Re: mt performance and packaging

[Sam S. <sd...@gn...>]

Vladimir Tzankov wrote:
> On 7/28/09, Reini Urban <ru...@x-...> wrote:
>> mt performance on cygwin esp with lot of IO or symbols is horribly slow.
> 
> Can you check how spinlock_t is defined in clisp.h?
> I suspect that it fall backs to plain mutex on cygwin (such huge
> difference in performance is not likely to be caused by IO).

Could the problem be that every intern now calls mutex_lock?