Port-vax archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Bountysource campaign for gcc-vax

On Sun, 2 May 2021, John Klos wrote:

> Thank you for this. I don't have much time to give this as much attention as
> it deserves, so this is the perfect amount of a how-to to at least get
> started. I'll give it a go, hopefully some time this week.

 I'm glad to be of help!

> -j 64? Must be nice ;)

 Quite so, though for a VAX cross-compiler build the speed up is only 
ninefold or so with `time' giving:

8896.42user 346.68system 7:18.98elapsed 2105%CPU (0avgtext+0avgdata 533760maxresident)k
0inputs+7809792outputs (16778major+18136652minor)pagefaults 0swaps

for `-j 64' vs:

3785.87user 263.91system 1:06:53elapsed 100%CPU (0avgtext+0avgdata 532608maxresident)k
0inputs+7257088outputs (293major+17848903minor)pagefaults 0swaps

for `-j 1' here.  Still it does make a huge difference.

> For NetBSD you can use `sysctl -n hw.ncpu`.

 Good to know, thanks!  Sadly lizzie only says:

$ sysctl -n hw.ncpu

On the positive side she also says:

$ uptime
 4:21AM  up 212 days,  9:33, 2 users, load averages: 0.02, 0.01, 0.00

after all that beating with GCC verification. :)  She tends to lose NTP 
synchronisation though as the system clock drifts very badly when the load 
becomes excessive, which I suppose is due to a kernel bug in setting the 
scheduling priority for `ntpd'.  It's now off by well over an hour.


Home | Main Index | Thread Index | Old Index