The 1st Time of DSP

DSP Hardware Frame of C6000

  The C6000 series consist of CPU, peripheral equipment(外设) and memorizer (data cache)，it’s our main equipment to finish the DSP learning, first we will touch with some assemble language.
  The load instruction is the core of assemble language, and which is related to the Digital Circuit. the instruction set differ from each in every digital equipment, while we need to get hold of a few of them.

SOP:two basic arithmetic operation

  There are two kinds of mathematics that the digital circuit is able(easy) to deal with: Multiplication and Addition, the subtract can be add a negative number, the divide can be described as multiply a count backwards. computer running is formed from limited time of addition and multiplication.
  These two operation is what the assemble language mainly deal with, they are known as SOP(sum of products), and about them, we need to know that the CPU is the operator. As a matter of fact, whether DSP or the assemble language, both have to deal with CPU directly, the assemble langue can be translated into a command to control the CPU, each line of it’s code means that the CPU will execute a once of the arithmetic operation of it’s whole bit.
  when it comes to CPU(central processing unit), some knowledge from digital circuit is required, or it will be hard for you to understand why it can be divided into some smaller function unit. In order to be corresponding to the detail described before, these small units are the addition unit(.L), multiplication(.M). but more than that, the CPU can’t run without input, so there must be a part to write some data into CPU, which is load unit(.D), and to have some senior function, we need some loops to avoid many repetitive code. So the register unit(the loop unit, .S) is essential for us.
  After learning the assemble language myself, the next step is to analyze some exist function code developed by TMS320C6000 chips. Here I will illustrate about some simple rules form the example, even it’s not the core knowledge:

  The line like MVKL is the function of the sentence, and MVK- will write 16-bits const or note into the register, while register is 32-bit, so the MVKL will write the low 16-bits then MKVH write the other high 16-bits. The order mustn’t be changed!! And the LDH is load the add and sub is addition.
  In fact, the line like .S is a category mark of the sentence, it’s correspond to former context. It’s divided these key word into a few kinds.
  The third line of these code is the object these sentence going to deal with. we can compare it to the variable in C language(not so hard ha).
  Tips: the loop is like pre-action and [B0] (true) is judgement which is composite to ![B0] (false)

CPU structure of C6000

  Actually, the key content is CPU itself, so important while so difficult.

  As we see, two universal register, each has four pathways linked to four function unit, which will disposal four instructions respectively. that’s the internal core of CPU which decided the main function, but for a whole CPU, that’s not enough. it will link with the external register to save the data, and the port to exchange information with out equipment.
  About the relation between CPU and external register, it’s based on the harvard architecture(read instruction will not interfere the execution). From the picture we can see some advantages of this CPU: it will allocating task reasonably to have a high efficiency.
  I don’t think we can explain it clearly in only one blog, so see you next chapter, but besides this, I have learned some other knowledge like: we can write some code to control the register, too. that’s so complicated, I strongly recommend that you have a probable view about it’s(CPU) digital structure.
  

see u!!