This document describes DOCA DPACC Compiler and instructions about DPA toolchain setup and usage.
DPACC is a high-level compiler for the DPA processor which compiles code targeted for the data-path accelerator (DPA) processor into a device executable and generates a DPA program.
The DPA program is a host library with interfaces encapsulating the device executable. This DPA program is linked with the host application to generate a host executable. The host executable can invoke the DPA code through FlexIO runtime API.
DPACC uses DPA compiler (
dpa-clang) to compile code targeted for DPA. dpa-clang is part of the DPA toolchain package which is an LLVM-based cross-compiling bare-metal toolchain. It provides Clang compiler, LLD linker targeting DPA architecture, and other utilities.
|Device||DPA as present on the BlueField DPU|
|Host||CPU that launches the device code to run on the DPA|
|Device function||Any C function that runs on the DPA device|
|DPA global function||Device function that is the point of entry when offloading any work on DPA|
|Host compiler||Compiler used to compile the code targeting the host CPU|
|Device compiler||Compiler used to compile code targeting the DPA|
|DPA program||Host library that encapsulates the DPA device executable (
- DPA device code – C programs, targeted to run on the DPA. DPA device code may contain one or more entry functions.
- Host application code – the corresponding host application. Please refer to NVIDIA DOCA DPA Subsystem Programming Guide for more details
- Runtime – FlexIO or DOCA DPA library provides the runtime
The generated DPA program, when linked with A host application, results in a host executable which also contains the device executable. The host application is in charge of loading the device executable on the device.
DPA device code is a C code with some restrictions and special definitions.
The DPA is programmed using a subset of the C11 language standard. The compiler documents any constructs that are not available. Language constructs, where available, retain their standard definitions.
- Use of C thread local storage is not allowed for any variables
- Identifiers with
_dpaccprefix are reserved by the compiler. Use of such identifiers may result in an error or undefined behavior.
- DPA processor does not have native floating-point support; use of floating point operations is disabled
A remote procedure call function is a synchronous call that triggers work in DPA and waits for its completion. These functions return a type
uint64_t value. They are annotated with a
A DPA global function is an event handler device function referenced from the host code. These functions do not return anything. They are annotated with a
For more information, refer to the NVIDIA DOCA DPA Subsystem Programming Guide.
- Global functions must have
voidreturn type and RPC functions must have
- Annotated functions cannot accept C pointers and arrays as arguments (e.g.,
void my_global (int *ptr, int arr))
- Annotated functions cannot accept a variable number of arguments
- Inline specifier is not allowed on snnotated functions
User-defined data types, when used as global function arguments, require special handling. They must be annotated with a
If the user-defined data type is
typedef statement must be annotated with a
__dpa_global__ attribute along the data type itself.
- They must have a copy of the definition in all translation units where they are used as global function arguments
- They cannot have pointers, variable length arrays, and flexible arrays as members
- Fixed-size arrays as C structure members are supported
- These characteristics apply recursively to any user-defined/
typedef'd types that are members of an annotated type
DPACC processes all annotated functions along with annotated types and generates host and device interfaces to facilitate the function launch.
DPA features such as fences and processor-specific instructions are exposed via intrinsics by the DPA compiler. All intrinsics defined in the header file
dpaintrin.h are guarded by the
DPA_INTRIN_VERSION_USED macro. The current
#define DPA_INTRIN_VERSION_USED (DPA_INTRIN_VERSION(1, 3)) #include <dpaintrin.h> … __dpa_thread_writeback_window(); // Fence for write barrier
For more information, please refer to NVIDIA DOCA DPA Subsystem Programming Guide.
|Host compiler||Compiler specified through
|Device compiler||The default device compiler is the "DPA compiler". Installing the DPACC package also installs the DPA compiler binaries
|FlexIO SDK and C library||Available as part of the DOCA software package. DPA toolchain does not provide C library and corresponding headers. Users are expected to use the C library for DPA from the FlexIO SDK.|
- DPACC version 1.5.0
- See NVIDIA DOCA DPA Subsystem Programming Guide for other component versions
DPACC can produce DPA programs in a single command by accepting all source files as input. DPACC also offers the flexibility of producing DPA object files or libraries from input files.
DPA object files contain both host stub objects (DPACC-generated interfaces) and device objects. These DPA object files can later be given to DPACC as input to produce the DPA library.
|Phase||Option Name||Default Output File Name|
|Compile input device code files to DPA object files||--compile or -c||.dpa.o appended to the name of each input source file|
|Compile and link the input device code files/DPA object files, and produce a DPA program||No specific option||No default name, output file name must be specified|
|Compile and build DPA library from input device code files/DPA object files||
||No default name, output library name must be specified|
DPACC can accept the following file types as input:
|File Extension||File Type||Description|
|.c||C source file||DPA device code|
|.dpa.o||DPA object file||Object file generated by DPACC, containing both host and device objects|
||DPA object archive||An archive of DPA object files. User can generate this archive from DPACC-generated DPA objects.|
Based on the mode of operations, DPACC can generate the following output files:
|Output File Type||Input Files|
|DPA object file||C source files|
|DPA program||C source files, DPA object files, and/or DPA object archives|
|DPA library (DPA host library and DPA device library)||C source files, DPA object files, and/or DPA object archives|
The following provides the commands to generate different kinds of supported output file types for each input file type:
|C source file||DPA program||
|DPA object||DPA program||
|DPA object archive||DPA program||
- DPACC-generated host stubs which facilitate invoking a DPA global function from the host application
- Device executable, generated by DPACC by compiling input DPA device code
DPA program library must be linked with the host application that contains appropriate runtime APIs to load the device executable onto DPA memory.
DPACC produces DPA object files in compile-only mode. A DPA object is an object file for the host machine. In a DPA object, the device object generated by compiling the input device code file is placed inside a specific section of the generated host stubs object. This process is repeated for each input file.
- DPA device library – contains device objects generated from input files
- DPA host library – contains host interface objects corresponding to the device objects in DPA device library
The DPA device library is consumed by DPACC during DPA-program generation and the DPA host library can optionally be linked with other host code and be distributed as the host library. Both libraries are generated as static archives.
The following diagram illustrates DPACC compile-and-link mode trajectory.
This is a one-step mode that accepts C source files or DPA object files and produces the DPA program. Specifying the output library name is mandatory in this mode. Example commands:
$ dpacc in1.c in2.c -o myLib1.a -hostcc=gcc # Takes C sources to produce myLib1.a library $ dpacc in3.dpa.o in4.dpa.o -o myLib2.a -hostcc=gcc # Takes DPA object files to produce myLib2.a library $ dpacc in1.c in3.dpa.o -o myLib3.a -hostcc=gcc # Takes C source and DPA object to produce myLib3.a library
This mode accepts C source code and produces
.dpa.o object files. These files can be given to DPACC to produce the DPA program. The mode is invoked by the
The user can explicitly provide the output object file name using the
$ dpacc -c input1.c -hostcc=gcc # Produces input1.dpa.o $ dpacc -c input3.c input4.c -hostcc=gcc # Produces input3.dpa.o and input4.dpa.o $ dpacc -c input2.c -o myObj.dpa.o -hostcc=gcc # Produces myObj.dpa.o
This mode accepts C source files or DPA object files and produces the DPA program. Specifying the output DPA library name is mandatory in this mode. Example commands:
$ dpacc in1.c in2.c -o libdummy1 -hostcc=gcc -gen-libs # Takes C sources to produce libdummy1_host.a and libdummy_device.a archives $ dpacc in3.dpa.o in4.dpa.o -o libdummy2 -hostcc=gcc -gen-libs # Takes DPA object files to produce libdummy2_host.a and libdummy2_device.a archives $ dpacc in1.c in3.dpa.o -o outdir/libdummy3 -hostcc=gcc -gen-libs # Takes C source and DPA object to produce outdir/libdummy3_host.a and outdir/libdummy3_device.a archives
Usage: dpacc <list-of-input-files> -hostcc=<path> [other options] Helper Flags: -h, --help Print help information about DPACC -V, --version Print DPACC version information -v, --verbose List the compilation commands generated by this invocation while also executing every command in verbose mode -dryrun, --dryrun Only list the compilation commands generated by DPACC, without executing them -keep, --keep Keep all intermediate files that are generated during internal compilation steps in the current directory -keep-dir, --keep-dir Keep all intermediate files that are generated during internal compilation steps in the given directory -optf, --options-file <file>,... Include command line options from the specified file
|List of one or more input files||All||List of C source files or DPA object file names. Specifying at least one input file is mandatory. A file with an unknown extension is treated as a DPA object file.|
|-hostcc, --hostcc <path>||All||Specify the host compiler. This is typically the native compiler present on the host system.|
|-o, --output-file <file>||Compile-and-link/library generation||Specify name and location of the output file.|
||Specify DPA application name for the DPA program. This option is required if multiple DPA programs are part of a host application because each DPA application must have a unique name. Default name is
||Enable link-time optimization (LTO) for device code. Specify this option during compilation along with an optimization level in
||Specify the list of options to pass to the device compiler.|
|-devicelink-options, --devicelink-options <options>, ...||Specify the list of options to pass during device linking stage.|
||Specify the list of device libraries including their names (in
|-I, --common-include-path <path>, ...||Specify include search paths common to host and device code compilation. FlexIO headers paths are included by DPACC by default.|
|-devicecc, --devicecc <path>||Specify the device compiler. By default, DPACC invokes dpa-clang.|
|-o, --output-file <file>||
Specify name and location of the output file.
|-hostcc-options, --hostcc-options <options>, ...||Specify the list of options to pass to the host compiler.|
||Generate a DPA library from input files|
- Only the default linker script is supported with LTO
- Using options
-devicecc-optionsis not supported when LTO is enabled
- Fat objects containing both LLVM bitcode and ELF representation are not supported
- Thin LTO is not supported
During compilation, LLVM generates the object as bitcode IR (intermediate representation) when LTO is enabled instead of ELF representation. The bitcode IR generated by the DPA compiler is only guaranteed to be compatible within the same version of DPACC. All objects involved in link-time optimization (enabled with
-flto) must be built with the same version of DPACC.
This example shows how to build DPA libraries using DPACC. Libraries for DPA typically contain two archives, one for the host and one for the device.
dpacc input.c -hostcc=gcc -o lib<name> -gen-libs -hostcc-options="-fPIC"
This command generates the output files
- Generating a static host library:
ar x lib<name>_host.a # Extract objects to generate *.o ar cr lib<name>.a <*src.host.o> *.o # Generate final static archive with all objects
- Generating a shared host library:
gcc -shared -o lib<name>.so <*src.host.o> -Wl,-whole-archive -l<name>_host -Wl,-no-whole-archive # Link the generated archive to build a shared library
The DPA device library generated by DPACC using
-gen-libs as part of a DPA library can be consumed by DPACC using the
dpacc input.c -hostcc=gcc -o libInput.a -device-libs="-L <path-to-library> -l<libName>"
Link-time optimizations can be enabled using
-flto along with an optimization level specified for device compilation.
dpacc input1.c -hostcc=gcc -c -flto -devicecc-options="-O2" dpacc input2.c -hostcc=gcc -c -flto -devicecc-options="-O2" dpacc input1.dpa.o input2.dpa.o -hostcc=gcc -o libInput.a
This example includes headers for device compilation using
devicecc-options and host compilation using
hostcc-options. You may also specify headers for any compilation on both the host and device side using the
dpacc input.c -hostcc=gcc -o libInput.a -I <common-headers-path> -devicecc-options="-I <device-headers-path>" -hostcc-options="-I <host-headers-path>"
dpa-clang is a compiler driver for accessing the Clang/LLVM compiler, assembler, and linker which accepts C code files or object files and generates an output according to different usage modes.
Refer to the following resources for detailed user guide and command line references:
- Clang Compiler User’s Manual
- Clang command line argument reference
- Target-dependent compilation options
dpa-clang <list-of-input-files> [other-options]
LLD is the default linker provided in the DPA toolchain. Linker-related options are passed to through the compiler driver.
For more information, please refer to the LLD command line reference.
dpacc-extract is a tool for extracting a device executable out of a DPA program or a host executable containing DPA program(s).
To execute dpacc-extract:
Usage: dpacc-extract <input-file> -o=<output-file> [other options] Helper Flags: -o, --output-file Specify name of the output file -app-name, --app-name <name> Specify name of the DPA application to extract -h, --help Print help information about dpacc-extract -V, --version Print dpacc-extract version information -optf, --options-file <file>,... Include command line options from the specified file
|Input file||DPA program or host executable containing DPA program. Specifying one input file is mandatory.|
||Specify name and location of the output device executable.|
||Specify name of the DPA application to extract. Mandatory if input file has multiple DPA apps.|
The dpa-objdump utility prints the contents of object files and final linked images named on the command line.
For more information, please refer to the Objdump command line reference. Commonly used dpa-objdump options:
dpa-ar is a Unix ar-compatible archiver.
For more information, please refer to the Archiver command line reference.
The dpa-nm utility lists the names of symbols from object files and archives.
For more information, please refer to the NM tool command line reference.
|-mcpu=nv-dpa-bf3||Option to choose micro-architecture and ABI for DPA processor.
|-mrelax/-mno-relax||Option to enable/disable linker relaxations.|
|-I <dir>||Option to include header files present in
||Option to link against libraries|
|-fdebug-macro||Option to emit macro debugging information. This option enables macro-debugging similar to GCC option
- Objects produced by LLD are not compatible with those generated by any other linker.
- The default debugging standard of the DPA compiler is DWARFv5. GDB versions <10.1 have issues processing some DWARFv5 features. Use the option
-devicecc-options="-gdwarf-4"with DPACC to debug with GDB versions <10.1.
This document is provided for information purposes only and shall not be regarded as a warranty of a certain functionality, condition, or quality of a product. NVIDIA Corporation nor any of its direct or indirect subsidiaries and affiliates (collectively: “NVIDIA”) make no representations or warranties, expressed or implied, as to the accuracy or completeness of the information contained in this document and assume no responsibility for any errors contained herein. NVIDIA shall have no liability for the consequences or use of such information or for any infringement of patents or other rights of third parties that may result from its use. This document is not a commitment to develop, release, or deliver any Material (defined below), code, or functionality.
NVIDIA products are sold subject to the NVIDIA standard terms and conditions of sale supplied at the time of order acknowledgement, unless otherwise agreed in an individual sales agreement signed by authorized representatives of NVIDIA and customer (“Terms of Sale”). NVIDIA hereby expressly objects to applying any customer general terms and conditions with regards to the purchase of the NVIDIA product referenced in this document. No contractual obligations are formed either directly or indirectly by this document.
NVIDIA products are not designed, authorized, or warranted to be suitable for use in medical, military, aircraft, space, or life support equipment, nor in applications where failure or malfunction of the NVIDIA product can reasonably be expected to result in personal injury, death, or property or environmental damage. NVIDIA accepts no liability for inclusion and/or use of NVIDIA products in such equipment or applications and therefore such inclusion and/or use is at customer’s own risk.
NVIDIA makes no representation or warranty that products based on this document will be suitable for any specified use. Testing of all parameters of each product is not necessarily performed by NVIDIA. It is customer’s sole responsibility to evaluate and determine the applicability of any information contained in this document, ensure the product is suitable and fit for the application planned by customer, and perform the necessary testing for the application in order to avoid a default of the application or the product. Weaknesses in customer’s product designs may affect the quality and reliability of the NVIDIA product and may result in additional or different conditions and/or requirements beyond those contained in this document. NVIDIA accepts no liability related to any default, damage, costs, or problem which may be based on or attributable to: (i) the use of the NVIDIA product in any manner that is contrary to this document or (ii) customer product designs.
No license, either expressed or implied, is granted under any NVIDIA patent right, copyright, or other NVIDIA intellectual property right under this document. Information published by NVIDIA regarding third-party products or services does not constitute a license from NVIDIA to use such products or services or a warranty or endorsement thereof. Use of such information may require a license from a third party under the patents or other intellectual property rights of the third party, or a license from NVIDIA under the patents or other intellectual property rights of NVIDIA.
Reproduction of information in this document is permissible only if approved in advance by NVIDIA in writing, reproduced without alteration and in full compliance with all applicable export laws and regulations, and accompanied by all associated conditions, limitations, and notices.
THIS DOCUMENT AND ALL NVIDIA DESIGN SPECIFICATIONS, REFERENCE BOARDS, FILES, DRAWINGS, DIAGNOSTICS, LISTS, AND OTHER DOCUMENTS (TOGETHER AND SEPARATELY, “MATERIALS”) ARE BEING PROVIDED “AS IS.” NVIDIA MAKES NO WARRANTIES, EXPRESSED, IMPLIED, STATUTORY, OR OTHERWISE WITH RESPECT TO THE MATERIALS, AND EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES OF NONINFRINGEMENT, MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE. TO THE EXTENT NOT PROHIBITED BY LAW, IN NO EVENT WILL NVIDIA BE LIABLE FOR ANY DAMAGES, INCLUDING WITHOUT LIMITATION ANY DIRECT, INDIRECT, SPECIAL, INCIDENTAL, PUNITIVE, OR CONSEQUENTIAL DAMAGES, HOWEVER CAUSED AND REGARDLESS OF THE THEORY OF LIABILITY, ARISING OUT OF ANY USE OF THIS DOCUMENT, EVEN IF NVIDIA HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. Notwithstanding any damages that customer might incur for any reason whatsoever, NVIDIA’s aggregate and cumulative liability towards customer for the products described herein shall be limited in accordance with the Terms of Sale for the product.
NVIDIA, the NVIDIA logo, and Mellanox are trademarks and/or registered trademarks of Mellanox Technologies Ltd. and/or NVIDIA Corporation in the U.S. and in other countries. The registered trademark Linux® is used pursuant to a sublicense from the Linux Foundation, the exclusive licensee of Linus Torvalds, owner of the mark on a world¬wide basis. Other company and product names may be trademarks of the respective companies with which they are associated.
© 2023 NVIDIA Corporation & affiliates. All rights reserved.