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

Re: Change in package architectures list.

Michael Schmitz <schmitz@mail.biophys.uni-duesseldorf.de> writes:

>> > Depending on the access pattern, a ratio of 2:1 swap to ram would be
>> > painful. The only machine where I could try that right away only has
>> > 256 MB RAM... and it's building gcc-snapshot right now.
>> I don't really know how much RAM might be required for a reasonable
>> gcc build time.  I only know a bit about the VM size.
> How do you get the peak VM usage data, then? That's been baffling me
> since some time. Anyway, I'll give it a try one way or other.

I didn't do anything very precise.  I just watched the build (via top,
htop, or proc -- I forget), and it looked like the virtual memory
footprint went a bit over 1GB.  So I know that the package isn't going
to build on any machine that has less than that amount of RAM and

However, just knowing that the virtual memory footprint is at least
1GB doesn't really indicate how much physical RAM might be needed to
avoid excessive swapping (and build times).

So basically, all I have is a rough estimate, and I also don't know
how gcc's memory usage might vary across platforms.

>> > I'll give it a shot on the 512 MB box as soon as we're largely clear
>> > of the backlog.
>> Before you can try that I'll have to generate a new source package,
>> one that includes pre-generated C for arm.  Do you have any idea how
>> long it might be before you would be able to try the build?
> I wish I knew - the current build runs for a week now, and it could well
> go on for a few days yet.

OK.  Are you talking about m68k or arm?

Rob Browning
rlb @defaultvalue.org and @debian.org; previously @cs.utexas.edu
GPG starting 2002-11-03 = 14DD 432F AE39 534D B592  F9A0 25C8 D377 8C7E 73A4

Reply to: