Re: [LTP] Re: VM test on 2.4.14pre3aa2 (compared to 2.4.14pre3aa1)

[Paul L. <pl...@au...>]

On Mon, 2001-10-29 at 00:47, Andrea Arcangeli wrote:
> I'm looking into optimizing this test. While it probably doesn't affect
> the above numbers given the bytes allocated are quite similar, the
> benchmark is not reliable, if you want to use it as benchmark you should
> apply this patch first to make sure to compare apples to apples (not to
> oranges). For example, without those fixes it allocates only 20mbytese
> of ram here so it cannot swapout despite I use -p 80, because it
> considers only the freeswap and freememory but on any real load all the
> free memory will be allocated in cache most of the time.
> 
> the bench in short measure how fast we can push stuff to disk (just the
> swapout, no the swapins).
This is true, however the intention of the -p flag was to allow you to
make the test crank memory utilization up to a certain percent including
what is already allocated, so if I wanted to make it allocate 90% of my
memory and 20% was already allocated, I could safely do a -p90 and it
wouldn't oom and fail the test because it tried to allocate more than I
wanted it to.  It also makes a convenient OOM test by just passing it
-p101.

That being said, I do think you make a valid point with this patch, I
just don't want it to be the default.  It would be nice to make that an
optional behaviour.  Were there other changes/optimizations you wanted
to make to this test Andrea, or was this change the one you were talking
about?

Thanks for taking a look at this test, and if you have other memory
tests that you think might benefit the LTP, please feel free to
contribute them! :)

Thanks,
Paul Larson
Linux Test Project