[Info-vax] INIT/HIGHWATER

[abrsvc]

On Feb 6, 1:33 am, hel... at astro.multiCLOTHESvax.de (Phillip Helbig---
undress to reply) wrote:
> In article
> <0fad844f-459a-4586-a6f3-e5a810d35... at k30g2000yqb.googlegroups.com>,
> Hein RMS van den Heuvel <heinvandenheu... at gmail.com> writes:
>
> > On Feb 5, 8:57 am, hel... at astro.multiCLOTHESvax.de (Phillip Helbig---
> > undress to reply) wrote:
> > > How much of a performance hit is /HIGHWATER?
>
> > > What about /ERASE?
>
> > Both are high overhead option
>
> OK.
>
> > I woudl leave both of for a hobby environment and even for a well
> > understood/controlled production environment.
>
> That's what I was thinking.  INIT/NOERASE is the default, but
> INIT/HIGHWATER is the default.  My guess is that many folks have the
> latter without needing it.
>
> Apart from /HEADERS=<50,000/GB>, perhaps I should add /NOHIGHWATER to my
> personal default INIT command.

THe main difference is WHEN the actual erase happend. With highwater
turned on, any block that is allocated will be erased BEFORE being
passed to the application.  WIth erase, the erasure will occur upon
release of the block.  A quick test will show this;

- Set the volume/highwater then execute the following sequence and
time it.
MC sysgen
sysgen> create testfile.dat/siz=10000


- Now turn off highwater and try it again.  You will notice a
difference.

The timeing difference will be similar for an erase function upon
delete or release.

Dan