This topic lists the options that provide new functionality in this release.
Some compiler options are only available on certain systems, as indicated by these labels:
Label |
Meaning |
|---|---|
i32 |
The option is available on systems using IA-32 architecture. |
i64em |
The option is available on systems using Intel® 64 architecture. |
If no label appears, the option is available on all supported systems.
If "only" appears in the label, the option is only available on the identified system.
For more details on the options, refer to the Alphabetical Compiler Options section.
For information on conventions used in this table, see Conventions.
New compiler options are listed in tables below:
The first table lists new options that are available on Windows* systems.
The second table lists new options that are available on Linux* and Mac OS* X systems. If an option is only available on one of these operating systems, it is labeled.
Windows* OS Options |
Description |
Default |
|---|---|---|
/Qansi-alias-check |
Enables or disables the ansi-alias checker. |
/Qansi-alias-check- |
/Qcilk-serialize |
Causes serialization of Intel® Cilk™ Plus code. |
OFF |
/Qdiag-sc-dir[:dir] |
Specifies a directory for the results of static security analysis. |
OFF |
/Qfp-trap:mode |
Sets the floating-point trapping mode for the main routine. |
/Qfp-trap:none |
/Qfp-trap-all:mode |
Sets the floating-point trapping mode for all routines. |
/Qfp-trap-all:none |
/Qguide[=n] |
Lets you set a level of guidance for auto-vectorization, auto-parallelization, and data transformation. |
OFF |
/Qguide-data-trans[=n] |
Lets you set the level of guidance for data transformation. |
OFF |
/Qguide-file[=filename] |
Causes the results of guided auto-parallelization to be output to a file. |
OFF |
/Qguide-file-append[=filename] |
Causes the results of guided auto-parallelization to be appended to a file. |
OFF |
/Qguide-opts=string |
Tells the compiler to analyze certain code and generate recommendations that may improve optimizations. |
OFF |
/Qguide-par[=n] |
Lets you set the level of guidance for auto-parallelization. |
OFF |
/Qguide-vec[=n] |
Lets you set the level of guidance for auto-vectorization. |
OFF |
/Qimf-absolute-error |
Defines the maximum allowable absolute error for math library function results. |
OFF |
/Qimf-accuracy-bits |
Defines the relative error for math library function results. |
OFF |
/Qimf-arch-consistency |
Ensures that the math library functions produce consistent results across different implementations of the same architecture. |
OFF |
/Qimf-max-error |
Defines the maximum allowable relative error for math library function results. |
OFF |
/Qimf-precision |
Defines the accuracy for math library functions. |
OFF |
/Qintel-extensions |
Enables or disables all Intel® C/C++ language extensions. |
/Qintel-extensions |
/Qopt-args-in-regs:keyword |
Determines whether calls to routines are optimized by passing parameters in registers instead of on the stack. |
/Qopt-args-in-regs:seen |
/Qopt-matmul |
Enables or disables a compiler-generated Matrix Multiply (matmul) library call. |
/Qopt-matmul- |
/Qparallel-source-info[-] |
Enables or disables source location emission when OpenMP* or auto-parallelization code is generated. |
/Qparallel-source-info |
|
/Qpatchable-addresses
|
Tells the compiler to generate code such that references to statically assigned addresses can be patched. |
OFF |
/Qprof-value-profiling[:keyword] |
Controls which values are value profiled. |
/Qprof-value-profiling:all |
/Qprofile-functions |
Inserts instrumentation calls at a function's entry and exit points. |
OFF |
/Qprofile-loops:keyword |
Inserts instrumentation calls at a function's entry and exit points, and before and after instrumentable loops. |
OFF |
/Qprofile-loops-report[:n] |
Controls the level of detail for the data collected when instrumentation occurs before and after certain loops. |
/Qprofile-loops-report:1 |
/Qregcall |
Tells the compiler that the __regcall calling convention should be used for functions that do not directly specify a calling convention. |
OFF |
/Qsimd |
Enables or disables the SIMD vectorization feature of the compiler. |
/Qsimd |
/Qzero-initialized-in-bss |
Determines whether the compiler places in the DATA section any variables explicitly initialized with zeros. |
/Qzero-initialized-in-bss |
Linux* OS and Mac OS* X Options |
Description |
Default |
|---|---|---|
-ansi-alias-check |
Enables or disables the ansi-alias checker. |
-no-ansi-alias-check |
-auto-p32 (i64em only; Linux* OS only) |
Instructs the compiler to analyze the program to determine if there are 64-bit pointers that can be safely shrunk into 32-bit pointers. |
OFF |
-cilk-serialize (Linux* OS only) |
Causes serialization of Intel® Cilk™ Plus code. |
OFF |
-diag-sc-dir[=dir] |
Specifies a directory for the results of static security analysis. |
OFF |
-ffriend-injection |
Causes the compiler to inject friend functions into the enclosing namespace. |
OFF |
-fzero-initialized-in-bss |
Determines whether the compiler places in the DATA section any variables explicitly initialized with zeros. |
-fzero-initialized-in-bss |
-fimf-absolute-error |
Defines the maximum allowable absolute error for math library function results. |
OFF |
-fimf-accuracy-bits |
Defines the relative error for math library function results. |
OFF |
-fimf-arch-consistency |
Ensures that the math library functions produce consistent results across different implementations of the same architecture. |
OFF |
-fimf-max-error |
Defines the maximum allowable relative error for math library function results. |
OFF |
-fimf-precision |
Defines the accuracy for math library functions. |
OFF |
-fms-dialect[=ver](Mac OS* X only) |
Enables support for a language dialect that is compatible with Microsoft* Windows, while maintaining link compatibility with gcc. |
-fno-ms-dialect |
-fp-trap=mode |
Sets the floating-point trapping mode for the main routine. |
-fp-trap=none |
-fp-trap-all=mode |
Sets the floating-point trapping mode for all routines. |
-fp-trap-all=none |
-fvar-tracking |
This is an alternate name for option -debug variable-locations. |
OFF |
-fvar-tracking-assignments |
This is an alternate name for option -debug semantic-stepping. |
OFF |
-guide[=n] |
Lets you set a level of guidance for auto-vectorization, auto-parallelization, and data transformation. |
OFF |
-guide-data-trans[=n] |
Lets you set the level of guidance for data transformation. |
OFF |
-guide-file[=filename] |
Causes the results of guided auto-parallelization to be output to a file. |
OFF |
-guide-file-append[=filename] |
Causes the results of guided auto-parallelization to be appended to a file. |
OFF |
-guide-opts=string |
Tells the compiler to analyze certain code and generate recommendations that may improve optimizations. |
OFF |
-guide-par[=n] |
Lets you set the level of guidance for auto-parallelization. |
OFF |
-guide-vec[=n] |
Lets you set the level of guidance for auto-vectorization. |
OFF |
-intel-extensions (Linux* OS only) |
Enables or disables all Intel® C/C++ language extensions. |
-intel-extensions( |
-no-libgcc |
Prevents the linking of certain gcc-specific libraries. |
OFF |
-opt-args-in-regs=keyword |
Determines whether calls to routines are optimized by passing parameters in registers instead of on the stack. |
-opt-args-in-regs=seen |
-opt-matmul |
Enables or disables a compiler-generated Matrix Multiply (matmul) library call. |
-no-opt-matmul |
-[no-]parallel-source-info |
Enables or disables source location emission when OpenMP* or auto-parallelization code is generated. |
-parallel-source-info |
-prof-value-profiling[=keyword] |
Controls which values are value profiled. |
-prof-value-profiling=all |
-profile-functions |
Inserts instrumentation calls at a function's entry and exit points. |
OFF |
-profile-loops=keyword |
Inserts instrumentation calls at a function's entry and exit points, and before and after instrumentable loops. |
OFF |
-profile-loops-report[=n] |
Controls the level of detail for the data collected when instrumentation occurs before and after certain loops. |
-profile-loops-report=1 |
-regcall |
Tells the compiler that the __regcall calling convention should be used for functions that do not directly specify a calling convention. |
OFF |
-simd |
Enables or disables the SIMD vectorization feature of the compiler. |
-simd |
-Wremarks |
Tells the compiler to display diagnostic messages that are remarks or comments. |
OFF |
-Wsign-compare |
Determines whether warnings are issued when a comparison between signed and unsigned values could produce an incorrect result when the signed value is converted to unsigned. |
-Wno-sign-compare |
-Wstrict-aliasing |
Determines whether warnings are issued for code that might violate the optimizer's strict aliasing rules. |
-Wno-strict-aliasing |
Copyright © 1996-2010, Intel Corporation. All rights reserved.