Switch to DuckDuckGo Search
   November 25, 2019  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >

Toggle Join/Part | bottom
[00:46:32] <igork> andyf: you missed gcc 64bit bootstrap and need additional update if you move to use it
[00:47:24] <igork> or you try to check gcc 64bit as default with your build?
[01:12:34] *** bahamat_ <bahamat_!~bahamat@72.2.113.233> has left #illumos ("Textual IRC Client: www.textualapp.com")
[01:13:22] *** bahamat <bahamat!~bahamat@72.2.113.233> has joined #illumos
[02:08:26] *** baojg <baojg!~baojg@2604:a880:0:1010::16:1> has joined #illumos
[03:18:03] *** wl- <wl-!~wl_@2605:6000:1b0c:5a5f::87c> has quit IRC (Quit: Leaving)
[03:24:11] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Quit: jcea)
[03:29:19] *** khng300 <khng300!~khng300@unaffiliated/khng300> has quit IRC (Read error: Connection reset by peer)
[03:29:51] *** khng300 <khng300!~khng300@unaffiliated/khng300> has joined #illumos
[03:30:20] *** insomnia <insomnia!~insomnia@shadowcat/actuallyamemberof/lollipopguild.insomnia> has joined #illumos
[03:44:16] *** pwinder <pwinder!~pwinder@86.2.210.254> has joined #illumos
[03:45:59] *** alanc <alanc!~alanc@inet-hqmc01-o.oracle.com> has quit IRC (Remote host closed the connection)
[03:46:26] *** alanc <alanc!~alanc@inet-hqmc01-o.oracle.com> has joined #illumos
[06:41:25] *** yomisei <yomisei!~void@ip4d16bc15.dynamic.kabel-deutschland.de> has quit IRC (Remote host closed the connection)
[06:43:54] *** yomisei <yomisei!~void@77.22.188.21> has joined #illumos
[07:22:38] *** amrfrsh <amrfrsh!~Thunderbi@185.212.171.68> has quit IRC (Quit: amrfrsh)
[07:26:08] *** cpk <cpk!~chris@185.172.87.163> has quit IRC (Ping timeout: 276 seconds)
[08:17:42] *** amrfrsh <amrfrsh!~Thunderbi@131.234.44.76> has joined #illumos
[08:26:40] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Quit: tsoome)
[08:38:56] *** awordnot <awordnot!~awordnot@c-73-210-60-203.hsd1.il.comcast.net> has quit IRC (Ping timeout: 276 seconds)
[08:40:23] <leoric> Is the following panic (XHCI runtime reset required) https://www.illumos.org/issues/11049 ring some bells for somebody?
[09:02:57] *** denk <denk!~denis@devel.tambov.ru> has quit IRC (Remote host closed the connection)
[09:06:49] *** awordnot <awordnot!~awordnot@c-73-210-60-203.hsd1.il.comcast.net> has joined #illumos
[09:08:59] *** denk <denk!~denis@devel.tambov.ru> has joined #illumos
[09:17:36] *** Diagon <Diagon!DiagonalAr@gateway/vpn/protonvpn/diagonalarg> has joined #illumos
[10:04:05] *** jimklimov <jimklimov!~jimklimov@37-48-40-41.nat.epc.tmcz.cz> has joined #illumos
[10:21:56] *** amrfrsh <amrfrsh!~Thunderbi@131.234.44.76> has quit IRC (Ping timeout: 240 seconds)
[10:23:16] *** amrfrsh <amrfrsh!~Thunderbi@131.234.44.76> has joined #illumos
[10:26:04] *** Diagon <Diagon!DiagonalAr@gateway/vpn/protonvpn/diagonalarg> has quit IRC (Quit: Leaving)
[10:28:58] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[11:08:50] *** bacterio <bacterio!bacterio@fsf/member/bacterio> has quit IRC (Read error: Connection reset by peer)
[11:10:26] *** jimklimov <jimklimov!~jimklimov@37-48-40-41.nat.epc.tmcz.cz> has quit IRC (Quit: Leaving.)
[11:11:17] *** amrfrsh <amrfrsh!~Thunderbi@131.234.44.76> has quit IRC (Quit: amrfrsh)
[11:44:30] <andyf> igork, morning, sorry I don't understand what you said here last night
[11:45:40] <igork> andyf: hi. about your gnu ld update - do you try to use gcc 64bit bootstrap with 64bit GNU AS from binutils ?
[11:45:57] <igork> it's mean, you have 64bit gnu as and gnu ld
[11:46:26] <andyf> In OmniOS bloody, we've had a 64-bit binutils for a long time
[11:46:36] <andyf> but we are now changing it to assume/output 64-bit by default
[11:47:10] <igork> do you have 64bit gcc default bootstrap?
[11:47:28] <andyf> That's the bit I don't understand..
[11:47:47] <andyf> our gcc7 and gcc8 still default to 32-bit, if that answers the question
[11:48:02] <andyf> gcc9 defaults to 64-bit (same as OpenIndiana)
[11:48:21] <igork> yes, it is answer
[11:48:28] <andyf> (it means, we don't need the extra -m32 bits yet)
[11:48:43] <igork> with gcc 64bit default you have to fix grub1 build to be in 32bit
[11:49:13] <andyf> at some point, when the official illumos-gate compiler is a 64-bit default target, more changes will be needed - but I think that is some time away and we'll have to coordinate across distributions anyway
[11:50:40] <andyf> I do have another update for gcc44-il-4 to cater for the new binutils, I should open a PR..
[11:52:23] <igork> you need 32bit GNU AS for gcc44
[11:52:47] <igork> it can't be built with 64bit gnu as
[11:53:50] <andyf> yes it can
[11:53:54] <andyf> but it needs a patch
[11:54:40] <andyf> I'll open a PR against illumos/gcc later when I have time
[11:55:19] <andyf> You said you had proposed a patch for 12016 before - is there an issue that I missed?
[12:06:58] <igork> i do not remember
[12:07:06] <igork> it was a very long time ago
[12:07:17] <igork> it was proposal for dilos needs, but was ignored
[12:18:57] *** amrfrsh <amrfrsh!~Thunderbi@185.212.171.68> has joined #illumos
[12:34:31] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Read error: Connection reset by peer)
[12:36:02] *** Kassandry <Kassandry!~Kassandry@kassandry.net> has quit IRC (Ping timeout: 240 seconds)
[12:39:19] *** tsoome_ <tsoome_!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[12:39:20] *** Kassandr1 <Kassandr1!~Kassandry@kassandry.net> has joined #illumos
[12:39:49] *** tsoome_ is now known as tsoome
[12:40:44] *** gitomat <gitomat!~nodebot@165.225.148.18> has quit IRC (Remote host closed the connection)
[12:43:07] *** joltman <joltman!~znc@unaffiliated/joltman> has quit IRC (Quit: ZNC - http://znc.in)
[12:44:16] *** kev009 <kev009!~kev009@ip72-222-200-117.ph.ph.cox.net> has quit IRC (Remote host closed the connection)
[12:44:50] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 240 seconds)
[12:46:24] *** KungFuJesus <KungFuJesus!~adamstyli@207.250.97.74> has quit IRC (Remote host closed the connection)
[12:46:26] *** yomisei <yomisei!~void@77.22.188.21> has quit IRC (Ping timeout: 240 seconds)
[12:48:18] *** kev009_ <kev009_!~kev009@ip72-222-200-117.ph.ph.cox.net> has joined #illumos
[12:48:47] *** joltman_afk <joltman_afk!znc@gateway/vpn/privateinternetaccess/joltman> has joined #illumos
[12:48:47] *** clapont_ <clapont_!~clapont@213.233.88.148> has joined #illumos
[12:48:47] *** KungFuJe2us <KungFuJe2us!~adamstyli@207.250.97.74> has joined #illumos
[12:48:47] *** joltman_afk is now known as joltman
[12:48:55] *** gitomat <gitomat!~nodebot@165.225.148.18> has joined #illumos
[12:51:38] *** yomisei <yomisei!~void@ip4d16bc15.dynamic.kabel-deutschland.de> has joined #illumos
[12:53:23] *** freezing <freezing!~weechat@ec2-13-57-230-193.us-west-1.compute.amazonaws.com> has joined #illumos
[12:54:36] *** arnoldoree <arnoldoree!~arnoldore@ranoldoree.plus.com> has joined #illumos
[12:59:02] *** arnoldoree <arnoldoree!~arnoldore@ranoldoree.plus.com> has quit IRC (Client Quit)
[13:02:03] <andyf> I'm getting DEBUG assertion fail on boot after 11827
[13:04:28] *** morgib <morgib!sid342182@gateway/web/irccloud.com/x-cayifpjikotwgakm> has quit IRC (Quit: Connection closed for inactivity)
[13:12:11] <sensille> can i reserve a cpu completely for a process? i created a psrset with one cpu, but it still shows significant sys usage in mpstat
[13:12:33] <sensille> even after psrset -f (disable interrupts)
[13:23:47] <jlevon> that sounds odd. what do you have running
[13:27:41] <sensille> tons of incoming network traffic
[13:28:01] <sensille> 300k udp packets/s
[13:29:19] <sensille> ah, if i stop the processing process, all processors go idle except the first 20 or so
[13:29:45] <sensille> so i make the set for one of the processors above that
[13:30:40] <sensille> how are the hyperthreads arranged? do 2 consecutive "cpu" make one physical cpu?
[13:32:51] <sensille> doesn't work, still 50% system on cpu 55 :(
[13:34:58] <sensille> aah, 55 is special, too. it works on 50
[13:42:40] *** bacterio <bacterio!bacterio@fsf/member/bacterio> has joined #illumos
[13:47:05] <jlevon> sensille: I guess you have the mac worker threads on those cpus then. I'm kind of surprised, I thought cpu set would re-assign those.
[13:49:53] <sensille> and most frustrating: even with both hyperthreads reserved, linux is still 10% faster for the same assembler code on the same hardware
[13:50:32] <sensille> cpu only, no memory access
[13:51:51] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[14:19:14] *** neirac <neirac!~cneira@pc-184-104-160-190.cm.vtr.net> has joined #illumos
[14:50:17] *** patdk-lap <patdk-lap!~patrickdk@208.95.164.6> has quit IRC (Ping timeout: 240 seconds)
[14:53:09] *** patdk-lap <patdk-lap!~patrickdk@208.95.164.6> has joined #illumos
[15:13:19] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has joined #illumos
[15:16:07] *** BH23 <BH23!~BH23@193.117.206.132> has joined #illumos
[15:17:13] *** wl_ <wl_!~wl_@2605:6000:1b0c:5a5f::87c> has joined #illumos
[15:47:18] <andyf> igork, https://github.com/illumos/gcc/pull/31
[15:47:54] <igork> andyf: thanks
[16:05:14] *** kkantor <kkantor!~kkantor@c-24-118-59-107.hsd1.mn.comcast.net> has quit IRC (Quit: WeeChat 1.4)
[16:05:35] *** kkantor <kkantor!~kkantor@c-24-118-59-107.hsd1.mn.comcast.net> has joined #illumos
[16:13:13] <andyf> pwinder, I'm building a DEBUG image now with your patch, I'll let you know
[16:14:56] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has quit IRC (Ping timeout: 240 seconds)
[16:15:06] *** man_u <man_u!~manu@89-92-76-134.hfc.dyn.abo.bbox.fr> has joined #illumos
[16:19:00] *** rzezeski <rzezeski!uid151901@gateway/web/irccloud.com/x-iqsizkmttblueurh> has joined #illumos
[16:27:13] <clapont_> hi. does anyone have the time to tell why at "zpool status" I see some zpools with the device like "/dev/vx/dmp/stor0_0" and "others with the device like "/dev/vx/dmp/stor0_1s0" ? I created them with "zpool create /dev/vx/dmp/stor0_1"... I know there are 8 slices (0-7); for each Lun I did "format+select+label+quit". thank you very much.
[16:29:08] <Agnar_> clapont_: s0 will be appended if you added the device as a slice, if you use the whole disk, just leave out the slice
[16:29:33] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has joined #illumos
[16:31:18] <clapont_> Agnar_: thank you for answer! I wanted to use the whole LUN so I took the format's default, just "select/label/Y/quit" then I did "zpool create /dev/vx/dmp/stor0_1" - meaning the whole disk but....
[16:32:36] <Agnar_> that will take the whole device
[16:32:50] <clapont_> also "zpool import" works only wish "-d /dev/vx/dmp/stor0_1s0", the "-d /dev/vx/dmp/stor0_1" doesn't find the zpool
[16:33:13] <Agnar_> well, that's a speciality of vxdmp
[16:34:21] <Agnar_> check is there are device entries for stor0_1 in /dev/vx/dmp
[16:34:21] <clapont_> "vxdisk list stor0_1" is showing the c#t#d# -no slice at the end
[16:34:24] <Agnar_> I doubt
[16:35:04] <Agnar_> btw, it should be zpool import -d /dev/vx/dmp/
[16:35:05] <clapont_> in /dev/vx/dmp/ I see the /dev/vx/dmp/stor0_1 + /dev/vx/dmp/stor0_1s0 + ... /dev/vx/dmp/stor0_1s7
[16:35:27] <Agnar_> ok, try importing via zpool import -f /dev/vx/dmp
[16:35:49] <clapont_> if I do with "zpool import -d /dev/vx/dmp/" it will import it fine and will show the same "zpool status" as "/dev/vx/dmp/stor0_1s0"
[16:35:56] <Agnar_> and just as a hint...you should hire commercial Veritas Support
[16:36:27] <Agnar_> yes, because it probably finds slice0 first and slice0 has the whole disk
[16:37:28] <clapont_> the thing is working, I just wish to understand it deeper and I did not find the explanation in the VXadmin - maybe I missed it
[16:38:10] <clapont_> Agnar_: again, thank you for your time! if you know/can recommand something to read, I would be very happy
[16:44:19] <gitomat> [illumos-gate] 12013 fix smatch build for GCC4 -- John Levon <john.levon at joyent dot com>
[17:10:41] *** insomnia is now known as EricCiaramella
[17:11:11] *** EricCiaramella is now known as Guest69624
[17:11:16] <gitomat> [illumos-gate] 12009 Memory leaks in blkdev when blkdev device is detached -- Paul Winder <paul at winders dot demon.co.uk>
[17:11:26] *** Guest69624 is now known as LexLuthor
[17:21:42] *** tsoome <tsoome!~tsoome@148-52-235-80.sta.estpak.ee> has joined #illumos
[17:26:16] <sensille> any ideas why a program can run faster on linux than on illumos, on identical hardware, when it's running in cpu cache only and is identical assembly? around 10%
[17:28:25] <tsoome> sensille: scheduler
[17:29:18] <LeftWing> sensille: Try putting it in the realtime class?
[17:29:25] <sensille> it's running on a single cpu, reserved by a psrset
[17:29:35] <sensille> with nice -n -1000
[17:29:53] <LeftWing> Yeah, the realtime class is a much bigger hammer
[17:29:59] <tsoome> sensille: if one system is preferring (for some reason) the app while other is trying to give fair chance to all apps, there will be difference.
[17:30:00] <LeftWing> It puts it at a priority above the kernel basically
[17:30:26] <sensille> what would be the command for that?
[17:30:34] <LeftWing> Hang on I'll find my notes
[17:31:07] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has quit IRC (Remote host closed the connection)
[17:31:23] *** cpk <cpk!~chris@185.172.87.163> has joined #illumos
[17:31:42] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has joined #illumos
[17:32:08] <rmustacc> priocntl -s -c RT -i pid <pid>
[17:32:20] <tsoome> man priocntl :)
[17:32:26] <tsoome> ah:D
[17:32:30] <sensille> the start of the investigation are 2 machines running the exact same load, linux vs. illumos, 20% idle vs. 75% idle
[17:32:35] <sensille> thanks
[17:33:25] <LeftWing> If they're two separate machines you may want to try Linux on both, say, to make sure they both run at the same speed
[17:33:43] <LeftWing> We've seen issues in the past with, say, different levels of overclock/turboboost based on uneven heat in the DC
[17:33:46] <rmustacc> What are the BIOS options of both systems and where are they physically?
[17:34:01] <rmustacc> The Intel turbo stuff will vary dramatically, particularly if the rest of the system is idle and you're looking at single thread performance.
[17:34:12] <rmustacc> Of course, there may be more background work going on that takes you down from an all cores turbo.
[17:34:28] <sensille> 28 cores, at 80% busy vs. 25% busy
[17:34:31] <rmustacc> So I would honestly, recommend disabling turbo-boost in the BIOS on all systems under test and make sure that energy/power management is set.
[17:34:38] <sensille> it's dramatic
[17:34:40] <tsoome> how is the app started? from GUI? from ssh session?
[17:34:50] <sensille> init
[17:35:22] <sensille> i assume most of the overhead gets lost in the network stack
[17:35:37] <sensille> but i'm surprised that even the single thread performance varies by 10%
[17:36:06] <sensille> realtime class didn't change anything, which is not surprising, as the cpu was already exclusive to the process
[17:36:47] <LeftWing> You may want to look with DTrace to see if it's going off cpu while you're measuring it
[17:36:50] <LeftWing> just to confirm
[17:37:01] <sensille> ok
[17:37:37] <LeftWing> https://illumos.org/books/dtrace/chp-sched.html#chp-sched
[17:38:08] <Agnar_> sensille: btw, maybe a stupid shot: have you compared the powertop output and had a look at cpu frequency?
[17:38:34] <sensille> reported cpu frequencies are identical
[17:39:03] <sensille> that would have been so nice ...
[17:40:03] <LeftWing> The thing you're measuring doesn't make any system calls?
[17:41:03] <sensille> no
[17:41:49] <LeftWing> I look forward to seeing the solution to this mystery :D
[17:42:18] <sensille> no off-cpu events
[17:42:42] <sensille> double checked, when not run in the psrset, i see events
[17:43:02] <sensille> LeftWing: would it be less mysterious if it involved memory accesses?
[17:43:27] <sensille> (which it doesn't)
[17:44:17] <LeftWing> Memory accesses to pages that haven't been faulted in already could obviously induce some variance
[17:44:24] <Agnar_> sensille: can you share the code? (not that I can help much, but I am curious and I have a RHEL and an illumos laptop here)
[17:45:26] <sensille> Agnar_: much too simple, adapted from the sysbench test: https://pastebin.com/0Ttyn31i
[17:47:05] <sensille> i can live with linux being a bit faster, but a factor of 3 overall is hard to justify
[17:48:46] <Agnar_> sensille: so, just "time ./a.out", right?
[17:48:55] <sensille> yeah
[17:49:21] <tsoome> psrsets can cause interesting side effects btw
[17:49:25] <sensille> or time psrset -e 2 priocntl -c RT -e ./a.out
[17:50:24] <sensille> tsoome: ok, on the production machine it takes 50% longer without the psrset, even with RT class
[17:51:33] <Agnar_> sensille: ok, my illumos is a core i5-3320M @2.6GHz, my RHEL is a i5-2520M @2.5GHz, on illumos it takes 5.4s real/5.39s user; on RHEL it takes 6.306s real/6.296s user/0.001s sys
[17:51:51] <sensille> :)
[17:51:59] <Agnar_> which sounds reasonable
[17:52:00] <sensille> 3.2 vs. 3.5s here
[17:52:43] <sensille> but it's hard to compare non-identical cpus
[17:53:36] <LeftWing> Is the program text you're running on both CPUs in the hot path identical?
[17:53:52] <rmustacc> If it is and it's all cpu bound, the next place to look is cpu perf counters, fwiw.
[17:53:55] <LeftWing> Or could there be differences in compiler version, optimisation level, libraries in use, etc?
[17:54:30] <Agnar_> sensille: sure, but I would have expected same runtimes if it would have been a illumos problem per se
[17:54:31] <sensille> yes, identical disassembly, apart from a few register names
[17:55:54] <sensille> rmustacc: ok, perf counters. are they accessible on both linux and illumos?
[17:56:08] <sensille> wanted to learn that for a long time :)
[17:56:24] <rmustacc> Yes, they are.
[17:56:31] <rmustacc> Through different tools.
[17:56:37] <sensille> sure
[17:56:44] <LeftWing> man cpc and ... cputrack I think?
[17:56:57] <rmustacc> I would expect perf(1) on Linux and probably cpustat.
[17:57:06] <rmustacc> But just focus on the processor you're bound to.
[17:58:05] <tsoome> on separate note, anything on 11938 ?:)
[17:58:35] <rmustacc> I did finish my review, right?
[17:58:52] <sensille> ok, thanks all
[17:58:54] <tsoome> yes, I did post the RTI:)
[17:59:27] <LeftWing> I'll have a look, one sec
[17:59:30] <tsoome> just not sure about the email flow:)
[18:01:48] <LeftWing> tsoome: /code/11938/usr/src/boot/sys/boot/libstand/i386/core: ELF 64-bit LSB core file ╚AMD64 Version 1, from 'smatch'
[18:01:53] <LeftWing> What's that one :P
[18:02:34] *** pwinder <pwinder!~pwinder@86.2.210.254> has quit IRC (Quit: This computer has gone to sleep)
[18:02:34] <tsoome> I have seen few of those, yep
[18:03:00] <LeftWing> Can you make sure we get some of them where jlevon (or others!) can see them?
[18:03:23] <tsoome> I’ll do when I see one from current smatch
[18:03:32] <LeftWing> OK thanks
[18:04:31] <LeftWing> In addition to making sure this fixes the busted machine, you confirmed this still works on a machine that used to work too?
[18:04:52] <danmcd> I have three from you tsoome marked as unread. Was going to get to them later today but if LeftWing or someone else has cycles now, sure!
[18:05:11] <tsoome> yes:) also it is already integrated to fbsd
[18:05:31] <LeftWing> OK cool
[18:06:44] <tsoome> and to be really honest, I have another spin in process:)
[18:07:09] *** chrisBF <chrisBF!519d0501@host81-157-5-1.range81-157.btcentralplus.com> has joined #illumos
[18:12:25] *** pwinder <pwinder!~pwinder@86.2.210.254> has joined #illumos
[18:14:10] <gitomat> [illumos-gate] 11938 loader.efi: HARDDRIVE_DEVICE_PATH may have subpaths -- Toomas Soome <tsoome at me dot com>
[18:17:54] *** man_u <man_u!~manu@89-92-76-134.hfc.dyn.abo.bbox.fr> has quit IRC (Quit: man_u)
[18:20:26] <gitomat> [illumos-gate] 12012 zdb: do not output binary org.illumos:checksum_salt on terminal -- Toomas Soome <tsoome at me dot com>
[18:20:58] <danmcd> ping LeftWing--> I think src.illumos.org may be behind reality. Couldn't find efi_reserved_sectors() today, e.g.
[18:21:18] <LeftWing> huh
[18:21:24] * LeftWing looks
[18:21:29] <danmcd> nvm. Too new.
[18:22:01] <danmcd> Wait... not too new.
[18:22:08] <danmcd> (Late thanksgiving this year is messing with my sense of time.0
[18:23:05] <LeftWing> 17:20:30.414Z ERROR opengrok_listener: git pull failed (repo=/var/opt/opengrok/src/linux-master, stdout="")
[18:23:07] <LeftWing> stderr: error: cannot open .git/FETCH_HEAD: Too many open files
[18:23:09] <LeftWing> sigh!
[18:23:15] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[18:24:40] <tsoome> 32-bit git?
[18:25:10] <LeftWing> yessss
[18:25:14] * LeftWing shakes fist
[18:26:13] * LeftWing tries LD_PRELOAD_32=/usr/lib/extendedFILE.so.1
[18:29:06] <danmcd> Ooof.
[18:29:42] <LeftWing> OK that's better
[18:30:04] <LeftWing> It's reindexing now.
[18:30:46] <toasterson> LeftWing (IRC): which distro is running that? If it's OI we could switch git.
[18:30:55] <toasterson> to 64bit
[18:30:59] <LeftWing> It's old pkgsrc tbh
[18:31:22] <toasterson> Well there I can't help.
[18:31:25] <LeftWing> :D
[18:32:08] <toasterson> we have too many buildsystems tbh.
[18:32:30] <toasterson> Laets maek another build system to unify them all (xkcd reference)
[18:33:24] <toasterson> https://xkcd.com/927/
[18:39:23] <sensille> ok, perf on linux reports the program run at 3.2GHz, while cpu clock is reported as 2.6GHz on both
[18:39:53] <sensille> does illumos has to support the boost?
[18:43:55] <sensille> can't match the inst_retired counts yet
[18:44:24] <rmustacc> So, IIRC, turbo boost itself is all bios/fw/hw based.
[18:44:36] <rmustacc> Basically based on different heat/workload characteristics it'll dynamically change it.
[18:44:52] <rmustacc> It's been a while since I've looked at it in any detail.
[18:45:15] <sensille> so it's unfair to compare 2 machines that are on a different basic workload
[18:45:58] <sensille> and i need to check the bios settings
[18:46:56] *** BH23 <BH23!~BH23@193.117.206.132> has quit IRC (Ping timeout: 240 seconds)
[18:47:19] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Remote host closed the connection)
[18:47:41] <rmustacc> It can be challenging to get accurate results. In the past I once benchmarked KVM faster than bare-metal linux on a compute bound workload because one had turbo on and the other off.
[18:52:22] <sensille> i really need idle hardware for the investigation
[18:53:01] <gitomat> [illumos-gate] 11993 format: reserved partition size calculation should use sector size -- Toomas Soome <tsoome at me dot com>
[19:08:35] <clapont_> sensille: I am curious are you trying to argument a Illumos vs Linux decision? or you are trying to optimize your Illumos?
[19:11:37] *** clapont_ is now known as clapont
[19:16:24] <tsoome> getting turbo boost support would be nice:)
[19:17:21] *** arnoldoree <arnoldoree!~arnoldore@ranoldoree.plus.com> has joined #illumos
[19:18:27] <rmustacc> There shouldn't be anything we need to do other than visibility, IIRC.
[19:19:31] <rmustacc> There's no software support that should be required.
[19:19:44] <rmustacc> Certainly it's worked on systems in the past.
[19:24:50] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Read error: Connection reset by peer)
[19:25:06] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[19:26:31] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[19:27:48] *** neirac <neirac!~cneira@pc-184-104-160-190.cm.vtr.net> has quit IRC (Remote host closed the connection)
[19:28:06] *** neirac <neirac!~cneira@pc-184-104-160-190.cm.vtr.net> has joined #illumos
[19:28:48] *** mpana <mpana!~mpana@193.226.149.49> has quit IRC (Quit: ZNC 1.6.3+deb1ubuntu0.1 - http://znc.in)
[19:31:55] *** neirac <neirac!~cneira@pc-184-104-160-190.cm.vtr.net> has quit IRC (Read error: Connection reset by peer)
[19:32:08] *** neirac <neirac!~cneira@pc-184-104-160-190.cm.vtr.net> has joined #illumos
[19:56:52] <tsoome> visibility is important there indeed.
[19:57:41] <tsoome> rmustacc: I am a bit lost about https://code.illumos.org/c/illumos-gate/+/184/11 - do you mean we should aslo list the size of reserved partition?
[20:02:58] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[20:03:20] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[20:17:49] *** chrisBF <chrisBF!519d0501@host81-157-5-1.range81-157.btcentralplus.com> has quit IRC (Remote host closed the connection)
[20:36:13] *** neirac <neirac!~cneira@pc-184-104-160-190.cm.vtr.net> has quit IRC (Remote host closed the connection)
[20:36:39] *** neirac <neirac!~cneira@pc-184-104-160-190.cm.vtr.net> has joined #illumos
[20:51:34] *** neirac <neirac!~cneira@pc-184-104-160-190.cm.vtr.net> has quit IRC (Quit: Leaving)
[20:58:24] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[21:00:02] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[21:03:27] *** LexLuthor <LexLuthor!~insomnia@shadowcat/actuallyamemberof/lollipopguild.insomnia> has quit IRC (Read error: Connection reset by peer)
[21:18:37] *** amrfrsh <amrfrsh!~Thunderbi@185.212.171.68> has quit IRC (Ping timeout: 240 seconds)
[21:19:11] *** despair86 <despair86!despair86@75-60-137-0.lightspeed.rcsntx.sbcglobal.net> has joined #illumos
[21:20:36] *** sengir <sengir!~sh42@178.251.69.48> has quit IRC (Ping timeout: 240 seconds)
[21:23:20] *** amrfrsh <amrfrsh!~Thunderbi@185.212.171.68> has joined #illumos
[21:27:17] *** sengir <sengir!~sh42@178.251.69.48> has joined #illumos
[21:31:15] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has quit IRC (Remote host closed the connection)
[21:32:15] *** Yogurt <Yogurt!~Yogurt@104-7-67-228.lightspeed.sntcca.sbcglobal.net> has joined #illumos
[21:56:39] *** BH23 <BH23!~BH23@santoroj.plus.com> has joined #illumos
[22:07:48] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Ping timeout: 245 seconds)
[22:12:11] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has joined #illumos
[22:16:33] *** jcea <jcea!~Thunderbi@2001:41d0:1:8a82:7670:6e00:7670:6e00> has quit IRC (Ping timeout: 245 seconds)
[22:38:11] *** insomnia <insomnia!~insomnia@shadowcat/actuallyamemberof/lollipopguild.insomnia> has joined #illumos
[23:05:26] *** Kurlon_ <Kurlon_!~Kurlon@cpe-67-253-141-249.rochester.res.rr.com> has joined #illumos
[23:07:44] *** arnoldoree <arnoldoree!~arnoldore@ranoldoree.plus.com> has quit IRC (Quit: Leaving)
[23:09:03] *** Kurlon <Kurlon!~Kurlon@bidd-pub-03.gwi.net> has quit IRC (Ping timeout: 245 seconds)
[23:14:06] <rmustacc> tsoome: I think mostly I was trying to think through was software relying on the existing value that's there or not.
[23:18:44] <gitomat> [illumos-gate] 11962 Port ZoL zfs diff tests to illumos -- LOLi <loli10K at users dot noreply.github.com>
[23:42:52] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
[23:43:41] *** cpk <cpk!~chris@185.172.87.163> has quit IRC (Ping timeout: 246 seconds)
[23:45:47] *** BH23 <BH23!~BH23@santoroj.plus.com> has quit IRC (Ping timeout: 246 seconds)
top

   November 25, 2019  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >