http://mailman.uclinux.org/pipermail/uclinux-dev/2003-September/020915.htmlon 9/7/03 7:49 PM, Hyok S. Choi at hyok.choi at samsung.com wrote: > Hi greg, > > Please call me Hyok. ^^ > I'm using JTAG emulators for runtime measuring.(T32) > First of all, I patched equal code for S3C2410 (logically) for Linux and > uClinux both. (2.4.21) and took step by step runtime analysis in > function base. > So, the main benchmark code was Linux Kernel. The key difference in > linux kernel boot-time was "mem_init" function, tough. > I have a plan to release a document for uClinux and Linux kernel boot > time analysis and optimization of both. (greg, give me an advice for > appropriate place to release) > > Thanks for your reading, > Hyok > > <EOT> > CHOI, HYOK-SUNG > Engineer (Linux System Software) > S/W Platform Lab, Digital Media R&D Center > Samsung Electronics Co.,Ltd. > tel: +82-31-200-8594 fax: +82-31-200-3427 > e-mail: hyok.choi at samsung.com <mailto:hyok.choi at samsung.com> > > [compile&run] > main(a){printf(a,34,a="main(a){printf(a,34,a=%c%s%c,34);}",34);} > > > > -----Original Message----- > From: uclinux-dev-bounces at uclinux.org > [mailto:uclinux-dev-bounces at uclinux.org] On Behalf Of Greg Ungerer > Sent: Monday, September 08, 2003 9:43 AM > To: uClinux development list > Subject: Re: [uClinux-dev] Re: MMU vs non MMU > > Hi Hyok S. Choi, > > Hyok S. Choi wrote: >> I've developed and optimized Linux and uClinux kernels for many >> platforms and several ARM chips include s3c2800, s3c3410, s3c2410, >> s5c7375, and so on. >> >> One of them(s3c2410) was to be both ported, Linux Kernel(w/MMU) AND >> uClinux Kernel(wo/MMU). >> >> It is based on 200MHz ARM920T core, and embeds several peripherals, > with >> I and D caches, and used in various applications include PDA, phone, >> network. >> >> The result was "the same", for both kernels, for plain codes, except > for >> some paging initializing codes in kernel in booting phase (Linux > kernel >> was faster). It is not from performance reason, but seems cache > effect, >> though I didn't look carefully for that, yet. > > What tests did you perform? > What benchmarks did you use? > > Regards > Greg |