Developer Guide

FPGA Optimization Guide for Intel® oneAPI Toolkits

Download PDF
ID 767853
Date 7/13/2023
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents
Document Table of Contents x
FPGA Optimization Guide for Intel® oneAPI Toolkits Introduction To FPGA Design Concepts Analyze Your Design Optimize Your Design FPGA Optimization Flags, Attributes, Pragmas, and Extensions Quick Reference Additional Information Document Revision History for the FPGA Optimization Guide for Intel® oneAPI Toolkits
Introduction To FPGA Design Concepts x
FPGA Architecture Overview Concepts of FPGA Hardware Design Methods of Hardware Design How Source Code Becomes a Custom Hardware Datapath Scheduling Mapping Parallelism Models to FPGA Hardware Memory Types
FPGA Architecture Overview x
Adaptive Logic Module (ALM) Lookup Table (LUT) Register Digital Signal Processing (DSP) Block Random Access Memory (RAM) Blocks
Concepts of FPGA Hardware Design x
Maximum Frequency (fMAX) Latency Pipelining Throughput Datapath Control Path Occupancy
How Source Code Becomes a Custom Hardware Datapath x
Mapping Source Code Instructions to Hardware Mapping Arrays and Their Accesses to Hardware
Scheduling x
Dynamic Scheduling Clustering the Datapath Handshaking Between Clusters
Mapping Parallelism Models to FPGA Hardware x
Data Parallelism Task Parallelism
Data Parallelism x
Executing Independent Operations Simultaneously Pipelining
Memory Types x
Kernel Memory Global Memory
Analyze Your Design x
Analyze the FPGA Early Image Analyze the FPGA Image
Analyze the FPGA Early Image x
Review the FPGA Optimization Report Access HLD FPGA Reports in JSON Format
Review the FPGA Optimization Report x
Loop Analysis Bottlenecks Viewer Area Estimates System Viewer Kernel Memory Viewer Schedule Viewer
Analyze the FPGA Image x
Quartus (Static) Summary Intel® FPGA Dynamic Profiler for DPC++ System-level Profiling Using the Intercept Layer for OpenCL* Applications
Quartus (Static) Summary x
Timing Failures
Intel® FPGA Dynamic Profiler for DPC++ x
Measure Kernel Performance Instrument the Kernel Pipeline with Performance Counters (-Xsprofile) Obtain Profiling Data During Runtime Reduce Area Resource Use While Profiling Profiler Analyses of Example SYCL* Design Scenarios Limitations
Obtain Profiling Data During Runtime x
Invoke the Profiler Runtime Wrapper to Obtain Profiling Data Use Intel® VTune™ Profiler
Use Intel® VTune™ Profiler x
Interpret Performance Counter Data
System-level Profiling Using the Intercept Layer for OpenCL* Applications x
Set Up the Intercept Layer for OpenCL* Applications
Optimize Your Design x
Throughput Resource Use
Throughput x
Single Work-item Kernels NDRange Kernels Memory Accesses Pipes Host
Single Work-item Kernels x
Single Work-item Kernel Design Guidelines Loops Single-Cycle Floating-Point Accumulator for Single Work-Item Kernels
Loops x
Refactor the Loop-Carried Data Dependency Relax Loop-Carried Dependency Transfer Loop-Carried Dependency to Local Memory Minimize the Memory Dependencies for Loop Pipelining Unroll Loops Fuse Loops to Reduce Overhead and Improve Performance Optimize Loops With Loop Speculation Remove Loop Bottlenecks Shannonization to Improve FMAX/II Optimize Inner Loop Throughput Improve Loop Performance by Caching On-Chip Memory
Single-Cycle Floating-Point Accumulator for Single Work-Item Kernels x
Strategies for Inferring the Accumulator
Memory Accesses x
Load-Store Units Global Memory Accesses Optimization Perform Kernel Computations Using Local or Private Memory Local and Private Memory Accesses Optimization Annotating Unified Shared Memory Pointers Zero-Copy Memory Access Additional Recommendations
Load-Store Units x
Load-Store Unit Styles Load-Store Unit Modifiers Load-Store Unit Controls
Global Memory Accesses Optimization x
Global Memory Bandwidth Use Calculation Manual Partition of Global Memory Partitioning Buffers Across Different Memory Types (Heterogeneous Memory) Partitioning Buffers Across Memory Channels of the Same Memory Type Ignoring Dependencies Between Accessor Arguments Contiguous Memory Accesses Static Memory Coalescing
Pipes x
Host Pipes
Host Pipes x
Host Pipe Declaration Host Pipe API Host Pipes RTL Interfaces
Host x
Multi-Threaded Host Application Utilizing Hardware Kernel Invocation Queue Double Buffering Host Utilizing Kernel Invocation Queue N-Way Buffering to Overlap Kernel Execution Prepinning Memory Simple Host-Device Streaming Buffered Host-Device Streaming
Double Buffering Host Utilizing Kernel Invocation Queue x
Applying Double-Buffering Using the Intercept Layer for OpenCL* Applications
Resource Use x
Data Types and Operations Kernel Variable Accesses
Data Types and Operations x
Optimize Floating-point Operation Avoid Expensive Functions Variable-Precision Integer and Floating-Point Support
Variable-Precision Integer and Floating-Point Support x
Advantages and Limitations of Arbitrary Precision Data Types Declare and Use the AC Data Types
Declare and Use the AC Data Types x
Declare the ac_int Data Type Declare the ac_fixed Data Type Declare the ac_complex Data Type Declare the ap_float Data Type
Declare the ap_float Data Type x
Conversion Rules for ap_float Operations with Explicit Precision Controls Comparison Operators Additional ap_float Functions Additional Data Types Provided by the ap_float.hpp Header File Quality of Results and the ap_float Data Type
FPGA Optimization Flags, Attributes, Pragmas, and Extensions x
Optimization Flags Optimization Targets Kernel Variables Kernel Attributes Memory Attributes Loop Directives Floating-Point Pragmas Latency Controls (Beta) FPGA Extensions
Optimization Flags x
Specify Schedule FMAX Target for Kernels (-Xsclock=<clock target>) Create a 2xclock Interface (-Xsuse-2xclock) Disable Burst-Interleaving of Global Memory (-Xsno-interleaving=<global_memory_name>) Force Ring Interconnect for Global Memory (-Xsglobal-ring) Force a Single Store Ring to Reduce Area (-Xsforce-single-store-ring) Force Fewer Read Data Reorder Units to Reduce Area (-Xsnum-reorder) Disable Hardware Kernel Invocation Queue (-Xsno-hardware-kernel-invocation-queue) Modify the Handshaking Protocol Between Clusters (-Xshyper-optimized-handshaking) Disable Automatic Fusion of Loops (-Xsdisable-auto-loop-fusion) Fuse Adjacent Loops With Unequal Trip Counts (-Xsenable-unequal-tc-fusion) Pipeline Loops in Non-task Kernels (-Xsauto-pipeline) Control Semantics of Floating-Point Operations (-fp-model=<value>) Modify the Rounding Mode of Floating-point Operations (-Xsrounding=<rounding_type>) Global Control of Exit FIFO Latency of Stall-free Clusters (-Xssfc-exit-fifo-type=<value>) Enable the Read-Only Cache for Read-Only Accessors (-Xsread-only-cache-size=<N>) Control Hardware Implementation of the Supported Data Types and Math Operations (-Xsdsp-mode=<option>) Generate Register Map Wrapper (-Xsregister-map-wrapper-type)
Optimization Targets x
Minimum Latency Flow Maximum Throughput Without Area Optimization Heuristics Flow
Kernel Attributes x
Specify Schedule FMAX Target for Kernels Specify a Workgroup Size Specify Number of SIMD Work Items Omit Hardware that Generates and Dispatches Kernel IDs Omit Hardware to Support the no_global_work_offset Attribute in parallel_for Kernels Reduce Kernel Area and Latency
Loop Directives x
disable_loop_pipelining Attribute initiation_interval Attribute ivdep Attribute loop_coalesce Attribute max_concurrency Attribute max_interleaving Attribute speculated_iterations Attribute unroll Pragma Loop Fuse Functions and nofusion Attribute max_reinvocation_delay Attribute (Beta)
FPGA Extensions x
Pipes Extension Asynchronous Parallelism Within Kernels (task_sequence) device_global Extension (Beta)
Pipes Extension x
Key Properties of a Pipe Accessing Pipes The pipe Class and its Use I/O Pipes Characteristics of Pipes Restrictions of Pipes Guidelines for Designing Pipes Pipe and Atomic Fence
Asynchronous Parallelism Within Kernels (task_sequence) x
Task Functions task_sequence Use Cases
Quick Reference x
Algorithmic C Data Types Floating Point Pragmas FPGA Accessor Properties FPGA Extensions FPGA Kernel Attributes FPGA Local Memory Function Latency Control Properties (Beta) FPGA LSU Controls FPGA Loop Directives FPGA Memory Attributes FPGA Optimization Flags Pipe API
Document Revision History for the FPGA Optimization Guide for Intel® oneAPI Toolkits x
Notices and Disclaimers
FPGA Optimization Guide for Intel® oneAPI Toolkits
Introduction To FPGA Design Concepts
Analyze Your Design
Optimize Your Design
FPGA Optimization Flags, Attributes, Pragmas, and Extensions
Quick Reference
Additional Information
Document Revision History for the FPGA Optimization Guide for Intel® oneAPI Toolkits

Document Revision History for the FPGA Optimization Guide for Intel® oneAPI Toolkits

Date Release Version Changes
July 2023 2023.2
  • Revised the note about SYCL streams support and mentioned about using sycl::ext::oneapi::experimental::printf for debugging kernels.
  • Updated all occurrences of -Xsno-interleaving=<global_memory_type> to -Xsno-interleaving=<global_memory_name>.
  • Revised the information in the following topics:
    • Optimization Targets
    • Minimum Latency Flow
    • Loop Analysis
    • Quartus (Static) Summary
    • Pipes
    • Simple Host-Device Streaming
    • Buffered Host-Device Streamting
  • Added the following new topics:
    • Generate Register Map Wrapper (-Xsregister-map-wrapper-type)
    • Maximum Throughput Without Area Optimization Heuristics Flow
    • Create a 2xclock Interface (-Xsuse-2xclock)
    • Host Pipes
    • Host Pipe Declaration
    • Host Pipe API
    • Host Pipes RTL Interfaces
March 2023 2023.1
  • Added max_reinvocation_delay loop attribute.
  • Updated the limitations of AC data types in Advantages and Limitations of Arbitrary Precision Data Types.
  • Added another option to the -Xshyper-optimized-handshaking compiler flag in Modify the Handshaking Protocol Between Clusters (-Xshyper-optimized-handshaking).
  • Updated the values of supported data types and math operations in Control Hardware Implementation of the Supported Data Types and Math Operations.
  • Rebranded Intel® Agilex™ as Intel Agilex® 7.
  • Added device_globals extension.
  • Added Minimum Latency Flow.
  • Added Optimization Levels.
  • Revised the information in Perform Kernel Computations Using Local or Private Memory.
  • Reorganized all extensions under FPGA Extensions.
  • Added examples of declaring ac_complex variables in Declare the ac_complex Data Type.
December 2022 2023.0
  • Revised all references to #include <CL/...> as #include <sycl/...>.
  • Revised all references to using namespace cl::sycl; as using namespace sycl;.
  • Revised all references to dpcpp compiler driver to icpx -fsycl.
  • Added a note about avoiding adding the result of the convert_to function to another ap_float variable in Conversion Rules for ap_float.
  • Added a note about constant propagation optimization technique in Operations with Explicit Precision Controls.
  • Added additional about values supported by N for the [[intel::num_simd_work_items(N)]] attribute in Specify Number of SIMD Work-Items.
  • Revised Latency Controls topic to include a section on using latency controls with stall-free loop.
  • Added information about ihc::FPsingle and ihc::FPdouble to Declare the ap_float Data Type and Additional Data Types Provided by the ap_float.hpp Header File.
  • Added new sections on explicit conversion functions in Declare the ac_fixed Data Type and Declare the ac_int Data Type topics.
  • Removed deprecation notice for hls_float data type renamed to ap_float.
  • Added additional information about using the Intel® oneAPI FPGA Reports tool in Review the Optimization Report (report.html).
September 2022 2022.3
  • Added a note about setting the safelen parameter with 0 or 1 in ivdep Attribute.
  • Added a note about applying the ivdep attribute to an array in ivdep Attribute.
  • Minor update to example code in max_interleaving Attribute.
  • Updated all images and made moderate updates in all topics in the Analyze the FPGA Early Image section.
  • Made minor updates in Pipes topic.
  • Updated Latency Controls (Beta) topic completely.
  • Added a note about using the buffer_location<index> on BSPs with heterogeneous memory support in combination with read-only cache in Partitioning Buffers Across Different Memory Types (Heterogeneous Memory).
  • Revised the guidance in Timing Failures.
  • Made minor revisions in Zero-Copy Memory Access and Prepinning Memory topics.
  • Added the following new topics:
    • System of Tasks (task_sequence) Extension
    • Task Functions
    • task_sequence Use Cases
April 2022 2022.2
  • Changed the document title Intel® oneAPI DPC++ FPGA Optimization Guide to FPGA Optimization Guide for Intel® oneAPI Toolkits.
  • Replaced all general references to DPC++ with SYCL.
  • Added new loop functions to the existing list in FPGA Loop Directives.
  • Added a new limitation and removed some existing limitations in Advantages and Limitations of Arbitrary Precision Data Types.
  • Replaced all references to https://hlslibs.org/ with a reference to the documentation at https://github.com/hlslibs/ac_types/blob/v3.7/pdfdocs/ac_datatypes_ref.pdf.
  • Made a minor update to the description in The pipe Class and its Use.
  • Modified all occurrences of [[cl::reqd_work_group_size(Z, Y, X)]] to [[sycl::reqd_work_group_size(Z, Y, X)]]. [[cl::reqd_work_group_size(Z, Y, X)]] is now deprecated.
  • Added the following new topics:
    • Control Hardware Implementation of the Supported Data Types and Math Functions (-Xsdsp-mode=<option>)
    • Latency Controls
December 2021 2022.1
  • Added new attributes to the existing list in FPGA Loop Directives and FPGA Optimization Flags.
  • Added the following new topics:
    • Global Control of Exit FIFO Latency of Stall-free Clusters (-Xssfc-exit-fifo-type=<value>)
    • nofusion Attribute
    • Enable the Read-Only Cache for Read-Only Accessors (-Xsread-only-cache-size=<N>)
    • Renamed hls_float data type to ap_float. Added support for the header files <sycl/ext/intel/ac_types/ap_float.hpp> and <sycl/ext/intel/ac_types/ap_float_math.hpp>. In a future release, the header files <sycl/ext/intel/ac_types/hls_float.hpp> and <sycl/ext/intel/ac_types/hls_float_math.hpp> will be removed.
    • Updated few limitations in Advantages and Limitations of Arbitrary Precision Data Types.
September 2021 2021.4
  • Changed all occurrences of the namespace sycl::ONEAPI:: to sycl::ext::oneapi::.
  • Changed all occurrences of the namespace sycl::INTEL:: to sycl::ext::intel::.
  • Changed all occurrences of the header #include <CL/sycl/INTEL/...> to #include <sycl/ext/intel/...>.
  • Changed clang++ to dpcpp.
  • Changed all occurrences of [[intel::ii(n)]] to [[intel::initiation_interval(n)]].
  • Minor update about the types of FIFOs reported in the System Viewer.
  • Added a note about multi-process execution in the installed FPGAs in Multi-Threaded Host Application.
  • Added an additional value on to #pragma fp contract(fast/off).
  • Updated the code snippets in Specify Schedule FMAX Target for Kernels and Memory Attributes.
  • Modified the flag description for AC type in Variable-Precision Integer and Floating-Point Support.
  • Modified the list of limitations in Advantages and Limitations of Arbitrary Precision Data Types.
  • Added information about debugging in Declare the ac_int Data Type.
  • Added support for simulation flow.
  • Changed all occurrences of Mentor Graphics* to Siemens EDA* (formerly Mentor Graphics) .
  • Changed the noinit attribute to no_init.
  • Modified the description of [[intel::kernel_args_restrict]] attribute in Global Memory Accesses Optimization.
  • Replaced the flag -Xsffpcontract=<value> with -fp-model=<value> and updated its description in Reducing Floating-Point Rounding Operations.
  • Removed few limitations of shift register in Kernel Variable Accesses.
  • Added the following new topics:
    • Modify the Rounding Mode of Floating-point Operations (-Xsrounding=<rounding_type>)
    • Reduce Kernel Area and Latency ([[intel::use_stall_enable_clusters]] attribute)
    • Timing Failures
July 2021 2021.3 Bug fixes.
June 2021 2021.3
  • Added more information about streaming data using I/O pipes and faking I/O pipes in I/O Pipes.
  • Added description about the local memory function in Kernel Memory.
  • Added a new section to describe the global memory bandwidth use calculation in Global Memory Accesses Optimization.
  • In System Viewer, added description about the global memory view, changed the Graph Viewer report name to System Viewer, updated all images and made minor updates in the remaining part of the topic.
  • Updated the description and image in Schedule Viewer.
  • Updated most of the code snippets to reflect the latest coding practices.
  • Made minor updates to the description of ivdep Attribute and loop_coalesce Attribute.
  • Added MLABs to the description in Perform Kernel Computations Using Local or Private Memory.
  • Modified the description in Strategies for Inferring the Accumulator to remove references to the -Xsfp-relaxed flag.
  • Updated the compilation options in Variable-Precision Integer and Floating-Point Support.
  • Updated the limitations in Advantages and Limitations of Arbitrary Precision Data Types.
  • Merged Ignoring Dependencies Between Accessor Arguments topic with Global Memory Accesses Optimization.
  • Updated the list of attributes in FPGA Kernel Attributes.
  • Added the following new topics:
    • Shannonization to Improve FMAX/II
    • Simple Host-device Streaming
    • Buffered Host-device Streaming
    • N-Way Buffering
    • Optimize Inner Loop Throughput
    • Improve Loop Performance by Caching On-Chip Memory
    • FPGA Local Memory Function
    • Pipelining Loops in Non-task Kernels (-Xsauto-pipeline)
    • Pipe and Atomic Fence
    • Reducing Floating-Point Rounding Operations (-Xsffp-contract=fast)
    • FPGA Accessor Properties
  • Removed the following topics because the respective flags are now deprecated:
    • Relax the Order of Floating-Point Operations (-Xsfp-relaxed)
    • Reduce Floating-Point Rounding Operations (-Xsfpc)
  • Removed Tree Balancing and Rounding Operations topic from Optimize Floating-point Operation.
March 2021 2021.2
  • Updated the topic Cluster the Datapath completely including the diagrams.
  • Updated the dependency graph in Mapping Source Code Instructions to Hardware.
  • Updated Hyper-Optimized Handshaking Data Flow diagram in Handshaking Between Clusters topic.
  • Made minor updates in Executing Independent Operations Simultaneously.
  • Updated the topic Pipelining completed including the diagrams.
  • Added a note about the enablement of fast math operations for floating point operations in Data Types and Operations.
  • Updated the topic Access HLD FPGA Reports in JSON Format to include details about the Bottlenecks viewer.
  • Made minor update to a note in Specify Number of SIMD Work-Items
  • Added the following new topics:
    • Floating Point Optimizations
    • Floating Point Pragmas
    • Specify Schedule FMAX Target for Kernels
    • Bottlenecks Viewer
    • Loop Bottlenecks
    • Variable-Precision Integer and Floating-Point Support
    • Advantages and Limitations of Arbitrary Precision Data Types
    • Declare and Use the AC Data Types
    • Declare the ac_int Data Type
    • Declare the ac_fixed Data Type
    • Declare the ac_complex Data Type
    • Declare the hls_float Data Type
    • Conversion Rules for hls_float
    • Operations with Explicit Precision Controls
    • Comparison Operators
    • Additional hls_float Functions
    • Additional Data Types Provided by hls_float.hpp
December 2020 2021.1 First major release.
Level Two Title