- Marcar como nuevo
- Favorito
- Suscribir
- Silenciar
- Suscribirse a un feed RSS
- Resaltar
- Imprimir
- Informe de contenido inapropiado
I am trying to convert fortran code using IMSL to code using Intel MKL. I couldn't find the Gaussian cumulative distribution function, not Gaussian random number generators. This is the first attemptto useIntel fortran. Please Help!
Sherry
Sherry
Enlace copiado
3 Respuestas
- Marcar como nuevo
- Favorito
- Suscribir
- Silenciar
- Suscribirse a un feed RSS
- Resaltar
- Imprimir
- Informe de contenido inapropiado
Hello Sherry,
Intel Math Kernel Library contains the functionality you need.
To compute value of cumulative normal distribution function you might want to call v?cdfnorm routine. Its interface is described in section "Special Functions" of chapter "VML Mathematical functions" in MKL Reference Manual. Using setmode function you can set computation accuracy mode which is required for your application (see description of the function in "VML Service Functions" section).
Generation of Gaussian random numbers can be done by means of
v?rnggaussian or v?rnggaussianmv functions. The second routine is intended to generate vectors from multivariate Gaussian distribution. Description of the generators is available in Section "Distribution GeneratorsContinuous Distributions", chapter "Statistical Functions" in MKL Reference Manual. Additional details about generation of random numbersare available in VSL Notes.
Please, note that both cumulative normal distribution function and Gaussian generators have vector interface and allow producing array of numbers for price of one call.
Detailed information about those functions,their interface and performanceis in the library documentation package which is available at
http://software.intel.com/en-us/articles/intel-math-kernel-library-documentation.
Thanks,
Andrey
Intel Math Kernel Library contains the functionality you need.
To compute value of cumulative normal distribution function you might want to call v?cdfnorm routine. Its interface is described in section "Special Functions" of chapter "VML Mathematical functions" in MKL Reference Manual. Using setmode function you can set computation accuracy mode which is required for your application (see description of the function in "VML Service Functions" section).
Generation of Gaussian random numbers can be done by means of
v?rnggaussian or v?rnggaussianmv functions. The second routine is intended to generate vectors from multivariate Gaussian distribution. Description of the generators is available in Section "Distribution GeneratorsContinuous Distributions", chapter "Statistical Functions" in MKL Reference Manual. Additional details about generation of random numbersare available in VSL Notes.
Please, note that both cumulative normal distribution function and Gaussian generators have vector interface and allow producing array of numbers for price of one call.
Detailed information about those functions,their interface and performanceis in the library documentation package which is available at
http://software.intel.com/en-us/articles/intel-math-kernel-library-documentation.
Thanks,
Andrey
- Marcar como nuevo
- Favorito
- Suscribir
- Silenciar
- Suscribirse a un feed RSS
- Resaltar
- Imprimir
- Informe de contenido inapropiado
Andrey,
I have version 9.0. It doesn't have v?cdfnorm. I think I can use special function v?erf. But I got the following message: undefined reference to `vderf_'
Thank you for your help.
Sherry
- Marcar como nuevo
- Favorito
- Suscribir
- Silenciar
- Suscribirse a un feed RSS
- Resaltar
- Imprimir
- Informe de contenido inapropiado
Quoting - xiaomingg
Andrey,
I have version 9.0. It doesn't have v?cdfnorm. I think I can use special function v?erf. But I got the following message: undefined reference to `vderf_'
Thank you for your help.
Sherry
Sherry, Andrey mentioned the version MKL 10.2, because of v. 9.0 no longer supported.
Responder
Opciones de temas
- Suscribirse a un feed RSS
- Marcar tema como nuevo
- Marcar tema como leído
- Flotar este Tema para el usuario actual
- Favorito
- Suscribir
- Página de impresión sencilla