Machine Learning

with FPGAs on the cloud,

Coral FPGA Resource manager

Program against your FPGAs like it’s a single pool of accelerators

FPGA-based Machine Learning Accelerators

Speedup your ML applications in the cloud or on-prem, instantly.

We develop high performance accelerators-as-a-service for widely-used cloud computing applications and data analytics such as: machine learning (classification, clustering, regression) compression, encryption, financial applications

The hardware accelerators of InAccel are fully compatible to Apache Spark ML framework and allows faster execution of Spark-based applications based on machine learning (ML, MLlib) libraries. InAcce's framework utilize FPGAs to accelerated the ML kernel without changing your Spark application.

InAccel develops hardware accelerators that are compatible with the Amazon AWS f1 platform. Use the hardware accelerator as an IP to speedup your application, seamlessly. InAccel is an official seller of Amazon AWS Marketplace. InAccel offers the accelerators in the form of IP cores (AFIs) and provides the integrated AMIs for the seamless integration in your code. ​

We develop custom hardware application for FPGAs tailored-made to your applications. The accelerators can be deployed in the Amazon AWS F1 and the Intel FPGA Platforms. ​Use the hardware accelerators to speedup your applications in terms of throughput and latency (execution time).

Machine Learning Accelerators

ML IP cores

  • Logistic Regression
  • K-means clustering
  • XGboost
High performance IP cores for FPGAs for Machine Learning training and prediction that can be deployed on the cloud or on-prem with zero code changes.

Compatible with cloud providers like AWS, Alibaba, Huawei and Nimbix.



Accelerated Machine Learning

InAccel provides hardware accelerators as FPGAs IPs for widely-used machine learning algorithms such as:

Classification and Regression

Logistic Regression
Naive Bayes



Gradient Boosted Trees


Recommendation engines

Alternating Least Squares (ALS)

Data Processing

Improve the throughput and the latency of your applications that are computation intensive using accelerators for data processing:


Zstandard (Zstd)





Financial algorithms

Speedup your application with hardware accelerators as IP block for the following algorithms




Try on AWS

Try now the Accelerated Machine Learning suite on Amazon AWS

InAccel advantages

Compatible with Amazon AWS

All accelerators are compatible with the Amazon AWS F1 instances. AWS compatibility allows easy and fast deployment of the accelerators and seamless integration with your current AWS applications.

Seamless integration with your code

InAccel provides all the required APIs for the seamless integration of the accelerators without any modifications on your original code.

Acceleration of your code

Accelerators from InAccel provide up to 15x speedup compared to contemporary processors in typical servers.

Supported APIs