Date: Sun, 28 Nov 2021 19:42:16 -0600 From: Jason Bacon <bacon4000@gmail.com> To: Steve Kargl <sgk@troutmask.apl.washington.edu> Cc: freebsd-hackers@freebsd.org Subject: Re: Call for Foundation-supported Project Ideas Message-ID: <836e9e78-7220-c3aa-796c-7f31b23aae6f@gmail.com> In-Reply-To: <20211129003635.GA81568@troutmask.apl.washington.edu> References: <861r36xzpe.fsf@phe.ftfl.ca> <20211128220732.GA81140@troutmask.apl.washington.edu> <d0c77bfe-6a37-e177-f64d-2e1d3fc23dc2@gmail.com> <20211129003635.GA81568@troutmask.apl.washington.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On 11/28/21 18:36, Steve Kargl wrote:
> On Sun, Nov 28, 2021 at 05:55:32PM -0600, Jason Bacon wrote:
>> On 11/28/21 16:07, Steve Kargl wrote:
>>>
>>> % ps -ww -p 77387
>>> PID TT STAT TIME COMMAND
>>> 77387 2 R 40:34.67 c++
>>>
>>> 40 minutes for 1 file with many exceeding 30 minutes seems a tad bit
>>> excessive.
>>
>> I would bet you have a hardware issue. I've never seen a buildworld take
>> more than a few hours and that was on some very old hardware.
>>
>
> It's certainly not the latest and greatest,
> CPU: Intel(R) Core(TM)2 Duo CPU T7250 @ 2.00GHz (1995.04-MHz K8-class CPU)
> ...
> ada0: <Patriot Burst SBFM91.0> ACS-4 ATA SATA 3.x device
> ada0: Serial Number 1B0607771A0800257271
> ada0: 300.000MB/s transfers (SATA 2.x, UDMA6, PIO 8192bytes)
>
> but this laptop has built FreeBSD for several years and each
> year and with new import of llvm the buildworld times go up,
> up, up...
>
> buildworld is essentially the only thing running on it.
>
> last pid: 4921; load averages: 2.17, 2.12, 2.29; b up 4+22:40:32 16:35:53
> 78 processes: 3 running, 75 sleeping
> CPU: 0.2% user, 98.0% nice, 1.4% system, 0.4% interrupt, 0.0% idle
> Mem: 1045M Active, 1543M Inact, 11M Laundry, 776M Wired, 395M Buf, 575M Free
> Swap: 3881M Total, 84M Used, 3798M Free, 2% Inuse
>
> PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND
> 4918 root 1 106 4 220M 153M RUN 1 2:17 97.05% c++
> 4763 root 1 106 4 1140M 924M CPU0 0 13:29 95.20% c++
> 4921 kargl 1 20 0 14M 3392K CPU1 1 0:00 1.47% top
> 1018 kargl 3 21 0 169M 23M select 0 50:22 1.14% Xorg
> 4889 kargl 1 20 0 22M 8636K select 0 0:00 0.16% xterm
>
CPU utilization close to 100% would pretty much rule out a disk
bottleneck, assuming it stays at that level most of the time. I would
watch it closely for a while, as well as the swap stats. I would guess
the compiler WCPU will plummet at some point if it's taking half an hour
to compile one file.
--
All wars are civil wars, because all men are brothers ... Each one owes
infinitely more to the human race than to the particular country in
which he was born.
-- Francois Fenelon
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?836e9e78-7220-c3aa-796c-7f31b23aae6f>