[Mono-dev] Mono GC tuning

Vlad Brezae vlbrez at microsoft.com
Thu Dec 22 19:42:15 UTC 2016


Hey Greg,

       Aside from the smaller major pause times,  the concurrent collector is typically faster since it’s doing part of its job concurrently with the mutator (which likely doesn’t use the cpu at 100%). You can see our results on our benchmarks at [1]. You should be fine with using it on 4.6 and newer.
 
        You should definitely consider using higher nursery sizes if most objects die young.

Vlad
       

[1] http://open.xamarin.com/benchmarker/front-end/compare.html#ids=17099+17101

On 22/12/2016, 16:34, "Mono-devel-list on behalf of Greg Young" <mono-devel-list-bounces at lists.dot.net on behalf of gregoryyoung1 at gmail.com> wrote:

    Thanks Alex, any idea on the throughput trade offs with concurrent?
    Obviously I will measure this but would be interested to see any
    benchmarks etc others may have.
    
    On Thu, Dec 22, 2016 at 12:47 PM, Alex Rønne Petersen <alex at alexrp.com> wrote:
    > Hi,
    >
    > The concurrent collector should be fairly stable on desktop at least.
    > It's what we use on CI.
    >
    > Regards,
    > Alex
    >
    > On Thu, Dec 22, 2016 at 1:26 PM, Greg Young <gregoryyoung1 at gmail.com> wrote:
    >> I want to spend a couple of days tuning our mono GC usage. What things
    >> should I be looking to tweak and what things should I be monitoring?
    >> Currently I have been looking at GC pauses via the profiler API.
    >>
    >> As most of our allocations are short lived (and long lived ones mostly
    >> live for the life time of the process) I would imagine trying to tune
    >> nurseries to be the right size is the main place for improvement? Also
    >> is concurrent/multi-threaded GC stable as of this point?
    >>
    >> Cheers,
    >>
    >> Greg
    >>
    >> --
    >> Studying for the Turing test
    >> _______________________________________________
    >> Mono-devel-list mailing list
    >> Mono-devel-list at lists.dot.net
    >> https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.dot.net%2Fmailman%2Flistinfo%2Fmono-devel-list&data=02%7C01%7Cvlbrez%40microsoft.com%7C5667af3e98744a9db03208d42a77b85f%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636180141074077788&sdata=WsW9iEV2dhYJx8BbjUx2ZLTnxYVLWDV3xOsGr%2Fq6PTM%3D&reserved=0
    
    
    
    -- 
    Studying for the Turing test
    _______________________________________________
    Mono-devel-list mailing list
    Mono-devel-list at lists.dot.net
    https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Flists.dot.net%2Fmailman%2Flistinfo%2Fmono-devel-list&data=02%7C01%7Cvlbrez%40microsoft.com%7C5667af3e98744a9db03208d42a77b85f%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636180141074077788&sdata=WsW9iEV2dhYJx8BbjUx2ZLTnxYVLWDV3xOsGr%2Fq6PTM%3D&reserved=0
    



More information about the Mono-devel-list mailing list