Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
14 years ago

Arccos calculation

Hello everyone,

I am using de0nano and need to perform arccos calculation.

I am aware that altera have floating point megafunction for arctan and sin/cos, but it does not have arcsin or arccos mega function.

What option do I have?

My data is two complement 12 bit width which means it has 4096 possibilities. If I have to use look up table(LUT) method, then I have to program it with 4096 cases.(which is way too much if I have to do it manually using LUT)

Thank you anyone for your help.

Gunardi :)

5 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    14 years ago

    If you're using HDL - you can create a LUT using the ROM template from the coding guidelines and you're allowed to use a function to initialise the ram. In VHDL you can use the arcsin and arccos functions from math_real package to generate the LUT.

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    14 years ago

    and - you could tie it up with the new VHDL floating point package to convert the real types to std_logic_vectors.

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    14 years ago

    @kaz: Thank you for your reply. It is just cosine value not Re/im input.

    @Tricky: Thank you for your reply.

    "If you're using HDL - you can create a LUT using the ROM template from the coding guidelines and you're allowed to use a function to initialise the ram."

    What is ROM template or ram? How to create it? Is it possible to use verilog not vhdl?

    Please be more specific, because I am a beginner in FPGA:)
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    14 years ago

    Given cosine input range from 0 ~ 4095 representing 1 ~ -1

    then use one of following options:

    1) use full LUT, simple, needs 4096 * bitwidth as you know. Use cosine input as direct address.

    2) use half LUT plus logic. since upper half is anti-symmetric with lower half. e.g.

    if address > 2047

    address = 4095 - address

    data = pi - data

    end

    but pi need be scaled correctly.

    3)use smaller LUT e.g. 256 locations then add logic to interpolate linearly values in between. practical and good approximation and intuitive.

    4) explore possibilty of using cordic approach