New PAX_LDPC v1.0 released — FPGA-verified QC-LDPC for TLC/QLC SSD Learn more →
PaxQubit
← IP Core Catalog
New RTL-complete

Signal Processing

ADSP-2181 Compatible DSP Core IP

A clean-room RTL DSP core that executes genuine ADSP-21xx object code — a drop-in path for systems built around an obsolete part.

69.91 MHz
Fmax on Arria II GX — about 2x the original part
3
Independent simulators, matching cycle counts
24-bit
Genuine ADSP-21xx instruction encoding

Overview

Legacy DSPs go out of production long before the systems built around them retire. This core solves that: it implements the genuine 24-bit ADSP-21xx instruction encoding, so binaries produced by the original toolchain run unmodified — no source port, no re-qualification of the algorithm.

Cycle accuracy was treated as a hard requirement, not a goal: the arithmetic units, status flags, sequencer and address generators reproduce the original timing so that sample-rate-critical audio and control loops behave identically. Interrupt entry masking, secondary register banks, circular buffers and the 40-bit MAC accumulator are all modelled.

Verification is layered and repeatable: a regression suite runs identically on three independent simulators — down to matching cycle counts — and synthesis is clean, with no combinational loops and no optimized-away registers. The assembler used for the test suite was written from the published instruction manual alone.

Parametric specification

Role Drop-in replacement core for obsolete ADSP-21xx family parts
Instruction set ADSP-21xx, full 24-bit encoding per the published manual
Cycle accuracy Hard requirement — timing matches the original device
Datapath 16-bit ALU / shifter, 40-bit MAC accumulator, primary + secondary register banks
Address generators Dual DAG with modulo (circular) and bit-reverse addressing
Peripherals SPORT, timer, autobuffer, BDMA, IDMA, flag I/O, interrupt controller
Verification Regression suite passing on three simulators with identical cycle counts
Synthesis 0 errors, no combinational loops, no optimized-away registers; Fmax 69.91 MHz (Arria II GX)
Toolchain In-house assembler written from the published instruction manual
Implementation Pure Verilog RTL, vendor-primitive-free
Status RTL complete and verified; validated against real legacy application code

Performance

69.91 MHz
Fmax on Arria II GX — about 2x the original part
3
Independent simulators, matching cycle counts
24-bit
Genuine ADSP-21xx instruction encoding

Deliverables

  • Synthesizable Verilog RTL — core, peripherals and top-level wrapper
  • Instruction-encoding specification as the single source of truth
  • Self-checking regression testbench and simulator flow
  • Assembler and toolchain notes; integration guide (on request)

Supported Devices

Intel/Altera Arria II GX (reference synthesis)ASIC — process-portable pure RTL

Targets / verified platforms. ASIC process numbers are characterized at integration.

Because the core executes the original object code, an existing application can be migrated without touching its source — the usual risk and cost of a legacy DSP replacement (re-porting, re-tuning, re-qualifying) largely disappears.

Synthesis figures are from the reference Arria II GX build; results vary with target device and constraints. Detailed specifications, register maps and test suites are available on request.

Documents

Full datasheet, integration guide and evaluation files are available on request. Request access below — we reply within one business day.

Request Datasheet

Applications

Life-extension of fielded systems whose DSP has gone obsolete
Audio effect and signal processing engines
Embedded control loops requiring deterministic timing
SoC integration — replacing an external DSP with an on-chip core

Request Quote

Request a Quote or Evaluation

Tell us about your project. Fields marked * are required. We respond within one business day.

We never share your details. Used only to respond to your request.

If the form does not work, email yimidi@daum.net directly.