From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jonathan Larmour To: "Pamulapati, Indrasena (CTS)" Cc: ecos-discuss@sources.redhat.com Subject: Re: [ECOS] Performance comparision Date: Tue, 06 Feb 2001 06:55:00 -0000 Message-id: <3A801046.69FBFD1E@redhat.com> References: <8F1D166D08ACD41196EE00B0D020944BABA7C6@CTSINENTSXUB> X-SW-Source: 2001-02/msg00068.html "Pamulapati, Indrasena (CTS)" wrote: > > > Hi, > > > > I am the same guy who asked about the problem of having a very high tick > > rate for addressing very small time units. > > I now perfectly understand the futility of increasing the tick rate to > > very high value and have decided to use external timers to address small > > time frames. > > > > But just for comparision, I ported my application to VRTXsa a different > > RTOS. And I used the same high tick rate ( about 10000 ticks per second) > > and the application was running fine. I was able to acheive the bit rate > > of about 20Mbps that I was suppose to acheive. But in eCos with such high > > tick value I was able to acheive only 4Mbps. > > > > In reply to my previous mail I understood that if we use a very high tick > > value a lot of processor resources will be spent on processing the tick > > only and my application will suffer. But why am i not facing the same > > problem in VRTXsa? > > > > Is it a limitation of eCos? It's impossible to tell because it depends on how the driver and handler was coded. Since at worst eCos allows you to write a VSR in assembly to handle the interrupt, I would be surprised if performance could be bettered in that situation. Jifl -- Red Hat, Rustat House, Clifton Road, Cambridge, UK. Tel: +44 (1223) 271062 Maybe this world is another planet's Hell -Aldous Huxley || Opinions==mine