← TRIGLER Lab
Compute Platform — Concept

Raptor Core

A heterogeneous hardened-CPU + reconfigurable-fabric compute module for AI inference and robotics control — spun out of Kestrel K1's accelerator research once the fit turned out to be a different product, not a component of it.

Document RAP-CORE-001
Revision A
Issued 2026-09-18
Status Concept — pre-production
2
Cores: hardened CPU + reconfigurable fabric
160
DSP blocks, Efinix Ti60-class fabric reference
<300µW
Always-on wake tier, QuickLogic EOS S3-class
0
OS/RTOS scheduler jitter on fabric control loops
01

Origin & Scope

This project started as a side question inside Kestrel K1's accelerator-comparison research: whether an updatable soft CPU core / eFPGA could work as a gaming handheld's main processor. It doesn't — the power/clock-speed tradeoff on that fabric class is structural, not a maturity gap, and a modern gaming UI/OS needs more than a soft core can deliver at Kestrel's power budget. But the same fabric stopped looking like a workaround and started looking like the right architecture once the workload shape changed from branchy scalar code to dataflow: dense multiply-accumulate, deterministic control timing, many parallel sensor/actuator I/O lines. That's a robotics/AI compute problem, not a handheld problem, so it's its own project rather than a Kestrel component.

Scope going forward: Raptor Core is a general-purpose heterogeneous compute module — not tied to any specific robot chassis or handheld — evaluated against AI inference, motor/sensor control, and sensor-bridging workloads.
02

Why Heterogeneous, Not Homogeneous

A soft CPU core on generic LUTs and reconfigurable routing is the worst case for interpreted, branchy scalar code — it's paying fabric-level overhead for work a hardened core does natively and more efficiently. Dataflow work is the opposite case: dedicated parallel pipelines built directly in the fabric, running next to a hardened CPU that handles orchestration. Raptor Core keeps both, each doing the job it's actually good at.

Raptor Core
Orchestration (ROS nodes, planning, OS)
Runs on a hardened CPU core — Cortex-A/R-class or RISC-V hard IP
Dataflow pipelines in fabric
NN inference (DSP-block MAC/BFLOAT16), motor/sensor control loops, sensor bridging — parallel hardware, not interpreted instructions
03

Workload Fit

WorkloadWhy the fabric helps
NN inference / sensor fusion mathDense MAC is what DSP blocks are built for — e.g. Efinix Ti60-class fabric has 160 DSP blocks, each capable of INT8 quad-multiply or a fused BFLOAT16 multiply-add in hardware
Motor/sensor control loopsCycle-accurate, deterministic timing — no OS/RTOS scheduler jitter, a real capability advantage independent of power
Sensor bridgingMany simultaneous parallel I/O interfaces without contending for a shared bus/DMA like an MCU would
04

Reference Parts

Existing heterogeneous hardened-CPU + fabric parts, as a starting reference point for what a Raptor Core module could be built on or benchmarked against — not a sourcing decision yet.

PartRole
AMD Kria K26 SOMHard Cortex-A53/R5 + fabric, ROS 2-ready — the reference heterogeneous pattern
Microchip PolarFire SoCHard RISC-V + fabric, real-time control + reconfigurable I/O as the core pitch
Efinix Sapphire RISC-V SoCHard core + fabric on one die — BFLOAT16 DSP path is the point here, not soft-core clock speed
Lattice CrossLink-NX / CertusPro-NXLow-power edge AI/vision sensor bridging
QuickLogic EOS S3Cortex-M4F + small eFPGA, always-on voice/sensor wake tier — legitimate fit, not a stretch
05

Product Shape

Form
Compute module / SOM, not a finished robot — a board other robotics projects integrate, the way Kria K26 or a Jetson module is integrated rather than sold as an end product.
Target users
Hobbyist and small-team robotics builds, drone/vision platforms, always-on sensor nodes needing deterministic control timing alongside on-device inference.
Not the target
General-purpose application processing, gaming, or anything needing a modern OS/UI on the fabric side — that's the workload class Raptor Core's architecture is structurally weak at (see §02).
06

Open Questions & Next Steps

This doc is the seed of the project, not a finished spec. No system architecture, BOM, or cost reference exists yet — those are the next docs, once the questions below have a first-pass answer.

QuestionWhy it matters
Buy a reference SOM (Kria/PolarFire) vs. design a custom board?Sets the whole cost and timeline profile, same fork Kestrel K1 hit with its own silicon question
How does this compare to existing robotics compute (Jetson Orin, Raspberry Pi + HAT accelerators)?Needed before claiming the heterogeneous approach is actually better for a given workload, not just architecturally interesting
What's the first concrete workload to validate against?"AI/robotics" is too broad to design a board for — needs one real control-loop or inference target to size the fabric against
07

Recommendation

Start here
Proof-of-concept on an off-the-shelf heterogeneous dev kit (Kria K26 or PolarFire SoC dev board) against one real workload — don't design custom silicon or a custom board before that validation exists.
Pattern to follow
Same shape as Kestrel K1's SteamOS POC: validate the architecture on existing hardware before committing to bespoke silicon or tooling.
Raptor Core — Concept Spec Rev A Figures are public datasheet specs and engineering estimates, not a sourced teardown