Intel® Stratix® 10 NX FPGAs

AI-optimized FPGA for high-bandwidth, low-latency artificial intelligence (AI) acceleration applications. The Intel® Stratix® 10 NX FPGA delivers accelerated AI compute solution through AI-optimized compute blocks with up to 143 INT8 TOPS at ~1 TOPS/W;2 in package 3D stacked HBM high-bandwidth DRAM; and up to 57.8G PAM4 transceivers.

Se also: Intel® Stratix® 10 NX FPGAs Design SoftwareDesign StoreDownloadsDocumentation, Community, and Support

Intel® Stratix® 10 NX FPGAs

Intel® Stratix® 10 NX FPGA is an AI-optimized FPGA for high-bandwidth, low-latency artificial intelligence (AI) acceleration applications. The Intel® Stratix® 10 NX FPGA delivers accelerated AI compute solution through AI-optimized compute blocks with up to 143 INT8 TOPS at ~1 TOPS/W1; in package 3D stacked HBM high-bandwidth DRAM; and up to 57.8G PAM4 transceivers.

Features & Benefits

Up to 143 INT8 TOPS or 286 INT4 TOPS1 for High Throughput AI Applications1

The Intel® Stratix® 10 NX FPGA embeds a new type of AI-optimized block called the AI Tensor Block. The AI Tensor Block is tuned for the common matrix-matrix or vector-matrix multiplications used in AI computations, with capabilities designed to work efficiently for both small and large matrix sizes. A single AI Tensor Block achieves 143 INT8 TOPS or 286 INT4 TOPS.1

High Low-Latency Memory Bandwidth for Memory Bound Solution

The integrated memory stacks allow for large, persistent AI models to be stored on-chip, which results in lower latency with large memory bandwidth to prevent memory-bound performance challenges in large models.

Scalable and Flexible I/O Connectivity Bandwidth

Intel® Stratix® 10 NX FPGA include up to 57.8 Gbps PAM4 transceivers to implement multi-node AI inference solutions, reducing or eliminating bandwidth connectivity as a limiting factor in multi-node designs. The Intel® Stratix® 10 NX FPGA also incorporates hard IP such as PCIe Gen3 x16 and 10/25/100G Ethernet MAC/PCS/FEC. These transceivers provide a scalable connectivity solution and the flexibility to adapt to market requirements.

Beyond Peak Performance: Comparing Intel® Stratix® 10 NX AI-Optimized FPGAs and GPUs

Watch this video presentation of the conference paper “Beyond Peak Performance: Comparing the Real Performance of AI-Optimized FPGAs and GPUs”, which focuses on comparing Intel® Stratix® 10 NX AI-optimized FPGAs and GPUs based on achievable performance and measured overheads in realistic inference systems. The conference paper was accepted at the 2020 IEEE International Conference on Field Programmable Technology (FPT).

Read conference paper

Real-Time Text to Speech Synthesis with Intel® Stratix® 10 NX FPGA Demo

Watch this demo for text to speech application implemented on an Intel® Stratix® 10 NX FPGA using WaveNet running on 256 live streams.

Read white paper

Pushing AI Boundaries with…

Learn how Intel® Stratix® 10 NX FPGA can be used for large AI models requiring low-latency, high compute and memory as well as scalability across multi-nodes.

Applications

Natural Language Processing

  • Speech recognition.
  • Speech synthesis.

Security

  • Deep packet inspection.
  • Congestion control identification.
  • Fraud detection.

Real Time Video Analytics

  • Content recognition.
  • Video pre & post-processing.

Intel® Stratix® 10 NX FPGA

Learn more about all device variations and how specifications compare.

Product table: Intel® Stratix® 10 NX FPGA

Porównaj
Nazwa produktu
Status
Data premiery
Elementy logiczne (LE)
Bloki przetwarzania sygnału cyfrowego (DSP)
Maksymalna wbudowana pamięć
Opcje opakowania
Cena
Intel® Stratix® 10 NX 2100 FPGA Launched Q2'20 2073000 3960 239.5 Mb F2597, F2912

Informacje o produktach i wydajności

1

W oparciu o oszacowania wewnętrzne firmy Intel.
Testy mierzą wydajność komponentów w określonych systemach i warunkach testowych. Różnice w sprzęcie, oprogramowaniu lub konfiguracji wpłyną na rzeczywistą wydajność systemów. Aby ocenić wydajność przed dokonaniem zakupu, należy zapoznać się z innymi źródłami informacji. Więcej szczegółowych informacji na temat wydajności i testów porównawczych można znaleźć na stronie www.intel.pl/benchmarks.
Technologie Intel® mogą wymagać zgodnego sprzętu, oprogramowania lub aktywacji usług.
Żaden produkt ani komponent nie jest w stanie zapewnić całkowitego bezpieczeństwa.
Wyniki zostały oszacowane lub zasymulowane. Rzeczywiste koszty i wyniki mogą się różnić.
© Intel Corporation. Intel, logo Intel i inne znaki Intel są znakami towarowymi firmy Intel Corporation lub jej spółek zależnych. Inne nazwy oraz marki mogą być przedmiotem praw osób trzecich.