Re: [PATCH v8 0/8] mm: zswap swap-out of large folios

[Yosry Ahmed <yosryahmed@google.com>]

[..]
> > > Performance Testing:
> > > ====================
> > >
> > > We list the data for 64K folios with before/after data per-compressor,
> > > followed by the same for 2M pmd-mappable folios.
> > >
> > >
> > >  -------------------------------------------------------------------------------
> > >  64K folios: zstd:
> > >  =================
> > >
> > >  zswap compressor                zstd          zstd         zstd      zstd v8
> > >                          before-case1  before-case2        after     vs.    vs.
> > >                                                                    case1  case2
> > >  -------------------------------------------------------------------------------
> > >  Total throughput (KB/s)    5,222,213     1,076,611    6,227,367     19%    478%
> > >  Average throughput (KB/s)    174,073        35,887      207,578     19%    478%
> > >  elapsed time (sec)            120.50        347.16       109.21      9%     69%
> >
> >
> > The diff here is supposed to be negative, right?
> > (Same for the below results)
>
> So this is supposed to be positive to indicate the throughput improvement
> [(new-old)/old] with v8 as compared to the before-case1 and before-case2.
> For latency, a positive value indicates the latency reducing, since I calculate
> [(old-new)/old]. This is the metric used throughout.
>
> Based on this convention, positive percentages are improvements in both,
> throughput and latency.

But you use negative percentages for sys time, we should at least be
consistent with this.