This website uses cookies. By using this site, you consent to the use of cookies. For more information, please take a look at our Privacy Policy.
Home > FPGA Technical Tutorials > FPGA-Based Prototyping Methodology > PROTOTYPING + VERIFICATION = THE BEST OF BOTH WORLDS > Comparing verification interface technologies

TABLE OF CONTENTS

Xilinx FPGA FPGA Forum

Comparing verification interface technologies

FONT SIZE : AAA

In order to compare different verification interfaces, Synopsys has created a  demonstration benchmark based on an image processing sub-system. The  demonstration system uses a JPEG algorithm to compress an image, which is stored  on a host computer. We first simulated the RTL in VCS running on the host  computer before moving the DUT from simulation to a synthesized target running  on an FPGA prototype, which enabled us to compare co-simulation with HDL  Bridge, SCE-MI and UMRBus. 

The comparisons and performance data are summarized in Table 34 later but in the  meantime, Table 32 summarizes the comparison between the four different  modeling techniques.

Summary of performance and benefits of four interface implementations.png

For this particular task, we can see that UMRBus is the fastest implementation.  SCE-MI, implemented on top of UMRBus is second fastest thanks to its use of  transaction-based communication. Co-simulation using cycle-based interfaces for  this task is about 300x slower than SCE-MI, and RTL simulation is over 4000x  slower than SCE-MI. While this data helps us to compare the different approaches,  it is not possible to generalize. Performance depends on many different parameters,  for example, how much computation is happening on hardware and how much is in  the simulator, also, what is the amount of data that is being exchanged between  different sides of the transactors, and what transport mechanism is used? Of course  we will also get different results for different kinds of design. 

The UMRBus channel is for the sole use of the co-simulation and SCE-MI  implementations. Using an alternative bus standard, such as PCIe, may deliver  worse performance as the channel is not a dedicated interface and it may have to  handle other traffic as well as the co-simulation data. 

As well as considering performance, it is worth comparing the debug capabilities of  each approach. UMRBus enables users to interact directly with the design from the  software world. Intensive debug activity in HDL Bridge or SCE-MI may reduce  overall performance. The number of signals that we can capture in HDL Bridge may  be limited.

It is also possible to perform debug by physically connecting to the FPGA using a  logic analyzer and physical probes. This allows probing of internal FPGA signals,  and it is useful for tracing free-running designs. There are a couple of potential  drawbacks with this approach to debugging – there may be a delay on debug traces,  and probe signal names may not exactly match the original RTL names.



  • XC2C64A-5QFG48C

    Manufacturer:Xilinx

  • CPLD CoolRunner -II Family 1.5K Gates 64 Macro Cells 263MHz 0.18um Technology 1.8V 48-Pin QFN EP
  • Product Categories: CPLDs

    Lifecycle:Active Active

    RoHS:

  • XC2C64A-7CP56C

    Manufacturer:Xilinx

  • CPLD CoolRunner -II Family 1.5K Gates 64 Macro Cells 159MHz 0.18um Technology 1.8V 56-Pin CSBGA
  • Product Categories: CPLDs

    Lifecycle:Active Active

    RoHS: No RoHS

  • XC2C64A-7F33C

    Manufacturer:Xilinx

  • Xilinx BGA
  • Product Categories:

    Lifecycle:Any -

    RoHS: -

  • XC5VFX100T-1FFG1738C

    Manufacturer:Xilinx

  • FPGA Virtex-5 FXT Family 65nm Technology 1V 1738-Pin FCBGA
  • Product Categories: FPGAs (Field Programmable Gate Array)

    Lifecycle:Active Active

    RoHS:

  • XC5VFX100T-2FF1738C

    Manufacturer:Xilinx

  • FPGA Virtex-5 FXT Family 65nm Technology 1V 1738-Pin FCBGA
  • Product Categories: Connecteurs

    Lifecycle:Active Active

    RoHS: No RoHS

Need Help?

Support

If you have any questions about the product and related issues, Please contact us.