SPRUJ28E November 2021 – September 2024 AM68 , AM68A , TDA4AL-Q1 , TDA4VE-Q1 , TDA4VL-Q1
The LDC node scheduler comprises 0 consumer sockets and 7 producer sockets. On consumer side, the LDC uses its own read DMA engine to fetch input data directly into internal buffer of LDC. On producer socket, it can be connected to 5 different consumers.
The producer sockets are mapped to the following output buffers:
Pipeline #n (n = 0, …, 6 configurable) is mapped to this scheduler. When a scheduler is enabled, the LDC read DMA engine loads required input data into the LDC memory. Prior to starting a LDC thread, the output buffer availability is checked to write output data. These producer sockets can optionally be disabled depending on the usecase.
To support on-the-fly MSC and NF operation after LDC, a pattern adapter is used to convert blocks into lines. Although produced data count is only two, the produced data count is still included, along with pattern adapter pair for unique end consumers.
On-the-fly support for LDC -> MSC / LDC -> NF connection can be achived via the SL2 memory (if space is available) or via the SoC level MSMC SRAM. Both flows are supported using the HTS scheduler.