Forum Discussion

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

How to find out a Tco from TimeQuest

Hi all!

I'm using Quartus TQ and I have a question about the best way to measure a Tco for a FPGA pin to pin. My task is specifically for the case in that example - https://www.altera.com/content/dam/altera-www/global/en_us/pdfs/literature/wp/wp_stxtco.pdf. In Figure 2. Sample tCO Design we see 3 FFs, so commands like report_path and report_timing as I figured out can't give result immediately - I need to iteratively use them for FF-to-FF path for example. Is it another option, for example to view the summary of paths from input to output pin using one command?

Thanks in advance.

14 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor

    --- Quote Start ---

    Hi kaz!

    Yes, exactly, as about tCO parameter it is what you wrote. I was interested if there is such parameter that sum tCO and integrated clock cycles on other register (a path from input pin to output pin with intermediate registers inside the path).

    --- Quote End ---

    data transition at first reg is sampled into reg2 then reg3 ...etc. This is latency issue and you the designer will know that. The tool does not change latency as it is vital for functionality.

    Any routing/logic delays between registers is absorbed into the sampling periods.(normally 1 period unless multicycle > 1)

    so what you want is [pin-to reg1 delay + number of regs in chain x clock period + last reg to pin delay]. =~ number of reg x clock period
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor

    --- Quote Start ---

    data transition at first reg is sampled into reg2 then reg3 ...etc. This is latency issue and you the designer will know that. The tool does not change latency as it is vital for functionality.

    Any routing/logic delays between registers is absorbed into the sampling periods.(normally 1 period unless multicycle > 1)

    so what you want is [pin-to reg1 delay + number of regs in chain x clock period + last reg to pin delay]. =~ number of reg x clock period

    --- Quote End ---

    Yes, kaz, I absolutely agree with you! I just thought that there is a command to find out that value automatically without manual calculation.

    Thank you very much!
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor

    That would be useful indeed for delay measurements. I used signaltap at some point just to measure delays.

    For mobile Tx/Rx systems all delay figures are important and have to be worked out by various engineers part of which is FPGA.
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor

    --- Quote Start ---

    That would be useful indeed for delay measurements. I used signaltap at some point just to measure delays.

    For mobile Tx/Rx systems all delay figures are important and have to be worked out by various engineers part of which is FPGA.

    --- Quote End ---

    Yes, kaz! So I think it would be great if such a total delay command would be inserted in TQ.