DSP default cache-sizes not optimal?

While debugging some DSP code yesterday I came a cross a performance oddity. Adding more code lowered the performance of an unrelated function.

By itself this is not *that* odd. It happens if the size of your code is larger than your first level code-cache and different functions start to kick each other out of the cache. However, in my little toy program this was unlikely. I had only around 20kb of code and the code-cache is 32kb in size.

Better safe than sorry I thought and took a look how the caches are configured. Big and pleasant surprise: Two of them are running at half the maximum size for no good reason:

In my case after DSP-boot I got:

Level 1 Data-Cache 32k
Level 1 Code-Cache 16k
Level 2 Cache      32k

However, the maximum possible cache sizes for the BeagleBoard are

Level 1 Data-Cache 32k (no change)
Level 1 Code-Cache 32k (16kb larger)
Level 2 Cache      64k (32kb larger)

So 48kb of valuable cache has been left unused. Changing the cache sizes is easy:

  #include < bcache.h >

  // and somewhere at the start of main()
  BCACHE_Size size;
  size.l1dsize = BCACHE_L1_32K;
  size.l1psize = BCACHE_L1_32K;
  size.l2size  = BCACHE_L2_64K;
  BCACHE_setSize (&size);

That still leaves you the 48kb of L1DSRAM for single cycle access and 32kb of L2RAM to talk with the video accelerators. Oh – and it gave a noticeable performance boost.

Btw- it’s very possible that this only applies to the DspLink configuration that I am using.


It turned out that the reason for the smaller cache-sizes is the default DspLink configuration. You can override this if you add the following lines to your projects TCF-file. Just put them somewhere between utils.importFile(“dsplink-omap3530-base.tci”); and prog.gen():

prog.module("GBL").C64PLUSL2CFG  = "64k";
prog.module("GBL").C64PLUSL1DCFG = "32k";
prog.module("GBL").C64PLUSL1PCFG = "32k";

var IRAM = prog.module("MEM").instance("IRAM");
IRAM.len = 32768;

This will configure the OMAP3530 DSP with:

L2-Cache:     64kb
L1 Data-Cache 32kb
L1 Code-Cache 32kb
L1SDRAM       48kb
IRAM (L2 Ram) 32kb
This entry was posted in Beagleboard, DSP, OMAP3530. Bookmark the permalink.

7 Responses to DSP default cache-sizes not optimal?

  1. Guillaume says:


    I am in the same situation. My .map gives this:
    name origin length used unused attr fill
    ———————- ——– ——— ——– ——– —- ——–
    CACHE_L2 10808000 00008000 00000000 00008000 RWIX
    CACHE_L1P 10e04000 00004000 00000000 00004000 RWIX
    L1DSRAM 10f04000 00004000 00000382 00003c7e RWIX
    CACHE_L1D 10f10000 00008000 00000000 00008000 RWIX

    But L1DSRAM is only 16k…


  2. Nils says:


    It seems like the .tcf-file is simply wrong. The L1DSRAM is 48kb, Take a look at the OMAP3530 Technical Reference Manual chapter 2.4.4 (L3 Interconnect View of the IVA2.2 Subsystem Memory Space). L1SDRAM goes from 0x10f04000 to 0x10f0ffff, directly followed by the 32kb of L1D-cache.


  3. Guillaume says:

    I’m trying to change the dsplink-omap3530-base.tci file, but the I always have compilation failure… (like MEM segment IRAM: overlaps with another segment or cache configuration).
    I haven’t figured out which other definition conflicts with this..


  4. Nils says:

    Hi Guillaume,

    How about this? http://torus.untergrund.net/code/dsplink-omap3530-base.tci

    The tconf works, but I haven’t tested if the dsp-binaries compiled with this script work. The map-file looks good though.


  5. Nils says:

    Just verified that it works.. Not only in theory but in practice. Yay!

  6. Joseph says:


    I am trying to implement an image processing algorithm on beagleboard. I am planning to use DSP in the beagleboard to do the job. Since I am a beginner with the beagleboard I have no idea how to write code for dsp and run it in the beaglebord. Could you please tell me the steps that I should follow to get this done?
    I have a beagleboard C4 running embinux android eclair on it.

  7. Abhijeet says:


    Can you please let me know what all modifications can one make to the cache inside the BeagleBoard. I am interested in checking for cache optimization specifically for Set Top Boxes but do not have enough information available to go ahead. I already have a beagleboard.

    Any inputs or pointers will be welcome.


Leave a Reply

Your email address will not be published. Required fields are marked *