d-Matrix Corsair
Vendor: d-Matrix (USA)
Category: ASIC Dedicated Accelerator
Architecture: DCM (Digital Chiplet Matrix)
Introduction
d-Matrix Corsair is a data center-grade inference platform from US AI inference chip company d-Matrix, based on digital in-memory computing technology, designed specifically for large-scale AI inference workloads. By embedding computation directly into memory arrays, d-Matrix eliminates data movement bottlenecks found in traditional architectures, delivering energy-efficient matrix computation acceleration.
Specifications
| Model | Compute | Memory | Interface | TDP | Process |
|---|---|---|---|---|---|
| Corsair | Inference optimized | In-memory compute architecture | PCIe | Data-center grade | Advanced node |
Official Website
Driver Downloads
Linux
Related Documentation
OS Support
| Windows | Linux | macOS | Android |
|---|---|---|---|
| ❌ | ✅ | ❌ | ❌ |
Version History
| Version | Release Date | Description |
|---|---|---|
| Corsair | Under development | Data center inference platform |
Performance Benchmarks
| Model | Task | Performance Metric |
|---|---|---|
| Corsair | LLM Inference | High energy efficiency |
| Corsair | Matrix Operations | In-memory compute acceleration |
Pricing
| Model | Reference Price | Notes |
|---|---|---|
| Corsair | Contact vendor | Enterprise data center product |
Quick Installation
Linux
# Install d-Matrix SDK
# See official documentation
Architecture Highlights
- DCM (Digital Chiplet Matrix): Chiplet-based modular architecture
- In-Memory Computing: Matrix operations executed directly in memory arrays, eliminating data movement bottlenecks
- Energy-Efficient Inference: Optimized for data center inference workloads
- Scalable Design: Supports flexible deployment from small-scale to large-scale
Model Compatibility
| Model/Framework | Support | Notes |
|---|---|---|
| Transformer LLM | ✅ | Inference optimized |
| Large-scale Inference | ✅ | Data center deployment |
| Quantized Models | ✅ | INT8/FP8 optimized |
Related Products
If you're evaluating alternatives, the following products may also fit your scenario:
- Groq LPU v1 — Groq (LPU Language Processor)
- SambaNova SN40L — SambaNova (ASIC Dedicated Accelerator)
- Cerebras WSE-3 — Cerebras (ASIC Dedicated Accelerator)
- NVIDIA GPU / CUDA — NVIDIA (GPU Graphics Processor)
- AWS Trainium / Inferentia — Amazon AWS (ASIC Dedicated Accelerator)
- Intel Gaudi (Habana) — Intel (ASIC Dedicated Accelerator)
- Cambricon Siyuan MLU — Cambricon (ASIC Dedicated Accelerator)