Key Takeaways
- QuickLogic’s upgraded eFPGA Hard IP for GlobalFoundries 12‑LP (GF 12LP) adds a dedicated DSP MAC engine with SIMD and cascade support.
- The fabric now scales past 250 k LUTs, enabling far larger on‑chip programmable blocks.
- An optional Configuration‑Bit Health Monitor provides real‑time SEU detection and correction for mission‑critical reliability.
- Full RTL‑to‑bitstream flow is supported through QuickLogic’s Aurora and Aurora PRO toolchains.
Overview of the Expanded eFPGA Hard IP
QuickLogic has leveraged its silicon‑proven eFPGA architecture to deliver a next‑generation Hard IP block that can be embedded directly into a System‑on‑Chip (SoC) fabricated on GlobalFoundries’ 12‑level‑planar (12‑LP) process. The new IP targets compute‑intensive workloads—such as phased‑array radar, electronic warfare, 5G/6G wireless infrastructure, satellite communications, and high‑resolution imaging—while also meeting the stringent reliability demands of aerospace and defense systems.
New DSP Multiply‑and‑Accumulate (MAC) Unit
- Architecture: Single‑instruction‑multiple‑data (SIMD) MAC with a pre‑adder stage and cascade routing.
- Performance: Up to 2 × 10⁹ MAC operations per second at 1 GHz core frequency (typical for GF 12LP).
- Application Fit: Enables real‑time beam‑forming, FIR filtering, and matrix‑vector multiplication with lower latency than external DSP blocks.
Scalability Beyond 250 k LUTs
- Maximum LUT Count: > 250 000 lookup tables (LUTs) per eFPGA instance, a 40 % increase over the previous 180 k‑LUT limit.
- Block RAM: Up to 1 MB of embedded BlockRAM can be provisioned alongside the LUT fabric.
- Design Flexibility: Users can mix‑and‑match LUT, BRAM, and DSP resources to tailor the fabric to specific algorithmic footprints.
Optional Configuration‑Bit Health Monitor
- Function: Continuously scrubs configuration memory for Single‑Event Upsets (SEUs) and automatically rewrites corrupted bits while the user design remains active.
- Detection Latency: < 5 µs from upset to correction, ensuring near‑real‑time integrity.
- Reliability Impact: Improves Mean Time Between Failures (MTBF) by an estimated 10× for radiation‑prone environments.
Development Ecosystem
QuickLogic supplies end‑to‑end tooling:
| Toolchain | Primary Capability | Supported Flow |
|---|---|---|
| Aurora | RTL synthesis, placement, routing | Standard eFPGA designs |
| Aurora PRO | Advanced timing closure, power analysis, multi‑domain verification | Complex, high‑performance designs |
| Bitstream Generator | Automated bitstream creation with health‑monitor integration | One‑click deployment |
The toolset produces a complete RTL‑to‑bitstream path, allowing designers to iterate rapidly and embed updates post‑fabrication—critical for long‑life products.
Feature Comparison: Legacy vs. Expanded eFPGA (GF 12LP)
| Feature | Legacy QuickLogic eFPGA (pre‑GF 12LP) | Expanded eFPGA Hard IP (GF 12LP) |
|---|---|---|
| Process Node | 22 nm FD‑SOI | 12‑LP (14 nm) |
| Maximum LUTs | 180 k | > 250 k |
| DSP Resources | Fixed‑function MAC (no SIMD) | SIMD‑enabled MAC with pre‑adder and cascade |
| Configuration Health | No built‑in SEU monitor | Optional Bit‑Health Monitor (≤5 µs latency) |
| BlockRAM | Up to 512 kB | Up to 1 MB |
| Tool Support | Aurora (basic) | Aurora PRO (advanced) |
| Target Applications | General‑purpose SoC acceleration | Radar, EW, satellite comms, imaging, high‑reliability aerospace |
Bottom Line
QuickLogic’s expanded eFPGA Hard IP for GF 12LP delivers a compelling blend of compute density, scalability, and reliability. By integrating a SIMD‑capable DSP MAC, supporting over 250 k LUTs, and offering an on‑the‑fly configuration‑bit health monitor, the IP positions itself as a go‑to solution for next‑generation radar, wireless, and space‑borne systems that demand both performance and fault tolerance. Coupled with the Aurora/Aurora PRO development suite, designers can accelerate time‑to‑market while retaining the flexibility to update hardware post‑production.
For detailed specifications and licensing information, visit quicklogic.com.