tag:blogger.com,1999:blog-4183569809600611634.post7685179842658217355..comments2024-01-24T00:45:45.524-08:00Comments on Sancho GGP Player: Sancho goes greenSteve Draperhttp://www.blogger.com/profile/06341780555576603467noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-4183569809600611634.post-82342872588906905772014-08-18T01:43:04.895-07:002014-08-18T01:43:04.895-07:00I think stack-based GC roots are identified for al...I think stack-based GC roots are identified for all collection types.<br /><br />G1 doesn't have such a black & white approach to stop-the-world vs concurrent collection. In most collection cycles, there are some parts where everything is stopped and some concurrent parts.<br /><br />Identifying GC roots in stack frames always requires a pause and since it's done in every collection, it's important not to have very deep stacks.Andrew Rosehttps://www.blogger.com/profile/12334890433091548139noreply@blogger.comtag:blogger.com,1999:blog-4183569809600611634.post-13052379256677203842014-06-19T07:51:08.264-07:002014-06-19T07:51:08.264-07:00Always good to hear more about sancho dev - thanks...Always good to hear more about sancho dev - thanks guys for sharing. Seems since end of coursera I've spent most my time trying to fix timeouts via eliminating allocations - today I introduced some new code for simultaneous games where I thought surely a few array creations won't make a difference, and was subsequently gifted a nice 198 seconds garbage collection during chinook! :) <br />Do you know if gc-roots on the stack are followed for incremental collections, or just full 'stop the world' ones?Richard Emsliehttps://www.blogger.com/profile/15336892000391057507noreply@blogger.com