<html xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns="http://www.w3.org/TR/REC-html40"> <head> <META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=us-ascii"> <meta name=Generator content="Microsoft Word 11 (filtered medium)"> <style>  </style> </head> <body lang=EN-US link=blue vlink=purple> <div class=Section1> Hi All,<o:p></o:p>    I am using Gstreamer for creating one of my video streaming application. My application creates multimedia session and then stream around 200 video sessions (a very processor intensive job).In doing so, the no. of threads created is around 1500 and I have given around 1536 KB as my stack size for each of these threads. I have one Quadcore(intel Xeon) system of around 8 GB of RAM(64 bit) with RHEL 5.5(kernel - 2.6.18-92.el5).<o:p></o:p>   After few hours of testing I see that my overall memory(RAM) is eating up. I did checked the memory and I saw that my cache size is growing. So far so good, as I can understand that system does cache instead of leaving some RAM free.<o:p></o:p>    My problem comes, when I see that after almost 20-25 hours of testing, RAM is continuously being eaten up and maximum part of that is going into cache. If I leave the system for few more hours I see whole of my memory is consumed up and now my swap started getting used. If I check the cache now, I clearly see most of RAM is now in form of cache. I still didn’t stop my application and allowed it run for few more hours and after sometime I see that process give segmentation fault while creating the pthread (create_pthread::unknown error 513).I also tried to observe the “core” file of application created in this way , but it doesn’t’ show anything. <o:p></o:p>   One thing I must mention that, in “top” I see almost constant values of resident memory, virtual memory and shared memory. There is no increase of memory usage here. So here I can assume that my application is not leaking the memory. I also observed that if I kill my application any time during this, the memory doesn’t go into free pool.<o:p></o:p> <o:p> </o:p>    Anyone, here can give any pointer for this kind of behavior. Is it can be because of some kernel BUG which is not freeing the memory to application usage.<o:p></o:p> No need to mentioned that any help or clue will be highly appreciated and regardedJ<o:p></o:p> <o:p> </o:p> <o:p> </o:p> Other Info regarding my system:<o:p></o:p> Linux testing 2.6.18-92.el5 #1 SMP Tue Apr 29 13:16:15 EDT 2008 x86_64 x86_64 x86_64 GNU/Linux<o:p></o:p>  <o:p></o:p> Gnu C                  4.1.2<o:p></o:p> Gnu make               3.81<o:p></o:p> binutils               2.17.50.0.6-6.el5<o:p></o:p> util-linux             2.13-pre7<o:p></o:p> mount                  2.13-pre7<o:p></o:p> module-init-tools      3.3-pre2<o:p></o:p> e2fsprogs              1.39<o:p></o:p> quota-tools            3.13.<o:p></o:p> PPP                    2.4.4<o:p></o:p> Linux C Library        > libc.2.5<o:p></o:p> Dynamic linker (ldd)   2.5<o:p></o:p> Procps                 3.2.7<o:p></o:p> Net-tools              1.60<o:p></o:p> Kbd                    1.12<o:p></o:p> Sh-utils               5.97<o:p></o:p> udev                   095<o:p></o:p> Modules Loaded         autofs4 hidp rfcomm l2cap bluetooth sunrpc ipv6 xfrm_nalgo crypto_api dm_mirror dm_multipath dm_mod video sbs backlight i2c_ec button battery asus_acpi acpi_memhotplug ac lp i2c_i801 ide_cd i2c_core shpchp e1000 tg3 cdrom parport_pc parport i5000_edac serio_raw edac_mc sg pcspkr ata_piix libata aacraid sd_mod scsi_mod ext3 jbd uhci_hcd ohci_hcd ehci_hcd<o:p></o:p> <o:p> </o:p> <o:p> </o:p> Best Rehards,<o:p></o:p> Parveen Jain<o:p></o:p> </div> </body> </html>