• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 98 subscribers
  • Views 256 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Nand Timings on the OMAP3530

scettindro6251
Prodigy 180 points
Other Parts Discussed in Thread: OMAP3530

Hello,

We Have a board with Omap3530, with a MT29F4G08AACWC nand Flash. Wait0 is pullup with 1k.

We are measuring Read  Trhoughput using :  "nanddump /dev/mtd4/ -f /dev/null/ -l 261619712".

My filesystem is Jffs2. We are using the last kernel 2.6.32 and the Uboot 2009.11.

The Throughput we measure is 3,4 MByte/s.

Looking at the http://wiki.davincidsp.com/index.php/AM35x-OMAP35x-PSP_03.00.00.04_Release_Notes

so the PSP 03.00.00.04 we found that similar measurements were done  with the same nand and filesystem, and are about 9Mbyte/s

We also took a look to the source code of the PSP 03.00.00.04 and found no differences with our code.

I have some questions:

1) Is possible to know how this throuput test was done? Maybe  we are not using the same method.

2) Is it possible to have a register dump for the GPMC register (for the CS0) from your system? The difference may be there.

 

Thanks

 

Sandro Mascetti

  • 0
    TI__Mastermind 41665 points

    scettindro said:
    My filesystem is Jffs2. We are using the last kernel 2.6.32 and the Uboot 2009.11.

    Are you using the same kernel that is included with the PSP 3.00.00.04 or did you get your kernel elsewhere? There may be dependencies in sources outside of the sources you were looking at that could impact performance, TI can only guarantee the performance levels of the PSP drivers if you are using the PSP kernel.

    Unfortunately I am not sure exactly how the testing was done for the NAND driver, since I see no special mention of any particular commands like nanddump, I would have assumed they were simply timing a copy in or out of the NAND filesystem but I could not say for sure. To get a more solid answer I am moving this thread to the embedded software > Linux forum, where the PSP team can take a look at it.

  • 0 in reply to Bernie Thompson TI
    Prodigy 180 points

    Hello

    We are using the official beagle board 2.6.29.omap1.r37 with nand driver omap2.c updated.

    Anyway we will test your official kernel to see if there are any changes.

     

    sandro