PATCH/RFC Pool allocations of QStringData objects..

[Maks Orlovich]

On Saturday 18 January 2003 11:25 am, Maks Orlovich wrote:
> > This number of course, deserves a pause. This means, if it's correct,
> > that we have created 100,000 QStringData and hence QString objects during
> > the startup! And subtracting frees from allocs, we have some 26,784
> > strings in memory!
>
> <snip>.

<Sigh>. I'll get it right. Sometime. One of those days. Yes, really. 
(The last numbers were run with glibc-2.3.1 malloc, not KDE malloc).

Base overhead:
KCounterProf::Cumulative: alloc = 3.62
KCounterProf::Cumulative: free = 2.57

KCounterProf::Cumulative: alloc = 3.87
KCounterProf::Cumulative: free = 2.47

KCounterProf::Cumulative: alloc = 3.65
KCounterProf::Cumulative: free = 2.65


KCounterProf::Cumulative: alloc = 3.52
KCounterProf::Cumulative: free = 3.02

Total (min): 6.3

Basic:

KCounterProf::Cumulative: alloc = 9.58
KCounterProf::Cumulative: free = 5.82

KCounterProf::Cumulative: alloc = 9.55
KCounterProf::Cumulative: free = 5.86

KCounterProf::Cumulative: alloc = 9.62
KCounterProf::Cumulative: free = 5.81

KCounterProf::Cumulative: alloc = 9.63
KCounterProf::Cumulative: free = 5.65

Total (min) : 15.28
w/o overhead: 8.98

Pooled:

KCounterProf::Cumulative: alloc = 5.37
KCounterProf::Cumulative: free = 2.82

KCounterProf::Cumulative: alloc = 5.71
KCounterProf::Cumulative: free = 2.81

KCounterProf::Cumulative: alloc = 5.33
KCounterProf::Cumulative: free = 2.78

KCounterProf::Cumulative: alloc = 5.35
KCounterProf::Cumulative: free = 2.81

Total(min):   8.16
w/o overhead: 1.86


Delta: 7.12ms, 79%.

Not much, but every bit helps..

(Thinking thoughts about how this would be like 70ms on a Pentium 100).