Frontpage
The most efficient data center accelerator for high-performance LLM and multimodal deployment
- 512 TFLOPS
- 64 TFLOPS (FP8) x 8 Processing Elements
- 48 GB
- HBM3 Memory Capacity
- 1.5 TB/s
- Memory Bandwidth
- 150 W
- Thermal Design Power
- 512 TFLOPS
- 64 TFLOPS (FP8) x 8 Processing Elements
- 48 GB
- HBM3 Memory Capacity
- 1.5 TB/s
- Memory Bandwidth
- 150 W
- Thermal Design Power
Performance results
Llama 2 7B
Energy Efficiency
Perf/Watt (tokens/s/W)
*Higher value is better
Batch=32, Input Length=2K, Output Length=2K
Batch=16, Input Length=2K, Output Length=2K
Latency
(ms)
*Lower value is better
Batch=1, Sequence Length=128
Batch=1, Sequence Length=128
Throughput
(tokens/s)
*Higher value is better
Batch=16, Input Length=2K, Output Length=2K
Batch=32, Input Length=2K, Output Length=2K
RNGD | H100 | L40S | |
---|---|---|---|
Technology | TSMC 5nm | TSMC 4nm | TSMC 5nm |
BF16/FP8 (TFLOPS) | 256/512 | 989/1979 | 362/733 |
INT8/INT4 (TFLOPS) | 512/1024 | 1979/- | 733/733 |
Memory Capacity (GB) | 48 | 80 | 48 |
Memory Bandwidth (TB/s) | 1.5 | 3.35 | 0.86 |
Host I/F | Gen5 x16 | Gen5 x16 | Gen4 x16 |
TDP (W) | 150 | 700 | 350 |
Disclaimer: Measurements by FuriosaAI internally on current specifications and/or internal engineering calculations. Nvidia results were retrieved from Nvidia website, https://developer.nvidia.com/d..., on February 14, 2024.
Purpose-built for tensor contraction
AI models structure data in tensors of various shapes. The RNGD chip fully exploits parallelism and data reuse by flexibly adapting to each tensor contraction with software-defined tactics and supporting model-wise operator fusion.
Uniquely designed for AI inference deployment, Furiosa TCP unlocks superior utilization, performance and energy efficiency.
AI models structure data in tensors of various shapes. The RNGD chip fully exploits parallelism and data reuse by flexibly adapting to each tensor contraction with software-defined tactics and supporting model-wise operator fusion.
Series RNGD
![Fai mock RNGDS](https://furiosa.imgix.net/fai-mock-RNGDS.png?auto=compress%2Cformat&crop=focalpoint&fit=crop&fm=webp&fp-x=0.5&fp-y=0.5&h=746&q=85&transformer=imgix&w=500)
RNGD-S 2025
Leadership performance for creatives, media and entertainment, and video AI
![Fai mock RNGD](https://furiosa.imgix.net/fai-mock-RNGD.png?auto=compress%2Cformat&crop=focalpoint&fit=crop&fm=webp&fp-x=0.5&fp-y=0.5&h=473&q=85&transformer=imgix&w=500)
RNGD Q3 2024
Versatile cloud and on-prem LLM and Multimodal deployment
- 512 TFLOPS
- 48GB HBM3 Memory Capacity
- 1.5TB/s Memory Bandwidth
- 150W TDP
![Fai mock RNGD Max 1](https://furiosa.imgix.net/fai-mock-RNGDMax-1.png?auto=compress%2Cformat&crop=focalpoint&fit=crop&fm=webp&fp-x=0.5&fp-y=0.5&h=260&q=85&transformer=imgix&w=500)
RNGD-Max 2025
Powerful cloud and on-prem LLM and Multimodal deployment
ENTERPRISE-READY & CLOUD-READY STACK
![Swstack basic](https://furiosa.imgix.net/swstack_basic.png?auto=compress%2Cformat&crop=focalpoint&fit=crop&fm=webp&fp-x=0.5&fp-y=0.5&h=697&q=85&transformer=imgix&w=750)
Workflow Integration
Our platform integrates seamlessly with your existing workflow, offering easy-to-use APIs, support for popular AI frameworks like PyTorch, ONNX, TensorFlow Lite, and compatibility with NumPy data structures.
Model Optimization
Leveraging our advanced compiler, we ensure your models achieve peak performance per watt. With our profiling tools identifying performance bottlenecks, we can further enhance model performance and efficiency.
Flexible Deployment
Furiosa hardware provides the flexibility to choose multiple processing elements (PEs) depending on your workloads. It’s designed for data center adaptability, incorporating containerization and Kubernetes, to facilitate the swift scaling of AI projects.