Re: [isabelle-dev] Slowdown of ConcurrentGC

2018-05-17 Thread Florian Haftmann
Hi Makarius, > I have studied this a bit further. The interval > 75130777ece4:d45b78cb86cf contains more than just the string type > change, but it is still the main thing. Here are also other sessions in > similar intervals 75130777ece4:0a6d6ba383dc, 75130777ece4:f76e8180c498: > > * faster:

Re: [isabelle-dev] Slowdown of ConcurrentGC

2018-05-12 Thread Makarius
On 12/05/18 00:01, Makarius wrote: > The Isabelle/AFP timing charts show a recent slowdown of ConcurrentGC: > > https://isabelle.sketis.net/devel/build_status/AFP_slow_64bit_6_threads/index.html#session_ConcurrentGC > > See also the attached copies of the chart and data file (the web version >

[isabelle-dev] Slowdown of ConcurrentGC

2018-05-11 Thread Makarius
The Isabelle/AFP timing charts show a recent slowdown of ConcurrentGC: https://isabelle.sketis.net/devel/build_status/AFP_slow_64bit_6_threads/index.html#session_ConcurrentGC See also the attached copies of the chart and data file (the web version updates itself dynamically). At the transition