产品
Programmable Accelerators →

Powered by the latest FPGAs and SoCs from Achronix, AMD, and Altera, our cards are designed and manufactured in-house for enterprise-class performance.

Browse All Programmable Accelerators (FPGA + SoC)

Direct RF (RFSoC)

Custom-Designed Products

Browse by Accelerator Manufacturer:

Achronix

AMD (formerly Xilinx)

Altera (formerly Intel)

Direct RF and Software-Defined Radio →

Card-Level Products

Our RFX family of cards leverage the AMD RFSoC chip with our own analog amplification/filtering and more.

RFX 880/770 (4 in/out + Up to 400G I/O)

RFX 881/771 (Up to 8 in/out + Up to 400G I/O)

RFX-8440

WaveBox RF Servers and Enclosures

Taking RFX in an integrated, modular approach with up to 12 analog in/out in a 1U server.

WaveBox RF

TeraBox™ Servers and Enclosures →

Finding the right server and configuring it can be time-consuming. We've designed TeraBox servers to be ready to go from the start, saving you time if you're developing and for deployment: a robust solution suitable for the toughest challenges.

浏览TeraBox服务器

WaveBox RF

Looking for analog + digital in one box? Our direct RF-focused WaveBox servers and enclosures are a perfect fit!

Browse WaveBox Servers

IP Cores + Example Designs →

BittWare Partners and internal projects gives you an easy way to get started quickly.

Learn about BittWare Partners

Data Movers + DMA + RDMA

Arkville by Atomic Rules

NVMe Bridge by Intelliprop

RDMA by Grovf

Network Offload

UDP Offload by Atomic Rules

TCP Offload by Chevin

精确时间

TimeServo by Atomic Rules

ÜberNIC by LMS

More Network Acceleration

MACsec by Xiphera

IPsec by Xiphera

100G NIC by Corundum

Open Source + Free

Open-Source 100G NIC by Corundum

SmartNIC Shell by BittWare

Loopback by BittWare

Traffic Generator by BittWAre

Pre-Programmed Solutions →

These solutions don't need FPGA programmers, rather they are software configurable and built on BittWare hardware!

Data/Packet Capture + Record

TK242 by Atomic Rules

P2P and Storage

Transparent Compression by Eideticom

Analytics Processor by Eideticom

Cryptographic Accelerator (coming soon)

Financial Services/Fintech

LMS has created ÜberNIC, pre-programmed with the entire network stack in hardware.

ÜberNIC Ultra

ÜberNIC Nano

AI + Machine Learning

Positron Atlas

边缘财团（EdgeCortix

Development Tools →

BittWare 软件开发工具包（SDK）

BittWorks II 工具包

High-Level Tool Flows

oneAPI

宇宙飞船(OpenCL)

Vendor Tools we Suport

Achronix: ACE Design Tools

AMD: Vivado Design Suite

Altera: Quartus Design Software
解决方案
Direct RF and SatCom →

From RFX PCIe cards to our modular integrated WaveBox RF soutions, we have your RFSoC needs covered!

Browse our RF Products:

RFX 880/770 (4 in/out + Up to 400G I/O)

RFX 881/771 (Up to 8 in/out + Up to 400G I/O)

RFX-8440

WaveBox RF

AI/ML and High-Performance Computing (HPC) →

The ability to tailor the application to the silicon is a major win for FPGAs in the HPC space. We've also seen AI/ML use cases where these programmable devices can run more efficiently than even competing GPUs.

Learn About Compute Acceleration

AI/ML Partner Solutions:

Positron AI Acceleration, Built on BittWare

EdgeCortix Edge Inference for AI/ML

英特尔FPGA人工智能套件

Our BWNN Article Explores Using Binary Weights for ML Acceleration

Network & Security →

Networking covers a wide range of use cases, which is why we also have a large portfolio of solutions!

Learn About Network Acceleration

Offload Engines

Need TCP/UDP offload? Our partners offer premium IP cores ready to integrate into your project.

TCP Offload Engine by Chevin

UDP Offload Engine by Atomic Rules

MACsec + IPsec

Featuring Xiphera's IP running these popular security protocols in hardware has never been easier!

MACsec @ Up to 100G

IPsec @ up to 200G

On-demand MACsec/IPsec webinar

RDMA

RDMA over Converged Ethernet (RoCE v2) system implementation and integration, from Grovf.

RDMA IP Core for RoCE v2

SmartNIC

LMS has created ÜberNIC, pre-programmed with the entire network stack in hardware.

ÜberNIC Ultra

ÜberNIC Nano

Open Source + Free

Open-Source 100G NIC by Corundum

SmartNIC Shell by BittWare

Loopback by BittWare

Traffic Generator by BittWare

Accelerators for Financial Services & Fintech →

Whether you're after ultra-low-latency trade performance or simply need a high-performance NIC optimized for fintech, BittWare has a suite of solutions from experts like LMS and Exegy.

LMS ÜberNIC

LMS ÜberNIC Nano

More Fintech Solutions:

nxFramework from Enyx

MAU Accelerator from Myrtle.ai

Broadcast and Video Solutions →

Building on accelerators like FPGAs is a smart way to get more from your investment. Broadcast video is moving away from legacy pre-configured pipelines to software-defined architecture but with hardware doing the heavy lifting.

了解更多

Storage & Peer to Peer Data Movement →

Move your algorithm to the data, not data to the algorithm with our Storage and P2P partner Eideticom.

Learn About Storage Acceleration Solutions

Jump Directly to Eideticom NoLoad Solutions:

Acceleration Processing Unit (APU)

Transparent Compression

Storage Webinars On Demand:
资源

BittWare博客

网络研讨会

文章和白皮书

视频

应用样本

产品支持
关于

关于BittWare

博客

Contact BittWare and Locations

销售渠道

工作机会

支持
搜索

学习资源

文章、白皮书、视频和应用样本

向BittWare学习

我们定期发布内容，可以帮助您选择最适合您需求的技术。浏览我们的文章库、白皮书、视频和一些FPGA应用样本。

需要以100G的速度采集或记录吗？

了解BittWare新的高速数据采集和记录器项目：

网络研讨会

Altera™ FPGA 上的 MACsec 和 IPsec

欢迎参加 Xiphera 关于 MACsec 和 IPsec 的 IP 讨论！现在提供包括问答在内的录音。

使用 COTS 采集/记录数据

我们的三位专家将讨论如何使用 COTS 进行数据采集和记录。由 Mouser Electronics 协办。

HPC 下一代英特尔Agilex FPGAs

特邀嘉宾César González解释了在巴塞罗那超级计算中心使用FPGA的一个应用实例+我们介绍了下一代的Agilex器件。

FPGA小组讨论

三位FPGA专家讨论了 "数据洪水 "以及FPGA帮助解决这一问题的各种方式。

Arkville PCIe Gen4数据转换器

本次网络研讨会介绍了Arkville IP核，用于使用英特尔Agilex FPGA在PCI Gen4上以高达220Gb/s的速度移动数据。

英特尔Agilex计算存储

本网络研讨会主要讨论如何使用英特尔Agilex FPGA和Eideticom的NoLoad IP进行存储加速。

英特尔oneAPI + 2D FFT演示

我们将探讨如何使用英特尔oneAPI编程模型来开发FPGA，并在520N-MX的演示中查看我们的2D FFT。了解这些高级工具如何使您更快地进入市场!

S7t-VG6 Achronix卡

BittWare的S7t-VG6 VectorPath卡具有400Gb/s的速度，并采用7纳米工艺制造。听听Achronix介绍一些创新的新芯片功能，如片上网络和以AI/ML为重点的硬IP!

TeraBox FPGA服务器

从超密集的1U 1400DN到新的200DE边缘服务器，我们的TeraBox系列FPGA服务器已经预先集成，便于开发，并准备好进行企业级规模部署。

计算存储加速

FPGA正在通过使计算更接近FLASH而改变NVMe存储。听听Eideticom与Los Alamos国家实验室的合作以及Myrtle.ai新的基于AI的SEAL推荐引擎。

文章和论文

FPGA神经网络

白皮书

我们研究了FPGA器件上的神经网络推理，说明了它们的优势和劣势。

同态加密加速

白皮书

FPGA加速实现了这一独特的解决方案，允许对加密数据进行计算，而无需解密或共享密钥。

快速开发定制化FPGA解决方案的两种方法

我们在定制解决方案方面有很长的历史。了解我们的一些最佳技巧，用两种方法解决定制需求，更快地进入市场。

交叉栏开关：在oneAPI中共享资源

白皮书

使用oneAPI，利用蝶形交叉线开关有效地共享FPGA资源。包括三个用例。

oneAPI + 使用 HBM2 的 2D FFT

白皮书

针对 oneAPI 的更新
我们在 520N-MX（带 HBM2）上使用英特尔的 oneAPI 加速 2D FFT。涉及主题：HBM2、oneAPI、FFT。

NVMe SSD写入性能

白皮书

在向基于FLASH的NVMe SSD存储长时间写入时，获得最佳性能的基准和细节。

高速数据采集

白皮书

了解有关我们如何通过PCIe实现100Gb/s的细节。涉及的主题：传感器/信号处理。

OVHcloud反DDoS

案例研究

了解OVHcloud如何在其反DDoS解决方案中使用BittWare FPGA卡。所涉主题：安全、网络、数据包处理。

时间戳

白皮书

我们看一下BittWare卡上非常精确的时间戳的使用。所涉主题：网络、数据包处理。

智能网卡外壳

白皮书

学习如何快速建立由FPGA驱动的网卡。涵盖的主题：网络、数据包处理。

回路实例

白皮书

如何使用 AMD CMAC 和配置数据包处理管道。涵盖主题：网络、数据包处理。

合成流量生成器

白皮书

我们如何将我们的XUP-P3卡变成一个用于回归测试的合成流量发生器。

溪流涌动

白皮书

为XUP-P3R客户提供免费的100G网络负载测试应用。涵盖的主题：网络、数据包处理。

解析器

白皮书

与P4相比，我们如何使用HLS C++实现一个数据包解析器。涵盖的主题：高级合成，网络。

通过织物的NVMe

白皮书

使用250族卡进行NVMe存储。所涉主题：闪存、NVMe、存储。

比较RTL和HLS

白皮书

通过建立一个网络RSS模块来比较HLS和RTL。涉及的主题：开发、网络、HLS、RSS。

BWNNs的加速

白皮书

使用OpenCL进行机器学习的Stratix 10编程。所涵盖的主题：OpenCL, 机器学习, Stratix 10.

FPGA加速的NVMe存储解决方案

白皮书

了解FPGA如何以多种方式提升闪存阵列的性能。涵盖的主题：存储加速、NAND闪存、NVMe。

使用FPGA对CNN进行加速

白皮书

在FPGA中使用可变精度来构建更好的机器学习推理网络。主题内容：机器学习、应用定制、Arria 10

用FPGA加快Kafka生产者的速度

白皮书

了解FPGA如何加速这一流行的流媒体平台。涵盖的主题：流媒体数据、压缩。

视频库

StreamSurge网络测试

我们展示了两块运行StreamSurge的XUP-P3R FPGA板，这是我们为客户提供的一个网络测试应用。

液体冷却的FPGA板

请看TeraBox服务器中的四个300+AMP液冷FPGA卡。

320A时的液体冷却

来自SC18展会的现场演示；我们带来了一块安装了液体冷却的XUPVVP板，在320A的FPGA核心功率下运行功率测试负载。

英特尔Stratix 10上的OpenCL

来自超级计算活动的介绍，包括帕德博恩大学的一个案例研究。

Xilinx VU13P的特写

开箱检查极其庞大的VU13P Virtex UltraScale+ FPGA。它甚至比XUP-P3R的VU9P多出50%的逻辑。

带有英特尔Arria 10的A10SA4

BittWare的A10SA4采用了强大的英特尔Arria 10 GX FPGA，在低矮的外形下提供了强大的加速功能。

TeraBox X5000TT扩展机箱

深入了解这一不寻常的解决方案，将几十块FPGA连接到一台主机上。

BWNN白皮书预览

预览BWNN白皮书，你可以在BittWare网站的资源部分找到。

A10PL4的查找/替换演示

使用Axonerve的IP进行查找/替换加速的演示，在基于Intel Arria 10的A10PL4卡上运行。

液体冷却8×卡

Carlos展示了我们如何在一台TeraBox服务器中安装8块XUP-VV4卡，包括液体冷却，并对它们进行远程监控。

LDA技术公司的e4 FPGA机箱

来自LDA技术公司的Sergey带我们了解独特的1U e4 FPGA服务器。

Powered by the latest FPGAs and SoCs from Achronix, AMD, and Altera, our cards are designed and manufactured in-house for enterprise-class performance.

Browse by Accelerator Manufacturer:

Card-Level Products

Our RFX family of cards leverage the AMD RFSoC chip with our own analog amplification/filtering and more.

WaveBox RF Servers and Enclosures

Taking RFX in an integrated, modular approach with up to 12 analog in/out in a 1U server.

Finding the right server and configuring it can be time-consuming. We've designed TeraBox servers to be ready to go from the start, saving you time if you're developing and for deployment: a robust solution suitable for the toughest challenges.

WaveBox RF

Looking for analog + digital in one box? Our direct RF-focused WaveBox servers and enclosures are a perfect fit!

BittWare Partners and internal projects gives you an easy way to get started quickly.

Data Movers + DMA + RDMA

Network Offload

精确时间

More Network Acceleration

Open Source + Free

These solutions don't need FPGA programmers, rather they are software configurable and built on BittWare hardware!

Data/Packet Capture + Record

P2P and Storage

Financial Services/Fintech

LMS has created ÜberNIC, pre-programmed with the entire network stack in hardware.

AI + Machine Learning

High-Level Tool Flows

Vendor Tools we Suport

From RFX PCIe cards to our modular integrated WaveBox RF soutions, we have your RFSoC needs covered!

Browse our RF Products:

The ability to tailor the application to the silicon is a major win for FPGAs in the HPC space. We've also seen AI/ML use cases where these programmable devices can run more efficiently than even competing GPUs.

AI/ML Partner Solutions:

Networking covers a wide range of use cases, which is why we also have a large portfolio of solutions!

Offload Engines

Need TCP/UDP offload? Our partners offer premium IP cores ready to integrate into your project.

MACsec + IPsec

Featuring Xiphera's IP running these popular security protocols in hardware has never been easier!

RDMA

RDMA over Converged Ethernet (RoCE v2) system implementation and integration, from Grovf.

SmartNIC