This Compute Units
sample is a tutorial that showcases a design pattern to allow you to make multiple copies of a kernel, called compute units. Using Compute Units To Duplicate Kernels
Area | Description |
---|---|
What you will learn | How to use this design pattern to create of duplicate compute units |
Time to complete | 15 minutes |
Category | Code Optimization |
The performance of a design can be increased sometimes by making multiple copies of kernels to make full use of the FPGA resources. Each kernel copy is called a compute unit.
This tutorial provides a header file that defines an abstraction for making multiple copies of a single-task kernel. This header can be used in any SYCL*-compliant design and can be extended as necessary.
This sample is part of the FPGA code samples. It is categorized as a Tier 3 sample that demonstrates a design pattern.
flowchart LR
tier1("Tier 1: Get Started")
tier2("Tier 2: Explore the Fundamentals")
tier3("Tier 3: Explore the Advanced Techniques")
tier4("Tier 4: Explore the Reference Designs")
tier1 --> tier2 --> tier3 --> tier4
style tier1 fill:#0071c1,stroke:#0071c1,stroke-width:1px,color:#fff
style tier2 fill:#0071c1,stroke:#0071c1,stroke-width:1px,color:#fff
style tier3 fill:#f96,stroke:#333,stroke-width:1px,color:#fff
style tier4 fill:#0071c1,stroke:#0071c1,stroke-width:1px,color:#fff
Find more information about how to navigate this part of the code samples in the FPGA top-level README.md. You can also find more information about troubleshooting build errors, links to selected documentation, etc.
Optimized for | Description |
---|---|
OS | Ubuntu* 20.04 RHEL*/CentOS* 8 SUSE* 15 Windows* 10, 11 Windows Server* 2019 |
Hardware | Intel® Agilex® 7, Agilex® 5, Arria® 10, Stratix® 10, and Cyclone® V FPGAs |
Software | Intel® oneAPI DPC++/C++ Compiler |
Note: Even though the Intel DPC++/C++ oneAPI compiler is enough to compile for emulation, generating reports and generating RTL, there are extra software requirements for the simulation flow and FPGA compiles.
For using the simulator flow, Intel® Quartus® Prime Pro Edition (or Standard Edition when targeting Cyclone® V) and one of the following simulators must be installed and accessible through your PATH:
- Questa*-Intel® FPGA Edition
- Questa*-Intel® FPGA Starter Edition
- ModelSim® SE
When using the hardware compile flow, Intel® Quartus® Prime Pro Edition (or Standard Edition when targeting Cyclone® V) must be installed and accessible through your PATH.
⚠️ Make sure you add the device files associated with the FPGA that you are targeting to your Intel® Quartus® Prime installation.
The code in this sample is a design pattern to generate multiple compute units using SYCL-compliant template metaprogramming.
The code included in the sample builds on the pipe_array tutorial. This sample includes the header file unroller.hpp
from the pipe_array tutorial, and provides the compute_units.hpp
header as well.
This sample defines a Source
kernel and a Sink
kernel. Data is read from host memory by the Source
kernel and sent to the Sink
kernel through a chain of compute units. The number of compute units in the chain between Source
and Sink
is determined by the following parameter:
constexpr size_t kEngines = 5;
Pipes pass data between kernels. The number of pipes needed is equal to the length of the chain: the number of compute units, plus one. This array of pipes is declared as follows:
using Pipes = PipeArray<class MyPipe, float, 1, kEngines + 1>;
The Source
kernel reads data from host memory and writes it to the first pipe:
void SourceKernel(queue &q, float data) {
q.submit([&](handler &h) {
h.single_task<Source>([=] { Pipes::PipeAt<0>::write(data); });
});
}
At the end of the chain, the Sink
kernel reads data from the last pipe and returns it to the host:
void SinkKernel(queue &q, float &out_data) {
// The verbose buffer syntax is necessary here,
// since out_data is just a single scalar value
// and its size can not be inferred automatically
buffer<float, 1> out_buf(&out_data, 1);
q.submit([&](handler &h) {
accessor out_accessor(out_buf, h, write_only, no_init);
h.single_task<Sink>(
[=] { out_accessor[0] = Pipes::PipeAt<kEngines>::read(); });
});
}
The following line defines functionality for each compute unit and submits each compute unit to the given queue q
. The number of copies to submit is provided by the first template parameter of SubmitComputeUnits
, while the second template parameter allows you to give the compute units a shared base name, e.g., ChainComputeUnit
. The functionality of each compute unit must be specified by a generic lambda expression that takes a single argument (with type auto
) that provides an ID for each compute unit:
SubmitComputeUnits<kEngines, ChainComputeUnit>(q, [=](auto ID) {
auto f = Pipes::PipeAt<ID>::read();
Pipes::PipeAt<ID + 1>::write(f);
});
Each compute unit in the chain from Source
to Sink
must read from a unique pipe and write to the next pipe. As seen above, each compute unit knows its ID; therefore, its behavior can depend on this ID. Each compute unit in the chain will read from pipe ID
and write to pipe ID + 1
.
Note: When working with the command-line interface (CLI), you should configure the oneAPI toolkits using environment variables. Set up your CLI environment by sourcing the
setvars
script located in the root of your oneAPI installation every time you open a new terminal window. This practice ensures that your compiler, libraries, and tools are ready for development.Linux*:
- For system wide installations:
. /opt/intel/oneapi/setvars.sh
- For private installations:
. ~/intel/oneapi/setvars.sh
- For non-POSIX shells, like csh, use the following command:
bash -c 'source <install-dir>/setvars.sh ; exec csh'
Windows*:
C:\"Program Files (x86)"\Intel\oneAPI\setvars.bat
- Windows PowerShell*, use the following command:
cmd.exe "/K" '"C:\Program Files (x86)\Intel\oneAPI\setvars.bat" && powershell'
For more information on configuring environment variables, see Use the setvars Script with Linux* or macOS* or Use the setvars Script with Windows*.
-
Change to the sample directory.
-
Build the program for the Agilex® 7 device family, which is the default.
mkdir build cd build cmake ..
Note: You can change the default target by using the command:
cmake .. -DFPGA_DEVICE=<FPGA device family or FPGA part number>
Alternatively, you can target an explicit FPGA board variant and BSP by using the following command:
cmake .. -DFPGA_DEVICE=<board-support-package>:<board-variant>
Note: You can poll your system for available BSPs using the
aoc -list-boards
command. The board list that is printed out will be of the form$> aoc -list-boards Board list: <board-variant> Board Package: <path/to/board/package>/board-support-package <board-variant2> Board Package: <path/to/board/package>/board-support-package
You will only be able to run an executable on the FPGA if you specified a BSP.
-
Compile the design. (The provided targets match the recommended development flow.)
-
Compile for emulation (fast compile time, targets emulated FPGA device).
make fpga_emu
-
Compile for simulation (fast compile time, targets simulator FPGA device):
make fpga_sim
-
Generate HTML performance report.
make report
The report resides at
compute_units.report.prj/reports/report.html
. You can visualize the kernels and pipes generated by looking at the "System Viewer" section of the report. Note that each compute unit is shown as a unique kernel in the reports, with namesChainComputeUnit<0>
,ChainComputeUnit<1>
, and so on. -
Compile for FPGA hardware (longer compile time, targets FPGA device).
make fpga
-
- Change to the sample directory.
- Build the program for the Agilex® 7 device family, which is the default.
mkdir build cd build cmake -G "NMake Makefiles" ..
Note: You can change the default target by using the command:
cmake -G "NMake Makefiles" .. -DFPGA_DEVICE=<FPGA device family or FPGA part number>
Alternatively, you can target an explicit FPGA board variant and BSP by using the following command:
cmake -G "NMake Makefiles" .. -DFPGA_DEVICE=<board-support-package>:<board-variant>
Note: You can poll your system for available BSPs using the
aoc -list-boards
command. The board list that is printed out will be of the form$> aoc -list-boards Board list: <board-variant> Board Package: <path/to/board/package>/board-support-package <board-variant2> Board Package: <path/to/board/package>/board-support-package
You will only be able to run an executable on the FPGA if you specified a BSP.
-
Compile the design. (The provided targets match the recommended development flow.)
-
Compile for emulation (fast compile time, targets emulated FPGA device).
nmake fpga_emu
-
Compile for simulation (fast compile time, targets simulator FPGA device):
nmake fpga_sim
-
Generate HTML performance report.
nmake report
The report resides at
compute_units.report.prj.a/reports/report.html
. You can visualize the kernels and pipes generated by looking at the "System Viewer" section of the report. Note that each compute unit is shown as a unique kernel in the reports, with namesChainComputeUnit<0>
,ChainComputeUnit<1>
, and so on. -
Compile for FPGA hardware (longer compile time, targets FPGA device).
nmake fpga
-
Note: If you encounter any issues with long paths when compiling under Windows*, you may have to create your
build
directory in a shorter path, for exampleC:\samples\build
. You can then build the sample in the new location, but you must specify the full path to the build files.
- Run the sample on the FPGA emulator (the kernels execute on the CPU).
./compute_units.fpga_emu
- Run the sample on the FPGA simulator device.
CL_CONTEXT_MPSIM_DEVICE_INTELFPGA=1 ./compute_units.fpga_sim
- Run the sample on the FPGA device (only if you ran
cmake
with-DFPGA_DEVICE=<board-support-package>:<board-variant>
)../compute_units.fpga
- Run the sample on the FPGA emulator (the kernels execute on the CPU).
compute_units.fpga_emu.exe
- Run the sample on the FPGA simulator device.
set CL_CONTEXT_MPSIM_DEVICE_INTELFPGA=1 compute_units.fpga_sim.exe set CL_CONTEXT_MPSIM_DEVICE_INTELFPGA=
Note: Hardware runs are not supported on Windows.
PASSED: The results are correct
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.