Skip to main content
Skip to main content

DMA Controller with AHB IIP

DMA Controller with AHB IIP

Overview

COMPETITIVE ADVANTAGE

The SivaKali Tech DMA Controller with AHB IIP is a highly reliable, silicon-proven IP core designed for high-performance system integration. Engineered to meet strict industry compliance standards, this core offers exceptional flexibility and ease of integration into both ASIC and FPGA designs. Forming the high-speed communication backbone of complex System-on-Chips. Validated on leading FPGA platforms and foundry nodes, it provides a low-risk, time-to-market advantage for developers.

COMPETITIVE ADVANTAGE

Deadlock Free: Robust routing logic prevents system hang-ups under heavy load conditions.

Low Latency Bridging: Efficient clock domain crossing and protocol conversion with minimal cycle overhead.

High Frequency: Pipelined architecture designed to close timing at high clock frequencies in modern nodes.

Scalable: Easily configurable for simple bus fabrics or complex, multi-layer network-on-chip (NoC) implementations.

FEATURES
  • Supports 1-16 channel DMA Transmit and DMA Receive Engine
  • Compliant with ARM AMBA 2 AHB Specification
  • Optional support for AMBA 3 AHB-Lite and AMBA 5 AHB Specification
  • Supports access for Ring and Chained Descriptor Structures
  • Configurable Transmit and Receive Engine based on Host Memory Data Width
  • Configurable support by DMA Transmit and Receive Engine for both of the endianness of the host memory (Little / Big Endian)
  • Supports configurable DMA Transmit and DMA Receive FIFO based on Host Memory Data width
  • Supports hardware DMA Control registers that can be written and read by CPU
  • Round Robin algorithm for arbitration between DMA Transmit and Receive Engine to access SoC Master Bus
  • SoC Master bus can be AXI/AHB/APB/OCP/Tilelink/Wishbone
  • Supports AHB Slave bus
  • Uses SoC Slave Interface to get Receive and Transmit descriptors and transfer the data to/from the system memory from/to FIFO inside the DMA controller
  • User logic to map data fetched from Host to IP core or from IP core to host
  • Supports following DMA transfers
    • Memory to Memory
    • Memory to Peripheral
    • Peripheral to Memory
    • Peripheral to Peripheral
  • Supports Sideband DMA request and Grant based triggering of transfers as on option for peripherals
  • Supports Scatter Gathers list
  • Supports 8/16/32/64/128/256 bit wide data transfers
  • Supports QoS per channel if SoC master interface supports Qos.
  • Supports programmable burst capability per SoC master
  • Supports both single data and burst data transfers, with burst size based on the burst length field in the DMA control registers
  • DMA supports full duplex operation, processing read and write transfers at the same time
  • Supports Link list based processor for autonomous operation
  • Interface widths can be controller for each SoC master interface
  • Supports different algorithms for selecting which channel to process and how long to process.
  • Generate full 32-bit addresses on the SoC master interface
  • Supports per Channel Interrupt output
  • Supports up to 64 MB transfer per Buffer Descriptor (BD)
  • Interrupts CPU on completion of a DMA transfer or an error
  • Fully synthesizable
  • Static synchronous design
FUNCTIONAL DESCRIPTION

CORE: Core module interconnects all the sub-modules in the DMA controller with AHB IIP. Ports of core module are the top level ports for the DMA controller with AHB IIP.

AHB SLAVE: AHB Slave is used to respond to a read or write operation within a given address-space range. It accepts address and control from the master and returns data/response.

AHB MASTER: AHB Master is use to initiate read and write operations by providing an address and control information.

DMA Arbiter: The arbiter looks for any write/read request and grants the bus to DMA Transmit or DMA Receive engine.

DMA FSM: This module implements DMA source state machine. It controls the fsms such as own bit check read local data, write data fetch

Router: This module implements DMA hardware router at top level.

Peripheral: It represents the destination devices that generate or consumes data from DMA transfers.

CSR: CSR module has all the registers. The contents of the registers are decoded and assigned to its respective output ports based on its functionality.

ASIC AND FPGA IMPLEMENTATION
ASIC TechnologyLogic ResourcesClock Frequency
TMSC 28nm45k100MHz

FPGA Device and FamilyLogic ResourcesClock Frequency
AMD-xcvu9p-flga2104-2L-e7500 LUT's100MHz

LICENSING OPTIONS
  • Single Site license for regional development teams.
  • Multi-Site license for global corporate deployments.
  • Single Design license for specific project cost-efficiency.
  • Unlimited Design license for high-volume product roadmaps.
DELIVERABLES
  • Complete Verilog/VHDL/SystemC Source Code.
  • UVM-compliant verification environment with a comprehensive test suite.
  • Production-ready synthesis, Lint, and CDC scripts.
  • IP-XACT RDL generated address maps.
  • Standard-compliant firmware and Linux/C driver packages.
  • Detailed documentation: User Guides, Release Notes, and ISO 26262 Safety Manual (SAM)/FMEDA.