KMP_AFFINITY
The KMP_AFFINITY environment variable uses the following general syntax:
Syntax |
---|
KMP_AFFINITY=[<modifier>,...]<type>[,<permute>][,<offset>] |
For example, to list a machine topology map, specify KMP_AFFINITY=verbose,none to use a modifier of verbose and a type of none.
The following table describes the supported specific arguments.
Argument |
Default |
Description |
---|---|---|
noverbose respect granularity=core |
Optional. String consisting of keyword and specifier.
|
|
none |
Required string. Indicates the thread affinity to use.
The logical and physical types are deprecated but supported for backward compatibility. |
|
0 |
Optional. Positive integer value. Not valid with type values of explicit, none, or disabled. | |
0 |
Optional. Positive integer value. Not valid with type values of explicit, none, or disabled. |
Type is the only required argument.
Does not bind OpenMP threads to particular thread contexts; however, if the operating system supports affinity, the compiler still uses the OpenMP thread affinity interface to determine machine topology. Specify KMP_AFFINITY=verbose,none to list a machine topology map.
Specifying compact assigns the OpenMP thread <n>+1 to a free thread context as close as possible to the thread context where the <n> OpenMP thread was placed. For example, in a topology map, the nearer a node is to the root, the more significance the node has when sorting the threads.
Specifying disabled completely disables the thread affinity interfaces. This forces the OpenMP run-time library to behave as if the affinity interface was not supported by the operating system. This includes the low-level API interfaces such as kmp_set_affinity and kmp_get_affinity, which have no effect and will return a nonzero error code.
Specifying explicit assigns OpenMP threads to a list of OS proc IDs that have been explicitly specified by using the proclist= modifier, which is required for this affinity type.
Specifying scatter distributes the threads as evenly as possible across the entire system. scatter is the opposite of compact; so the leaves of the node are most significant when sorting through the machine topology map.
Types logical and physical are deprecated and may become unsupported in a future release. Both are supported for backward compatibility.
For logical and physical affinity types, a single trailing integer is interpreted as an offset specifier instead of a permute specifier. In contrast, with compact and scatter types, a single trailing integer is interpreted as a permute specifier.
Specifying logical assigns OpenMP threads to consecutive logical processors, which are also called hardware thread contexts. The type is equivalent to compact, except that the permute specifier is not allowed. Thus, KMP_AFFINITY=logical,n is equivalent to KMP_AFFINITY=compact,0,n (this equivalence is true regardless of the whether or not a granularity=fine modifier is present).
For both compact and scatter, permute and offset are allowed; however, if you specify only one integer, the compiler interprets the value as a permute specifier. Both permute and offset default to 0.
The permute specifier controls which levels are most significant when sorting the machine topology map. A value for permute forces the mappings to make the specified number of most significant levels of the sort the least significant, and it inverts the order of significance. The root node of the tree is not considered a separate level for the sort operations.
The offset specifier indicates the starting position for thread assignment.
Modifiers are optional arguments that precede type. If you do not specify a modifier, the noverbose, respect, and granularity=core modifiers are used automatically.
Modifiers are interpreted in order from left to right, and can negate each other. For example, specifying KMP_AFFINITY=verbose,noverbose,scatter is therefore equivalent to setting KMP_AFFINITY=noverbose,scatter, or just KMP_AFFINITY=scatter.
Does not print verbose messages.
Prints messages concerning the supported affinity. The messages include information about the number of packages, number of cores in each package, number of thread contexts for each core, and OpenMP thread bindings to physical thread contexts.
Information about binding OpenMP threads to physical thread contexts is indirectly shown in the form of the mappings between hardware thread contexts and the operating system (OS) processor (proc) IDs. The affinity mask for each OpenMP thread is printed as a set of OS processor IDs.
KMP_LIBRARY
KMP_LIBRARY = { throughput | turnaround | serial }, Selects the OpenMP run-time library execution mode. The options for the variable value are throughput, turnaround, and serial.
The compiler with OpenMP enables you to run an application under different execution modes that can be specified at run time. The libraries support the serial, turnaround, and throughput modes.
The serial mode forces parallel applications to run on a single processor.
In a dedicated (batch or single user) parallel environment where all processors are exclusively allocated to the program for its entire run, it is most important to effectively utilize all of the processors all of the time. The turnaround mode is designed to keep active all of the processors involved in the parallel computation in order to minimize the execution time of a single job. In this mode, the worker threads actively wait for more parallel work, without yielding to other threads.
Avoid over-allocating system resources. This occurs if either too many threads have been specified, or if too few processors are available at run time. If system resources are over-allocated, this mode will cause poor performance. The throughput mode should be used instead if this occurs.
In a multi-user environment where the load on the parallel machine is not constant or where the job stream is not predictable, it may be better to design and tune for throughput. This minimizes the total time to run multiple jobs simultaneously. In this mode, the worker threads will yield to other threads while waiting for more parallel work.
The throughput mode is designed to make the program aware of its environment (that is, the system load) and to adjust its resource usage to produce efficient execution in a dynamic environment. This mode is the default.
KMP_BLOCKTIME
KMP_BLOCKTIME = value. Sets the time, in milliseconds, that a thread should wait, after completing the execution of a parallel region, before sleeping.Use the optional character suffixes: s (seconds), m (minutes), h (hours), or d (days) to specify the units.Specify infinite for an unlimited wait time.
KMP_STACKSIZE
KMP_STACKSIZE = value. Sets the number of bytes to allocate for each OpenMP* thread to use as the private stack for the thread. Recommended size is 16m. Use the optional suffixes: b (bytes), k (kilobytes), m (megabytes), g (gigabytes), or t (terabytes) to specify the units. This variable does not affect the native operating system threads created by the user program nor the thread executing the sequential part of an OpenMP* program or parallel programs created using -parallel.
OMP_NUM_THREADS
Sets the maximum number of threads to use for OpenMP* parallel regions if no other value is specified in the application. This environment variable applies to both -openmp and -parallel. Example syntax on a Linux system with 8 cores: export OMP_NUM_THREADS=8
OMP_DYNAMIC
OMP_DYNAMIC={ 1 | 0 } Enables (1, true) or disables (0,false) the dynamic adjustment of the number of threads.
OMP_SCHEDULE
OMP_SCHEDULE={ type,[chunk size]} Controls the scheduling of the for-loop work-sharing construct. type can be either of static,dynamic,guided,runtime chunk size should be positive integer
OMP_NESTED
OMP_NESTED={ 1 | 0 } Enables creation of new teams in case of nested parallel regions (1,true) or serializes (0,false) all nested parallel regions. Default is 0.
for purposes of optimization, it is assumed that pointer-based variables do not overlay the storage of any other variable.
]]>Default: The compiler generates code for the small memory model.
Usage
The following command line requests position independent code be generated, and the option -mcmodel=medium be passed to the assembler and linker:
$ pgfortran -mcmodel=medium myprog.f
Description
The default small memory model of the linux86-64 environment limits the combined area for a user's object or executable to 1GB,
with the Linux kernel managing usage of the second 1GB of address for system routines, shared libraries, stacks, and so on.
Programs are started at a fixed address, and the program can use a single instruction to make most memory references.
The medium memory model allows for larger than 2GB data areas, or .bss sections. Program units compiled using either -mcmodel=medium
or -fpic require additional instructions to reference memory.
The effect on performance is a function of the data-use of the application. The -mcmodel=medium switch must be used at both compile
time and link time to create 64-bit executables. Program units compiled for the default small memory model can be linked into medium
memory model executables as long as they are compiled with the option -fpic, or position-independent.
The linux86-64 environment provides static libxxx.a archive libraries, that are built both with and without -fpic, and dynamic
libxxx.so shared object libraries that are compiled with -fpic. Using the link switch -mcmodel=medium implies the -fpic switch
and utilizes the shared libraries by default. The directory $PGI/linux86-64/
memory model codes; and the directory $PGI/linux86-64/
-mcmodel=medium?executables.
Note:
-mcmodel=medium -fpic is not allowed to create shared libraries. However, you can create static archive libraries (.a) that are -fpic.
]]>Enables vectorization with SIMD instructions, cache alignment, and flushz for 64-bit targets.
Description
When you use this option, a generally optimal set of options is chosen for targets that support SIMD capability. In addition, the appropriate -tp option is automatically included to enable generation of code optimized for the type of system on which compilation is performed. This option enables vectorization with SIMD instructions, cache alignment, and flushz.
Note: Auto-selection of the appropriate -tp option means that programs built using the -fastsse option on a given system are not necessarily backward-compatible with older systems.
Note: C/C++ compilers enable -Mautoinline with-fast.
]]>