On 10/05/18 00:10, Makarius wrote:
> In the past couple of weeks I have sporadically tried to work around
> this resource problem, but failed so far.
> The latest attempt is Isabelle/f6a22490cca8. As usual, it "works for me
> on my usual test machines", but there is a remaining chance of problems
> coming back on other configurations.
So far this change looks good on various test machines with only 2 cores.
In Isabelle/a8f40dd73c61 I have made the parallelism a little bit more
ambitious, and keep watching resulting performance charts at
All this are just workarounds to get back into a situation where all
sessions work routinely most of the time.
isabelle-dev mailing list