https://www.dropbox.com/s/ltnedy59gv5i70j/VID_20140723_084207.mp4
This is what I have done in my iMX6 solo @800M 16bit ddr board 2014-11-12 11:22 GMT+08:00 liyaoshi <[email protected]>: > This is depends on your SOC main clock and cache size and nand /eMMC read > speed > If on 512K L2 cache and over than 1G ， lzo MAYBE will beat uncompressed > but in less than 500M and with very fast nand 16bit EDO ddr mode /eMMC > 50M@8bit with ADMA > what I get test result is raw (uncompressed) kernel image will be the > better choice on iMX6 > > You can see the test result from the link , > > None69656444.62674904.626749Ti dont have eDMA and ADMA enable as default > in previous u-boot driver > While this is patched in 2014.04 version on omap5 and dra7x driver > It wont take over than 1s while loading 6M size kernel image > What I test on imx6 , u-boot with ADMA driver enable on eMMC 4.4 @8bit 50M > Will over than 15MBytes /s > And with APBH DMA enable in 8bit nand driver @100M EDO mode enable > Will over than 15MBytes/s too > > So , what I suggest , you should check your u-boot driver , and make sure > ADMA enable for eMMC/sd > > > > > > 2014-11-12 9:37 GMT+08:00 Robert Nelson <[email protected]>: > >> On Tue, Nov 11, 2014 at 7:09 PM, liyaoshi <[email protected]> wrote: >> > If this is for automotive ，systemd is not so useful >> > last mode is required when in automotive enviroment >> > while as I know systemd can not turnning boot sequence for different >> modules >> > If not about automotive ,skip this . >> > And，if only talking about 0.5s or 1s , its only about u-boot and kernel >> > >> > what I test in imx 6 , whats' my test result is boot from nand slc >> onfi1@80M >> > Arm@800M with 16bit DDR@400m >> > >> > unless you remove kernel printk, its possible boot less than 1s >> > >> > remove any print out from serial ,boot loader turn mmu can cache on >> > >> > load kernel with uncompressed and make small than 4M,without prink >> >> but also double check with lzo... As it can beat uncompressed... >> >> http://free-electrons.com/blog/arm-xz-kernel-decompression-benchmarks/ >> >> Regards, >> -- >> Robert Nelson >> http://www.rcn-ee.com/ >> >> -- >> For more options, visit http://beagleboard.org/discuss >> --- >> You received this message because you are subscribed to the Google Groups >> "BeagleBoard" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> For more options, visit https://groups.google.com/d/optout. >> > > -- For more options, visit http://beagleboard.org/discuss --- You received this message because you are subscribed to the Google Groups "BeagleBoard" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
