This directory contains utility header libraries optimized for SYCL*-compliant FPGA designs. Find usage examples in the ReferenceDesigns and Tutorials directories.
| Filename | Description | Use case examples |
|---|---|---|
constexpr_math.hpp |
Defines utilities for statically computing math functions (for example, Log2 and Pow2). | ReferenceDesigns/merge_sort/ReferenceDesigns/qrdReferenceDesigns/qri |
memory_utils.hpp |
Generic functions for streaming data from memory to a SYCL pipe and vise versa. | ReferenceDesigns/decompress/ |
metaprogramming_utils.hpp |
Defines various metaprogramming utilities (for example, generating a power of 2 sequence and checking if a type has a subscript operator). | ReferenceDesigns/decompress/include/unrolled_loop.hpp |
onchip_memory_with_cache.hpp |
Class that contains an on-chip memory array with a register backed cache to achieve high performance read-modify-write loops. | Tutorials/DesignPatterns/onchip_memory_cache/ReferenceDesigns/decompress/ReferenceDesigns/db/ |
pipe_utils.hpp |
Utility classes for working with pipes, such as PipeArray. | Tutorials/DesignPatterns/pipe_array/ReferenceDesigns/merge_sort/ReferenceDesigns/gzip/ |
rom_base.hpp |
A generic base class to create ROMs in the FPGA using and initializer lambda or functor. | ReferenceDesigns/anr/ |
tuple.hpp |
Defines a template to implement tuples. | ReferenceDesigns/cholesky_inversion/ReferenceDesigns/qri/ReferenceDesigns/cholesky/ |
unrolled_loop.hpp |
Defines a templated implementation of unrolled loops. | Tutorials/DesignPatterns/pipe_array/ReferenceDesigns/cholesky/ReferenceDesigns/anr/ |
exception_handler.hpp |
Defines an exception handler to catch SYCL asynchronous exceptions. | All the samples use it |
| Filename | Description | Use case examples |
|---|---|---|
streaming_cholesky.hpp |
Cholesky decomposition of matrices with pipe interfaces. | ReferenceDesigns/cholesky |
streaming_cholesky_inversion.hpp |
Cholesky-based inversion of matrices with pipe interfaces. | ReferenceDesigns/cholesky_inversion |
streaming_covariance_matrix.hpp |
Standardized covariance matrix computation using pipe interfaces. | ReferenceDesigns/pca |
streaming_eigen.hpp |
Eigen values and Eigen vectors computation of square matrices using pipe interfaces. | ReferenceDesigns/pca |
streaming_matmul.hpp |
Systolic-array-based matrix multiply with pipe interfaces. | ReferenceDesigns/matmul |
streaming_qrd.hpp |
QR decomposition of matrices with pipe interfaces. | ReferenceDesigns/qrd |
streaming_qri.hpp |
QR-based inversion of matrices with pipe interfaces. | ReferenceDesigns/qri |
Code samples are licensed under the MIT license. See License.txt for details.
Third-party program Licenses can be found here: third-party-programs.txt.