardware into the FPGA bitstream is an astronomical 50% increase in LUT4s | 09:59 |
| lkcl | there are actual instructions already emulated in the linux kernel, the code to do so was started appx 20 years ago by paul mackerras | 10:06 |
| lkcl | and has grown since | 10:06 |
| lkcl | things like load-quad, and load-store-byte-reverse | 10:07 |
| lkcl | which are basically the difference between a v2.06/7/8 ISA and a v3.0 | 10:07 |
| lkcl | we very specifically need to stick to v3.0 because of the EULA from the OpenPOWER Foundation. | 10:08 |
| lkcl | *everyone* creating new Power ISA hardware has to do that | 10:09 |
| *** octavius_ <octavius_!~octavius@2a01:4c8:1065:db27:9ddd:f455:df27:168c> has joined #libre-soc | 10:18 | |
| *** Ritish31 <Ritish31!~Ritish@2401:4900:1f2d:31dd:9095:f091:e7a6:cd3e> has joined #libre-soc | 10:20 | |
| *** octavius <octavius!~octavius@92.40.168.143.threembb.co.uk> has quit IRC | 10:20 | |
| *** Ritish <Ritish!~Ritish@2401:4900:1f2d:31dd:9095:f091:e7a6:cd3e> has quit IRC | 10:21 | |
| *** Ritish31 <Ritish31!~Ritish@2401:4900:1f2d:31dd:9095:f091:e7a6:cd3e> has quit IRC | 10:21 | |
| *** Ritish <Ritish!~Ritish@2401:4900:1f2d:31dd:9095:f091:e7a6:cd3e> has joined #libre-soc | 10:21 | |
| programmerjake | note that lq/stq can't really be emulated when the memory address is aligned, they are 16-byte atomic ops, so need to be implemented using a 16-byte atomic instruction. no, a global mutex isn't sufficient. | 10:22 |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.160.32> has joined #libre-soc | 10:25 | |
| programmerjake | linux just runs lq/stq again when trying to emulate lq/stq: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=350779a29f11f80ac66a8b38a7718ad30f003f18 | 10:26 |
| sadoon[m] | Fun: my local file server is giving me ECC errors that I can't reproduce with memtest86.. | 10:40 |
| sadoon[m] | Bah I'm just gonna sell this stupid AMD threadripper and get an actual server motherboard | 10:42 |
| programmerjake | maybe try memtest86+? https://www.phoronix.com/news/memtest86-plus-6.10 | 10:48 |
| sadoon[m] | I meant to say memtest86+ | 10:50 |
| sadoon[m] | :') | 10:50 |
| sadoon[m] | Luckily I have an extra Xeon 1650v3 and motherboards are getting extremely cheap for those models | 10:50 |
| sadoon[m] | If I can find one locally it'd be great | 10:50 |
| programmerjake | maybe reseat your ram and cpu? they can sometimes have bad connections due to thermal cycling or other reasons | 10:52 |
| programmerjake | could also be the memory interface is being run slightly too fast or power issues | 10:53 |
| sadoon[m] | You might be right but I was already planning to replace this motherboard as it has gotten annoying for other reasons too | 10:53 |
| programmerjake | so you're looking for an excuse to replace it :) | 10:54 |
| sadoon[m] | Kinda :P | 10:54 |
| sadoon[m] | It's always been buggy given AMD threadripper 1st gen was a skunkworks project at AMD as well as their first Zen product | 10:55 |
| sadoon[m] | Always had CPU and Memory clock issues | 10:55 |
| programmerjake | oh, i gave my zen 1 cpu to my dad, then he got rid of it too...it always had mysterious issues | 10:56 |
| sadoon[m] | Yep, it was a very rushed product | 10:56 |
| programmerjake | for me was terrible for building code due to random segfaults, couldn't compile llvm in docker since it would crash halfway most of the time | 10:57 |
| programmerjake | for my dad, he ran into some other issue (lockups? icr very well) | 10:57 |
| sadoon[m] | I remember that fiasco | 10:58 |
| sadoon[m] | the segfault thing | 10:58 |
| sadoon[m] | Wow can you imagine that was 6 years ago? feels like yesterday they released Zen 1 | 10:58 |
| programmerjake | XD | 10:58 |
| programmerjake | i remember being thoroughly impressed at how much faster it was than my 8-core bulldozer cpu i had before (which was also impressive compared to my previous computer, a dual-socket 1-core/2-thread per socket pentium 4 workstation i found at a thrift store) | 11:02 |
| programmerjake | the pentium 4 computer was slower single-threaded than my laptop, which had 2/3 the clock frequency | 11:03 |
| sadoon[m] | Heheh | 11:04 |
| programmerjake | turion64 laptop | 11:04 |
| sadoon[m] | Intel despite being absolute scumbags have always had the server market rock solid | 11:04 |
| sadoon[m] | And workstations* | 11:04 |
| programmerjake | so basically i went from junkheap to junkheap to zen1 | 11:04 |
| programmerjake | intel & servers: yeah, till recently | 11:07 |
| sadoon[m] | Yup | 11:07 |
| sadoon[m] | I'd say they had it right until EPYC came around and they got stuck at 14nm | 11:08 |
| programmerjake | imho their core count lagging was more of an issue for servers, they're often not as sensitive to ultimate single-threaded performance | 11:10 |
| programmerjake | well, gtg -- need sleep before the meeting today :) | 11:11 |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.160.32> has quit IRC | 11:21 | |
| *** Ritish <Ritish!~Ritish@2401:4900:1f2d:31dd:9095:f091:e7a6:cd3e> has quit IRC | 11:32 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.170.188> has joined #libre-soc | 11:39 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.170.188> has quit IRC | 11:43 | |
| sadoon[m] | Night! | 12:18 |
| sadoon[m] | Instead of buying a server I found a 2000ish era Toshiba laptop with Windows ME | 12:19 |
| sadoon[m] | Couldn't resist | 12:19 |
| octavius_ | Ew, Windows :P | 12:24 |
| *** octavius__ <octavius__!~octavius@92.40.168.143.threembb.co.uk> has joined #libre-soc | 12:26 | |
| *** octavius_ <octavius_!~octavius@2a01:4c8:1065:db27:9ddd:f455:df27:168c> has quit IRC | 12:28 | |
| *** octavius__ <octavius__!~octavius@92.40.168.143.threembb.co.uk> has quit IRC | 12:29 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.170.110> has joined #libre-soc | 12:48 | |
| sadoon[m] | Don't be so down | 12:48 |
| sadoon[m] | I'll try reactOS on it :D | 12:48 |
| lkcl | toshywoshy, thank you | 13:41 |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@176.59.170.110> has quit IRC | 16:54 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@62.176.2.159> has joined #libre-soc | 16:55 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@62.176.2.159> has quit IRC | 17:31 | |
| *** ghostmansd[m] <ghostmansd[m]!~ghostmans@62.176.2.159> has joined #libre-soc | 17:32 | |
| *** josuah <josuah!~irc@46.23.94.12> has quit IRC | 19:42 | |
| *** irc <irc!~irc@46.23.94.12> has joined #libre-soc | 19:43 | |
| *** irc <irc!~irc@46.23.94.12> has joined #libre-soc | 19:45 | |
| *** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has quit IRC | 20:22 | |
| *** lxo <lxo!~lxo@gateway/tor-sasl/lxo> has joined #libre-soc | 20:23 | |
| *** irc is now known as josuah | 20:23 | |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!