External Memory Interface Handbook Volume 3: Reference Material: For UniPHY-based Device Families

ID 683841
Date 3/06/2023
Public
Document Table of Contents

9.1.3. Traffic Generator and BIST Engine

The traffic generator and built-in self test (BIST) engine for Avalon-MM memory interfaces generates Avalon-MM traffic on an Avalon-MM master interface. The traffic generator creates read and write traffic, stores the expected read responses internally, and compares the expected responses to the read responses as they arrive. If all reads report their expected response, the pass signal is asserted; however, if any read responds with unexpected data a fail signal occurs.

Each operation generated by the traffic generator is a single write or block of writes followed by a single read or block of reads to the same addresses, which allows the driver to precisely determine the data that should be expected when the read data is returned by the memory interface. The traffic generator comprises a traffic generation block, the Avalon-MM interface and a read comparison block. The traffic generation block generates addresses and write data, which are then sent out over the Avalon-MM interface. The read comparison block compares the read data received from the Avalon-MM interface to the write data from the traffic generator. If at any time the data received is not the expected data, the read comparison block records the failure, finishes reading all the data, and then signals that there is a failure and the traffic generator enters a fail state. If all patterns have been generated and compared successfully, the traffic generator enters a pass state.

Figure 70. Example Driver Operations


Within the traffic generator, there are the following main states:

  • Generation of individual read and writes
  • Generation of block read and writes
  • The pass state
  • The fail state

Within each of the generation states there are the following substates:

  • Sequential address generation
  • Random address generation
  • Mixed sequential and random address generation

For each of the states and substates, the order and number of operations generated for each substate is parameterizable—you can decide how many of each address pattern to generate, or can disable certain patterns entirely if you want. The sequential and random interleave substate takes in additions to the number of operations to generate. An additional parameter specifies the ratio of sequential to random addresses to generate randomly.