[Gcl-devel] Re: Nqthm times worsening under 2.7.0?

[Camm Maguire]

Greetings!  It is entirely possible that the recent type propagation
changes  has temporarily introduced a performance anomaly.  As always,
pinning this down to a specific function is most helpful and will be
addressed quickly, time permitting.

There is still some compiler type work ahead, so 2.7.0 is still
somewhat in a state of flux.  I'll describe the goals if anyone is
interested. 

Take care,


Robert Boyer <address@hidden> writes:

> Nqthm times seem down about 8% from two months ago:
> 
>    2.7.0 Yesterday:
> 
>    real time       :   2374.030 secs
>    run-gbc time    :   2099.820 secs
>    child run time  :    152.600 secs
>    gbc time        :     52.010 secs
> 
>    2.7.0 Dec 13, 2005
> 
>    real time       :   2180.760 secs
>    run-gbc time    :   1935.740 secs
>    child run time  :    149.420 secs
>    gbc time        :     51.450 secs
> 
> No idea whether this is noise, cache mysteries, or what.
> 
> Bob
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah