AI Accelerator Design · All levels

Weight-Stationary Dataflow

Systolic Arrays: In weight-stationary scheduling, each PE holds a weight value (or small filter fragment) for multiple cycles while activations stream through the mesh. Keeping weights local reduces expensive weight fetch traffic and is especially attractive for convolutional and batched inference workloads where filters are reused heavily across many input windows. The tradeoff is that activation movement and partial-sum routing can become dominant bandwidth or latency bottlenecks if tiling is not aligned to on-chip buffer capacity. Effective designs co-optimize filter blocking, activation prefetch, and reload cadence so stationary weights remain fully productive instead of idling between tiles.

What this topic teaches

Weight-Stationary Dataflow converts accelerator architecture concepts into release-ready engineering decisions. In weight-stationary scheduling, each PE holds a weight value (or small filter fragment) for multiple cycles while activations stream through the mesh. Keeping weights local reduces expensive weight fetch traffic and is especially attractive for convolutional and batched inference workloads where filters are reused heavily across many input windows. The tradeoff is that activation movement and partial-sum routing can become dominant bandwidth or latency bottlenecks if tiling is not aligned to on-chip buffer capacity. Effective designs co-optimize filter blocking, activation prefetch, and reload cadence so stationary weights remain fully productive instead of idling between tiles.

Senior-engineer framing question

When Weight reuse factor and off-array weight bandwidth per tera-operations for representative inference layers. regresses, can you isolate the first failing execution mechanism, collect decisive evidence, assign owners, and close with rollback-safe validation?

diagram

ACCELERATOR EXECUTION FLOW - Weight-Stationary Dataflow

request ingress and model metadata
      |
      v
graph lowering and kernel selection
      |
      v
tile/dataflow scheduling and memory placement
      |
      v
tensor execution + synchronization barriers
      |
      v
result assembly + quality/SLA validation
      |
      v
release decision and rollback guardrails

Evidence to collect

Primary metric: Weight reuse factor and off-array weight bandwidth per tera-operations for representative inference layers..
Primary artifact: Dataflow mapping report showing weight residency, activation traffic, and buffer pressure by layer type..
Owners to include: accelerator architect, compiler mapping owner, SRAM subsystem owner, runtime scheduler owner, model performance owner.
One reproducible failing workload and one stable comparator run.
One fixed-metadata run with compiler/runtime/hardware tags locked.

Bandwidth lens

diagram

BANDWIDTH LENS - Weight-Stationary Dataflow

working-set pressure
  ^
  |                saturation zone
  |          ----------------------------
  |      o   unstable tail latency
  |   o      tuning candidate
  | o        baseline behavior
  +-------------------------------------> optimization iteration

Primary metric tracked:
Weight reuse factor and off-array weight bandwidth per tera-operations for representative inference layers.

Ownership layers

diagram

OWNERSHIP LAYERS - Weight-Stationary Dataflow

+----------------------+--------------------------------+--------------------------------+
| Team                 | Primary responsibility         | Closure artifact               |
+----------------------+--------------------------------+--------------------------------+
| accelerator architect | mechanism and architecture intent| design rationale + tradeoffs   |
| compiler mapping owner | mapping, runtime, and execution   | profile traces + bottleneck map|
| SRAM subsystem owner | correctness, risk, and signoff    | test report + closure memo     |
+----------------------+--------------------------------+--------------------------------+

Key takeaways

Start with mechanism classification before changing tuning knobs.
Use one proving artifact for each major claim in review discussions.
Close with explicit owners, validation matrix, and rollback criteria.

Common pitfalls

Optimizing only peak throughput while p99 latency or quality regresses.
Mixing evidence captured from mismatched runtime or thermal conditions.
Declaring closure without production-like replay and guardrail checks.

AI accelerator deep dive

Systolic efficiency is governed by feed quality, tile fit, and bubble control.

Concept diagram

diagram

SYSTOLIC WAVEFLOW

operand stream -> wavefront launch -> PE mesh compute -> reduction/writeback
                         ^ bubbles and feed stalls reduce realized throughput

Metric graph

diagram

UTILIZATION LOSSES

tile mismatch      ███████
feed stalls        █████████
sync bubbles       █████

Metrics and artifacts to collect

mesh occupancy timeline
fill-drain overhead
tile mismatch histogram
DRAM stall attribution

Mini case study

Increasing mesh size did not help until scheduling and tile alignment removed persistent wavefront bubbles.

Debug branches

Measure bubble source first
Classify compute vs memory starvation
Tune tile policy before frequency changes

Senior review question

Ask: which first-principles bottleneck class explains the symptom, and what artifact proves it reproducibly?

Key takeaways

Tie every accelerator claim to a reproducible workload slice and one primary metric trend.
Prefer bounded fixes with clear owner and rollback boundary over broad tuning bundles.

Common pitfalls

Optimizing synthetic kernels without production-shape validation.
Reading average latency while ignoring p95 and p99 behavior.
Declaring sparse or precision wins without fallback and quality evidence.

Practice this lesson