×
×
×
LOGIN×
SIGNUP×

Please fill all the fields in the form below and submit it to request a 30 days free trial account. Once approved, we will notify you by email with instructions on how to access your free account.

Please note that free accounts are available exclusively to registered companies.

How It Works?

  1. Login to siwave.io cloud based app

    The compiler is hosted on tier1 cloud service providers network and uses secure HTTPS protocol as well as strong authorization and authentication protocols for maximum security.

    Signup to get a no-obligation 30 days free trial account


  2. Load Keras model

    All builtin TFLite operations for both float and quantized Keras models are supported. Get an instant feedback on the validity of the Keras model when it's loaded.


  3. Set system parameters and select target platform

    A super intuitive user interface with the minimal number of options to get going. No need to have any FPGA/ASIC experience or read a user manual to use the compiler, just load the Keras model, set the system parameters (Frequency and Latency) and select target platform to generate an RTL inference in just a couple of minutes. The inference can also be optimized for either frequency or area.

    The compiler generates agnostic RTL that can target any FPGA vendor/device or ASIC implementation

    System Parameters
    System Parameters

  4. Download IP and integrate into top level system

    Deliverables include the encrypted SystemVerilog inference model, a SystemVerilog testbench and a datasheet. A Quartus Platform Designer ready IP component and a Vivado IP core are also provided to facilitate the integration of the inference in a top level system. We also offer optional system design and integration service for generating a complete FPGA or ASIC system.

    The bus interface uses a simple ready/valid handshake protocol that can handle burst sizes up to 4G bytes

    Platform Designer IP
    Quartus Platorm Designer IP
    Vivado IP core
    Vivado IP Core

What sets us apart from the competition


3X Smaller Inferences


The entire inference is implemented in platform agnostic RTL which means there is  no runtime environment and  no microcode to execute on a microprocessor which translates to a considerable  reduction in inference size, in addition, to a much  faster deployment and  less maintenance effort.


2X Higher Performance


The entire inference is implemented in RTL and, therefore, there is  no time consuming context switching between ML hardware and host processor to execute unsupported operations. Siwave's compiler  optimizes each operation based on the system frequency and latency instead of slicing tensors to fit operations into a fixed matrix which translates to a considerable  speedup in inference execution time compared to our competitors when using same frequency or, alternatively, the potential to run the inference at  higher frequency which will, consequently, lead to a  lower latency as demonstrated in the table below.


Comparing AMD/Xilinx DPUCZDX8G BP512 v3.0 (1 core/1 ALU/No Softmax) with Siwave all RTL convolution module on a Zynq UltraScale+ MPSoCs device (xczu9eg-ffvb1156-2-i)
LUTLUTRAMFFBRAMDSPMHz
AMD DPUCZDX8G (BP512)8%2%6%8%4%325
SIWAVE CONVOLUTION RTL MODULE3%0%1%8%1%500