Re: NetBSD/vax 11.0RC1

On Mon, Feb 16, 2026 at 11:46:54AM -0500, emanuel stiebler wrote:
> On 2026-02-16 11:33, emanuel stiebler wrote:
> > On 2026-02-16 11:19, Ken Wellsch wrote:
> > 
> > > Maybe this is actually working correctly and I'm just not letting it
> > > run (thrash) along long enough :)
> > 
> > OK, let it run longer, and as Martin suggests, just one thread
> > (my is running since yesterday on the gimp-match-***, so prbably it just
> > takes forever ...)
> > 
> > BTW, you have 1GB swap?
> > 
> OK, I just tried "top" here, while it is running the gimp-match***
> 
> load averages:  1.15,  1.19,  1.08;               up 1+23:02:15
> 11:42:55
> 34 processes: 2 runnable, 31 sleeping, 1 on CPU
> CPU states: 93.4% user,  0.0% nice,  6.6% system,  0.0% interrupt,  0.0%
> idle
> Memory: 63M Act, 31M Inact, 9076K Wired, 18M Exec, 9936K File, 1948K Free
> Swap: 1024M Total, 115M Used, 909M Free / Pools: 12M Used / Network:
> 
>   PID USERNAME PRI NICE   SIZE   RES STATE       TIME   WCPU    CPU COMMAND
>  4679 emu       31    0   147M   72M RUN        37:10 90.63% 90.63% cc1plus
>  4328 root      85    0  4924K 1304K wait        0:00 11.00%  0.54% sh
>  1700 emu       85    0    14M 4876K poll        0:01  0.05%  0.05%
> sshd-session
> 
> Shouldn't "top" report the full memory of the system?
> 
> I'm wondering about the "Memory: 63M Act", I have 128MB in the system?
> (I'm missing something here obviously :) )

Yes, last fall I tried with 11-BETA and I didn't capture as much data
(actually just the final "top" screen views).

I see that I first tried to run the emulation with 128M (since that is
the most for which I have any real hardware) and this was what that
showed:

load averages:  1.00,  1.00,  1.00;               up 0+06:14:40    20:01:31
20 processes: 1 runnable, 18 sleeping, 1 on CPU
CPU states:  0.2% user,  0.0% nice, 99.8% system,  0.0% interrupt,  0.0% idle
Memory: 68M Act, 33M Inact, 11M Exec, 8376K File, 1008K Free
Swap: 1024M Total, 275M Used, 749M Free / Pools: 14M Used / Network: 1K In

  PID USERNAME PRI NICE   SIZE   RES STATE       TIME   WCPU    CPU COMMAND
23400 root      81    0   319M   92M RUN        35:15 69.43% 69.43% cc1plus
    0 root     126    0     0K 5684K pgdaemon    0:57  0.93%  0.93% [system]
17701 root      81    0    11M    0K wait        0:15  0.00%  0.00% nbgmake
 3580 root      81    0  5728K    0K wait        0:03  0.00%  0.00% nbgmake
 1401 root      85    0    12M    0K wait        0:01  0.00%  0.00% login
20017 root      79    0  4588K    0K wait        0:01  0.00%  0.00% nbmake
14330 root      43    0  5256K  188K CPU         0:00  0.00%  0.00% top
  167 kwellsch  86   -2  5888K    0K wait        0:00  0.00%  0.00% su
 1198 root      85    0    12M    0K wait        0:00  0.00%  0.00% login
 1725 kwellsch  85    0  5036K    0K wait        0:00  0.00%  0.00% sh
 1574 root      85    0  5036K    0K wait        0:00  0.00%  0.00% sh
  813 root      85    0  5036K    0K wait        0:00  0.00%  0.00% sh
 1496 root      85    0  5032K    0K wait        0:00  0.00%  0.00% sh
    1 root      85    0  4980K    0K wait        0:00  0.00%  0.00% init
 1302 root      83    0  5032K    0K wait        0:00  0.00%  0.00% sh
20970 root      81    0  7800K    0K wait        0:00  0.00%  0.00% c++
 1847 root      80    0  5596K    0K wait        0:00  0.00%  0.00% nbmake
20329 root      80    0  5596K    0K wait        0:00  0.00%  0.00% nbmake
12859 root      80    0  4572K    0K wait        0:00  0.00%  0.00% nbmake
15404 root      79    0  5056K    0K wait        0:00  0.00%  0.00% sh


It looks like it is still creeping along but then you notice the 100%
system time ... now I may have thought, oh why did I do that, it will
be thrashing so badly on swap, it is burning my SSD into the ground.

So I restarted again I see after and just reran the build (prior completed
work left as is to speed things along) but with 512M this time ... you
get to see the same thing I captured better this weekend with RC1.  But
I know I saw this same thing with 10.1 when it came out. ... same gimple-
match.cc file ...

load averages:  1.00,  0.97,  0.85;               up 0+00:37:44    07:46:29
20 processes: 1 runnable, 18 sleeping, 1 on CPU
CPU states:  0.2% user,  0.0% nice, 99.8% system,  0.0% interrupt,  0.0% idle
Memory: 226M Act, 111M Inact, 18M Exec, 34M File, 133M Free
Swap: 1024M Total, 1024M Free / Pools: 18M Used / Network: 

  PID USERNAME PRI NICE   SIZE   RES STATE       TIME   WCPU    CPU COMMAND
 9403 root      79    0   316M  279M RUN        25:30 50.34% 50.34% cc1plus
    0 root     126    0     0K 8228K pgdaemon    0:49  0.00%  0.00% [system]
11044 root      80    0  9828K    0K wait        0:16  0.00%  0.00% nbgmake
 1226 root      82    0  5728K    0K wait        0:03  0.00%  0.00% nbgmake
  762 root      85    0    12M    0K wait        0:01  0.00%  0.00% login
  687 root      85    0    12M    0K wait        0:01  0.00%  0.00% login
 1722 root      81    0  4588K    0K wait        0:01  0.00%  0.00% nbmake
 2520 root      43    0  5256K  224K CPU         0:00  0.00%  0.00% top
  933 kwellsch  86   -2  5888K    0K wait        0:00  0.00%  0.00% su
  759 root      85    0  5036K    0K wait        0:00  0.00%  0.00% sh
  914 root      85    0  5036K    0K wait        0:00  0.00%  0.00% sh
  880 kwellsch  85    0  5036K    0K wait        0:00  0.00%  0.00% sh
  888 root      85    0  5032K    0K wait        0:00  0.00%  0.00% sh
    1 root      85    0  4980K    0K wait        0:00  0.00%  0.00% init
  445 root      84    0  5032K    0K wait        0:00  0.00%  0.00% sh
 1141 root      82    0  4572K    0K wait        0:00  0.00%  0.00% nbmake
 1222 root      81    0  5596K    0K wait        0:00  0.00%  0.00% nbmake
 3185 root      81    0  5596K    0K wait        0:00  0.00%  0.00% nbmake
 8575 root      81    0  5056K    0K wait        0:00  0.00%  0.00% sh
10729 root      80    0  7800K    0K wait        0:00  0.00%  0.00% c++