To be honest, we've not done anything in depth recently. Last time I tried with GC1 (which is probably about 18 months ago now), I found it to be a bit slower than the CMS we use in the startup scripts... however the performance of different collectors is probably use case dependent... and it's also likely that v6 brokers and later behave quite differently to earlier brokers (due to the way direct buffers are used - and pooled).
If you do any work in this area, I'd be very interested in the results. -- Rob On 10 February 2017 at 00:00, Ramayan Tiwari <[email protected]> wrote: > Hi All, > > Has anyone done any perf testing around using different GC algorithms with > the Java broker or is there any recommendation on that? > > Thanks > Ramayan >
