Date: Mon, 18 Dec 2006 17:37:37 +0200 From: Achilleas Mantzios <achill@matrix.gatewaynet.com> To: freebsd-java@freebsd.org Subject: Re: Performance of Java on FBSD vs. others... Message-ID: <200612181737.39000.achill@matrix.gatewaynet.com> In-Reply-To: <B142E761BC54C0CA22CD9BFD@rambutan.pingpong.net> References: <20061110203714.GA89006@ace.b020.ceid.upatras.gr> <B142E761BC54C0CA22CD9BFD@rambutan.pingpong.net>
next in thread | previous in thread | raw e-mail | index | archive | help
=CE=A3=CF=84=CE=B9=CF=82 =CE=A4=CF=81=CE=AF=CF=84=CE=B7 12 =CE=94=CE=B5=CE= =BA=CE=AD=CE=BC=CE=B2=CF=81=CE=B9=CE=BF=CF=82 2006 17:25, =CE=BF/=CE=B7 Pal= le Girgensohn =CE=AD=CE=B3=CF=81=CE=B1=CF=88=CE=B5: > Hi! > > This mail thread died a while back. Any news, did you succeed in getting = it > faster? > > We are experiencing similar slowness with tomcat - it seems it runs much > faster on Linux. I have not done any tests yet, it is just a feeling, but > developers here are already moaning and want to replace FreeBSD with Linu= x. > I'd love to get the BSD boxes running as fast instead... Any input > appreciated. > Sorry, that i dont have smth specifically useful to say here, but i think that some kind of feedback would give a better feeling to all people in the list, that extra (cycological) support that could keep a person from making the final switch to linux. Raising such an important issue as the current thread, i think deserves something more than complete silence. If Nikos Ntarmos is around, anything related would be appreciated. Are these clock related issues worth intergrating into the patch? Any progress ? (I dont mean to sound like yet aonother "i want it all now, so you do it no= w=20 for me please" kind of newbie, i am writing this if nothing else to give some life to the thread). > Regards, > Palle > > --On fredag, november 10, 2006 22.37.14 +0200 Nikos Ntarmos > > <ntarmos@ceid.upatras.gr> wrote: > > Hi all. > > > > I've spent the last week or so running the exact same Java code on the > > exact same hardware, under different OSs. The code i'm running is > > heavily cpu-bound (no i/o other than light logging), uses a lot of > > memory (~1G), and does not rely on threading libraries (i.e. the program > > creates two threads when it starts and that's all about it). > > > > I'm running a small and a large version of the program (wrt memory > > requirements and run-time). The average execution time for the small > > version goes from ~200'' under win32, to ~220'' under linux, and to > > ~800'' under freebsd. For the large version, the numbers are ~1600'' > > for win32, ~1800'' for linux, and ~4700'' for freebsd. > > > > This is all on a 1.7GHz Pentium-M laptop with 1G RAM (dmesg at > > http://ntarmos.dyndns.org/Computers/dmesgs/ace.dmesg), running -current > > as of Sep 29. The JDKs tested are diablo-jdk15 (1.5.0.07.01) and > > sun-jdk15 (1.5.0p6) for FreeBSD, and sun-jdk15 (1.5.0_07) for linux and > > win32. > > > > Does anybody else see such a performance drop across OSs? Is there some > > specific (known) reason why freebsd lags behind the other two OSs by > > this 3x-4x factor? Any pointers as to some (sysctl or other) knob I've > > missed? > > > > Thanks in advance. > > > > \n\n > > _______________________________________________ > > freebsd-java@freebsd.org mailing list > > http://lists.freebsd.org/mailman/listinfo/freebsd-java > > To unsubscribe, send any mail to "freebsd-java-unsubscribe@freebsd.org" > > _______________________________________________ > freebsd-java@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-java > To unsubscribe, send any mail to "freebsd-java-unsubscribe@freebsd.org" =2D-=20 Achilleas Mantzios
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200612181737.39000.achill>