In this article
The development of logic that processes discrete digital values in a custom chip. It includes defining the architecture, writing RTL, verifying behavior and implementing the design in a target technology.
Turning behavior into hardware
A digital ASIC implements operations with logic gates, registers and memories chosen for a particular application. The starting point is a description of what the product must do: which data arrives, how quickly it must be processed, which interfaces connect it, and what happens when inputs are missing or invalid. A useful specification separates functional behavior from implementation preferences. A requirement to deliver a result within a latency limit is testable; a request for a very fast chip is not. This distinction gives architecture and verification teams a shared basis for decisions.
Architecture before RTL
Engineers divide the work among processing elements, control logic, memories and interconnect. They consider concurrency, pipeline depth, buffering and backpressure before writing large amounts of RTL. An accelerator that performs arithmetic quickly can still stall if its memory system cannot supply operands. A controller with a small gate count can still miss a real-time deadline if interrupts and transactions interact badly. Early models should therefore exercise complete workloads and stressful operating cases. These models are planning tools rather than promises of final silicon performance.
The implementation boundary
Synthesizable RTL is translated into cells from a target library, then implemented physically. Decisions about resets, clock domains, memories and power states affect both stages. A design that simulates correctly can be difficult to place or route, or fail timing under its actual constraints. Physical feedback should reach architecture owners while changes remain practical. The handover includes the RTL revision, constraints, library assumptions, verification results and documented exceptions. A directory containing source files without this context is not a complete implementation package.
A practical design review
Consider a packet-processing block with an input interface, lookup memory and output queue. Review what happens when the output stops accepting data, the lookup produces an error, or a reset arrives midway through a packet. Check whether counters overflow and whether software can recover from a fault. These questions often reveal more useful requirements than a single throughput target. Record the expected behavior and add corresponding tests. A review becomes valuable when its findings change the specification, the implementation or the evidence used to release the design.
Questions engineers ask
Is digital ASIC design the same as writing Verilog?
No. RTL coding is one part of a broader process that includes architecture, verification, physical implementation and production preparation. A correct source description still needs constraints, testability and implementation evidence.
Can a digital ASIC change after fabrication?
Its manufactured logic is fixed, but programmable registers, embedded software and configurable datapaths can provide planned flexibility. Those mechanisms must be designed and verified before fabrication; they do not make arbitrary hardware changes possible afterward.
Requirements that can be checked
Turn each important behavior into a requirement with an observable result. Describe valid inputs, invalid inputs, ordering rules, throughput expectations and the response to interrupted transactions. Identify which clocks and resets are involved. A good review asks whether another engineer could write an independent test from the description. Where the answer is unclear, improve the requirement before adding more implementation detail. Keep assumptions about the surrounding system visible, including how long an interface can stall and whether input data can arrive while a block is unavailable. These details determine whether the implementation and its test environment are solving the same problem.
Verification across normal and unusual operation
Check ordinary workloads alongside stressful combinations of events. Reset during activity, a full queue, simultaneous requests and an error response can expose behavior that an ideal sequence misses. Use appropriate simulation, assertions or formal methods for the properties being examined. Preserve reproducible inputs and expected results. If a check is excluded, record why and who reviewed the decision. Coverage helps organize evidence, but its interpretation depends on what the tests actually check. A regression that executes many cycles without observing meaningful outcomes can provide less confidence than a smaller set of well-targeted checks tied to requirements.
Implementation handover
The source design needs constraints, technology assumptions and integration context before a receiving team can use it reliably. Include the intended operating modes, clock relationships, IP versions and any known limitations. Check that source files, models and reports describe the same configuration. Define how changes are communicated while implementation is in progress. When physical results challenge an assumption, route the finding back to the responsible architecture or logic owner. A controlled handover keeps the project from optimizing an obsolete version or interpreting an undocumented exception as an intentional requirement.
Evidence for the next decision
Separate predicted results from measured behavior and state the conditions behind both. An early synthesis estimate can support architecture exploration, but it is not equivalent to a final implemented result. A prototype can support software integration without proving final ASIC timing. Before a release, identify the evidence required for the particular stage and list unresolved issues with owners. Preserve the approved revision and enough context to reproduce the relevant checks. This makes later debugging, engineering changes and product maintenance more efficient than relying on a single successful run whose inputs are no longer clear.
A practical cross-discipline review
Before committing the next stage, bring the design, verification and product-engineering owners together around the same revision. Confirm the assumptions that pass between them: operating modes, interfaces, test access and the evidence expected at release. List unresolved issues with a responsible owner and a concrete way to close them. An open question is easier to manage when its consequence is understood than when it is hidden inside a general progress estimate. Preserve the decision record so a later change can be evaluated against the original reasoning. This review is useful even for a narrow specialist task, because the surrounding product still determines whether the delivered result is usable.
Include the receiving team's acceptance question in the review: what must be demonstrated before this result can be used safely in the next engineering stage? Agree how that demonstration will be recorded and how a subsequent change triggers a focused reassessment of the affected evidence.
Explore the related Avnet offering or discuss your project requirements.
