Logs: liberachat/#xmonad
| 2021-11-01 13:10:16 | <L29Ah> | and on the lower side it's as cold as noisy your fan is |
| 2021-11-01 13:10:17 | <liskin> | it's normally limited to 80, because I haven't bothered with any of the throttled hacks, I like my laptop cool |
| 2021-11-01 13:10:36 | <liskin> | L29Ah: 41 without a fan |
| 2021-11-01 13:10:47 | L29Ah | sometimes wishes to have dock plugs for liquid cooling |
| 2021-11-01 13:10:49 | <Solid> | mine almost melts through my desk when compiling :) |
| 2021-11-01 13:10:51 | <liskin> | it does kick in around 43 or something |
| 2021-11-01 13:11:18 | <L29Ah> | liskin: nice, what's the device? |
| 2021-11-01 13:11:33 | <L29Ah> | and ambient temperature |
| 2021-11-01 13:11:33 | <liskin> | but for some strange reason the CPU started switching to a 65C limit lately |
| 2021-11-01 13:11:51 | <liskin> | L29Ah: T25 |
| 2021-11-01 13:11:57 | <liskin> | vanilla, no mods |
| 2021-11-01 13:12:08 | <liskin> | ambient temp around 21 I think |
| 2021-11-01 13:12:59 | <liskin> | it's already out of warranty so I may eventually need to replace it with a 51nb something |
| 2021-11-01 13:13:09 | <liskin> | or just get used to the "modern" keyboard :-/ |
| 2021-11-01 13:13:49 | <L29Ah> | i'm pissed with the firmware quality of 51nb, but still it seems to be the only decent option on the market these days |
| 2021-11-01 13:14:12 | <liskin> | yeah, that's what I'm worried about |
| 2021-11-01 13:14:33 | <liskin> | as I understand it, even the T480/25 frankenpad has issues with not all keys working as they should |
| 2021-11-01 13:14:49 | <L29Ah> | 16:11:51]<liskin> L29Ah: T25 |
| 2021-11-01 13:14:49 | <L29Ah> | can you please show the steady output of https://github.com/l29ah/intel-powermon/ at max load? |
| 2021-11-01 13:15:07 | <liskin> | so I'm kind of hoping they release something for the 30th anniversary, but at this point nobody cares any more :-/ |
| 2021-11-01 13:15:30 | <liskin> | L29Ah: I have a meeting now but remind me later today |
| 2021-11-01 13:15:35 | <L29Ah> | alright |
| 2021-11-01 13:27:35 | × | gruntsplatter quits (~sogens@gateway/vpn/pia/sogens) (Quit: WeeChat 3.3) |
| 2021-11-01 13:31:37 | → | cfricke joins (~cfricke@user/cfricke) |
| 2021-11-01 13:50:28 | <liskin> | L29Ah: https://gist.github.com/1f8e901a82bd89b48c5beafa2c64c06d - building xmonad-contrib with -j, temp slowly raised up to the 80 ceiling during the first minute or so |
| 2021-11-01 13:51:00 | <liskin> | I can try to raise the limit to 95 and see what happens |
| 2021-11-01 13:51:32 | <L29Ah> | this sounds like your CPU is throttled into oblivion |
| 2021-11-01 13:52:09 | <liskin> | well it's a U, of course it's throttled into oblivion :-) |
| 2021-11-01 13:52:10 | <L29Ah> | but still the cooling system is somewhat better than X201's |
| 2021-11-01 13:52:35 | <liskin> | but I'm not getting messages about it being throttled, which is a bit weird |
| 2021-11-01 13:52:50 | <liskin> | I used to get those in dmesg before, but they stopped once I undervolted |
| 2021-11-01 13:53:08 | <liskin> | so there are probably two different throttling mechanisms |
| 2021-11-01 13:53:29 | <L29Ah> | 8% better |
| 2021-11-01 13:53:48 | <L29Ah> | liskin: yeah by trottling i mean absolute frequency cap |
| 2021-11-01 13:54:03 | <L29Ah> | not the situational cap due to overheating |
| 2021-11-01 13:54:12 | <liskin> | L29Ah: oh, right |
| 2021-11-01 13:55:16 | <L29Ah> | can be illustrated by `sysbench cpu run --threads=9 --time=5` |
| 2021-11-01 13:55:22 | <L29Ah> | 42233 here |
| 2021-11-01 13:55:25 | <liskin> | hm, those numbers look almost the same with the limit raised to 95 |
| 2021-11-01 13:55:48 | <L29Ah> | but does actual temperature rise to 95? |
| 2021-11-01 13:55:58 | <liskin> | not really, barely gets to 80 |
| 2021-11-01 13:56:03 | <liskin> | I'll need to run stack in a loop |
| 2021-11-01 13:56:25 | <liskin> | or power the nvidia up to stress the cooling system a bit |
| 2021-11-01 13:58:55 | <liskin> | https://gist.github.com/0e84c03f0dea1fa066395ed54b7b7769 - ghc -j + primusrun glxgears |
| 2021-11-01 13:59:09 | <liskin> | 3.5k fan rpm, 83C tops |
| 2021-11-01 13:59:57 | <liskin> | I think I might need to write to another MSR to override the wattage cap, not just the thermal cap |
| 2021-11-01 14:03:14 | × | kr1x quits (~kriket@170.249.0.122) (Ping timeout: 260 seconds) |
| 2021-11-01 14:05:16 | → | kr1x joins (kriket@gateway/vpn/protonvpn/kr1x) |
| 2021-11-01 14:07:53 | <liskin> | hm, it's weird, I raised the wattage caps, but it still draws under 20 watts, and when I check with turbostat, all 4 cores are at their turbo frequencies all the time |
| 2021-11-01 14:08:18 | <liskin> | maybe ghc is too lazy :-) |
| 2021-11-01 14:10:29 | × | kr1x quits (kriket@gateway/vpn/protonvpn/kr1x) (Ping timeout: 268 seconds) |
| 2021-11-01 14:11:34 | → | kr1x joins (~kriket@170.249.0.122) |
| 2021-11-01 14:16:22 | <FOSSHuman[m]> | https://github.com/AdnanHodzic/auto-cpufreq |
| 2021-11-01 14:17:20 | <FOSSHuman[m]> | This can greatly extend the battery life of a laptop, it automatically adjusts the CPU frequency |
| 2021-11-01 14:30:12 | × | kr1x quits (~kriket@170.249.0.122) (Ping timeout: 268 seconds) |
| 2021-11-01 14:31:40 | <L29Ah> | FOSSHuman[m]: i'm skeptical of the utility of this program |
| 2021-11-01 14:31:53 | → | kr1x joins (kriket@gateway/vpn/protonvpn/kr1x) |
| 2021-11-01 14:32:19 | <L29Ah> | as it assumes that CPU is the sole power consumer, and you don't need the results of CPU work |
| 2021-11-01 14:33:03 | <L29Ah> | and/or additional waiting for the result is acceptable |
| 2021-11-01 14:35:00 | <L29Ah> | sure you can increase computations per joule if you lower the frequency somewhat, but it becomes interestingly nonlinear when the power consumption of other system components, notably LCD backlight, is brought into the equation |
| 2021-11-01 14:36:57 | <L29Ah> | and advanced techniques like race-to-idle bring even more confusion |
| 2021-11-01 14:37:27 | L29Ah | is disappointed by the lack of benchmarks in README |
| 2021-11-01 14:43:21 | × | kr1x quits (kriket@gateway/vpn/protonvpn/kr1x) (Remote host closed the connection) |
| 2021-11-01 14:47:50 | <liskin> | L29Ah: I just checked the picom slowness, and I had to revert both the commits you mentioned to get to a reasonable performance |
| 2021-11-01 14:49:11 | → | kr1x joins (~kriket@170.249.0.122) |
| 2021-11-01 14:50:45 | <L29Ah> | liskin: try noting that in the issue, as it seems like the maintainer have dismissed my concerns |
| 2021-11-01 14:50:57 | <L29Ah> | fwiw i have no clue how it works, if it's picom or xmonad or other apps' windows to blame, i just bisected it |
| 2021-11-01 14:52:29 | <liskin> | I think I have an idea what's going on |
| 2021-11-01 14:52:53 | <liskin> | in https://github.com/yshui/picom/issues/381 I see no answer from the maintainer |
| 2021-11-01 14:56:02 | <liskin> | L29Ah: without the free_paint_glx change reverted, glx_find_fbconfig is called more than 10 times on every xmonad layout resize |
| 2021-11-01 14:56:58 | <liskin> | dunno what that function does exactly, but it feels like it should be called once, not a hundred times every time something changes |
| 2021-11-01 14:59:06 | <liskin> | the other commit just makes glx_find_fbconfig slower |
| 2021-11-01 15:04:52 | <liskin> | I think glx_find_fbconfig just needs some memoization/caching and it will be good |
| 2021-11-01 15:05:10 | <liskin> | the range of its inputs is quite small |
| 2021-11-01 15:18:12 | × | kr1x quits (~kriket@170.249.0.122) (Quit: .) |
| 2021-11-01 15:44:42 | × | cfricke quits (~cfricke@user/cfricke) (Quit: WeeChat 3.3) |
| 2021-11-01 16:25:30 | × | qbt quits (~qbt@user/edun) (Quit: Leaving.) |
| 2021-11-01 16:42:10 | × | alternateved quits (~user@staticline-31-183-149-3.toya.net.pl) (Ping timeout: 268 seconds) |
| 2021-11-01 17:03:14 | → | seschwar joins (~seschwar@user/seschwar) |
| 2021-11-01 19:05:01 | → | qbt joins (~qbt@user/edun) |
| 2021-11-01 19:05:25 | × | qbt quits (~qbt@user/edun) (Client Quit) |
| 2021-11-01 19:28:45 | → | gruntsplatter joins (~sogens@gateway/vpn/pia/sogens) |
| 2021-11-01 19:53:20 | × | steve__ quits (~steve@ool-182c2aa4.dyn.optonline.net) (Read error: Connection reset by peer) |
| 2021-11-01 19:53:37 | → | steve__ joins (~steve@ool-182c2aa4.dyn.optonline.net) |
| 2021-11-01 19:55:27 | → | qbt joins (~qbt@user/edun) |
| 2021-11-01 19:55:32 | × | qbt quits (~qbt@user/edun) (Client Quit) |
| 2021-11-01 21:00:19 | → | Extropy joins (~extropy@user/extropy) |
| 2021-11-01 21:00:32 | → | alternateved joins (~user@staticline-31-183-149-3.toya.net.pl) |
| 2021-11-01 21:03:37 | → | charlie39 joins (~charlie39@223.238.124.49) |
| 2021-11-01 21:11:09 | ← | charlie39 parts (~charlie39@223.238.124.49) () |
| 2021-11-01 21:32:15 | × | mc47 quits (~mc47@xmonad/TheMC47) (Remote host closed the connection) |
| 2021-11-01 22:43:10 | × | alternateved quits (~user@staticline-31-183-149-3.toya.net.pl) (Ping timeout: 260 seconds) |
| 2021-11-01 23:08:02 | <vrs> | debian testing has... 0.15, apparently |
| 2021-11-01 23:08:18 | <geekosaur> | right, they haven't updated yet |
| 2021-11-01 23:08:33 | <geekosaur> | (there was no 0.16 xmonad, only xmonad-contrib) |
| 2021-11-01 23:10:32 | <geekosaur> | sid hasn't updated either, from the looks of it |
| 2021-11-01 23:10:38 | × | gruntsplatter quits (~sogens@gateway/vpn/pia/sogens) (Ping timeout: 260 seconds) |
| 2021-11-01 23:23:26 | <L29Ah> | Gentoo has 0.17, yay i can now get off the git |
| 2021-11-01 23:25:32 | × | geekosaur quits (~geekosaur@xmonad/geekosaur) (Remote host closed the connection) |
| 2021-11-01 23:27:11 | → | geekosaur joins (~geekosaur@xmonad/geekosaur) |
All times are in UTC.