September 2012 - mpiwg-rma - mailman3.cels.anl.gov

Re: [Mpi3-rma] Target displacement in dynamic windows
by Jim Dinan 18 Sep '12

18 Sep '12

Here is the full, proposed erratum: --snip-- In Section 11.2.4, page 412, line 26 the following sentence is added: For dynamic windows, the target_disp argument to RMA communication operations is not restricted to non-negative values. In addition, the following advice to users is added on line 29. Users are cautioned that displacement arithmetic can overflow in variables with MPI_Aint type and result in unexpected values on some platforms. This issue will be addressed in a future version of MPI. --snip-- ~Jim. On 9/18/12 10:25 AM, Jim Dinan wrote: > It is page 412, line 26 in the clean build of approved/MPI-3.0. > > ~Jim. > > On 9/17/12 9:23 PM, Rajeev Thakur wrote: >> The errata will need to identify the page and line at which the exact >> text should be inserted or changed. >> >> Rajeev >> >> On Sep 17, 2012, at 4:12 PM, Jim Dinan wrote: >> >>> Hi All, >>> >>> In today's meeting, we arrived at the following approach to >>> addressing the target displacements issue for dynamic windows: >>> >>> The MPI 3.0 errata should include the following two statements: >>> >>> 1. For dynamic windows, the target_disp argument to RMA communication >>> operations is not restricted to non-negative values. >>> >>> 2. Users are cautioned that displacement arithmetic can overflow in >>> variables with MPI_Aint type and result in unexpected values on some >>> platforms. This issue will be addressed in a future version of MPI. >>> >>> In MPI 3.1, we will look at two possible mechanisms for providing >>> safe offset arithmetic: >>> >>> 1. MPI_Get_pointer(MPI_Aint address, void *ptr), which will allow a >>> user to convert between addresses and pointers. >>> >>> 2. MPI_Auint -- an address integer which can represent the full range >>> of MPI_Aint as a positive integer. This could be an unsigned >>> integer, but for Fortran compatibiliy, it may need to be a singed >>> integer that's larger than MPI_Aint. The following conversion >>> functions would be provided: >>> >>> MPI_Get_auint(MPI_Aint addr, MPI_Auint *uaddr) >>> MPI_Get_aint(MPI_Auint uaddr, MPI_Aint *addr) >>> >>> Please send comments. Thanks, >>> >>> ~Jim. >> > _______________________________________________ > mpi3-rma mailing list > mpi3-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma

1 0

Re: [Mpi3-rma] Target displacement in dynamic windows
by Jim Dinan 18 Sep '12

18 Sep '12

It is page 412, line 26 in the clean build of approved/MPI-3.0. ~Jim. On 9/17/12 9:23 PM, Rajeev Thakur wrote: > The errata will need to identify the page and line at which the exact text should be inserted or changed. > > Rajeev > > On Sep 17, 2012, at 4:12 PM, Jim Dinan wrote: > >> Hi All, >> >> In today's meeting, we arrived at the following approach to addressing the target displacements issue for dynamic windows: >> >> The MPI 3.0 errata should include the following two statements: >> >> 1. For dynamic windows, the target_disp argument to RMA communication operations is not restricted to non-negative values. >> >> 2. Users are cautioned that displacement arithmetic can overflow in variables with MPI_Aint type and result in unexpected values on some platforms. This issue will be addressed in a future version of MPI. >> >> In MPI 3.1, we will look at two possible mechanisms for providing safe offset arithmetic: >> >> 1. MPI_Get_pointer(MPI_Aint address, void *ptr), which will allow a user to convert between addresses and pointers. >> >> 2. MPI_Auint -- an address integer which can represent the full range of MPI_Aint as a positive integer. This could be an unsigned integer, but for Fortran compatibiliy, it may need to be a singed integer that's larger than MPI_Aint. The following conversion functions would be provided: >> >> MPI_Get_auint(MPI_Aint addr, MPI_Auint *uaddr) >> MPI_Get_aint(MPI_Auint uaddr, MPI_Aint *addr) >> >> Please send comments. Thanks, >> >> ~Jim. >

1 0

Re: [Mpi3-rma] Target displacement in dynamic windows
by Jim Dinan 18 Sep '12

18 Sep '12

Hi Hubert, We discussed adding an alternative representation for manipulating window offsets, that would be overflow-safe (e.g. unsigned or "big enough"). The user would have to convert to this representation and then back to MPI_Aint in order to pass the displacement to an RMA communication routine. So, the proposed solution will work with the existing interface at the expense of a few extra calls. The fundamental problem is that in MPI-2, the window size if given by an (signed) MPI_Aint argument, which must be positive. Thus, in MPI-2, the window size and offset are both bounded at MPI_AINT_MAXVAL. Dynamic windows are not bounded in this way, so we need some safe way to do offset arithmetic on the full range of addresses. AFAIK, C doesn't define a signed representation and it doesn't define what will happen when you overflow a signed integer. So, I don't any solution where a user can do arithmetic on an MPI_Aint will work. Does Fortran define this behavior? Here's another proposal for safe displacement arithmetic would be to define a new routine: MPI_Address_add(MPI_Aint base, MPI_Aint disp, MPI_Aint new_base) Which returns new_base = base + disp. This would not require the addition of any new types and would make it MPI's job to do the arithmetic in a safe and portable way. This should also be Fortran-friendly. ~Jim. On 9/18/12 5:32 AM, Hubert Ritzdorf wrote: > Hi, > > the MPI_Auint has the problem that this doesn't work for the RMA communication interfaces and > that you have to create new interfaces for this. > > I think another possibility would be to change the base address for dynamic windows to > MPI_WIN_BASE_DYNAMIC which should be a predefined address and which could be > MPI_MAX_AINT + 1, where MPI_MAX_AINT is the maximal positive value of an MPI_Aint.. > > Advantages: > (*) You can address the entire memory range with the MPI_Aint (positive and negative) displacement > No differences between Fortran and C. > No integer overflow > (*) Same value for Fortran and C > Currently, you have to differentiate between dynamic windows created by C and Fortran > with all the possible problems in mixing C and Fortran routines/functions since the base > addresses are different. > (*) Improved scalability > Fixed predefined value which will be (probably) the same in all processes. > In Fortran, the value of MPI_BOTTOM (address of a global variable) can be different for > different processes (example: MPMD applications). > > Address calculations would simple (ptr = ((char *) MPI_WIN_BASE_DYNAMIC) + displacement + 1) > and could be placed into a macro for C. A function could work for Fortran. > What you need is a single additional function MPI_Get_address_dynamic() which returns > the displacement relative to MPI_WIN_BASE_DYNAMIC and not relative to MPI_BOTTOM. > > Hubert > > PS: I know its late. In the errata, we could replace MPI_BOTTOM by MPI_WIN_BASE_DYNAMIC and can > add the function. MPI_Get_address_dynamic() in addition to the non-negative > displacements. The MPI implementers would have the possibility to realize this and we to discuss it further. > If we stay with MPI_BOTTOM as base for dynamic windows, it's fixed. > > ________________________________________ > From: mpi3-rma-bounces(a)lists.mpi-forum.org [mpi3-rma-bounces(a)lists.mpi-forum.org] on behalf of Jim Dinan [dinan(a)mcs.anl.gov] > Sent: Monday, September 17, 2012 11:12 PM > To: MPI 3.0 Remote Memory Access working group > Subject: [Mpi3-rma] Target displacement in dynamic windows > > Hi All, > > In today's meeting, we arrived at the following approach to addressing > the target displacements issue for dynamic windows: > > The MPI 3.0 errata should include the following two statements: > > 1. For dynamic windows, the target_disp argument to RMA communication > operations is not restricted to non-negative values. > > 2. Users are cautioned that displacement arithmetic can overflow in > variables with MPI_Aint type and result in unexpected values on some > platforms. This issue will be addressed in a future version of MPI. > > In MPI 3.1, we will look at two possible mechanisms for providing safe > offset arithmetic: > > 1. MPI_Get_pointer(MPI_Aint address, void *ptr), which will allow a user > to convert between addresses and pointers. > > 2. MPI_Auint -- an address integer which can represent the full range of > MPI_Aint as a positive integer. This could be an unsigned integer, but > for Fortran compatibiliy, it may need to be a singed integer that's > larger than MPI_Aint. The following conversion functions would be provided: > > MPI_Get_auint(MPI_Aint addr, MPI_Auint *uaddr) > MPI_Get_aint(MPI_Auint uaddr, MPI_Aint *addr) > > Please send comments. Thanks, > > ~Jim. > _______________________________________________ > mpi3-rma mailing list > mpi3-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma > > > Click https://www.mailcontrol.com/sr/KWlIxjc1E8DTndxI!oX7Ug!f1TOPbjCaDuFAgQbitRSh… to report this email as spam. > > _______________________________________________ > mpi3-rma mailing list > mpi3-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma >

1 0

Re: [Mpi3-rma] Target displacement in dynamic windows
by William Gropp 18 Sep '12

18 Sep '12

Rich is suggesting fixing it in the 3.0 with the Forum's approval. Bill William Gropp Director, Parallel Computing Institute Deputy Director for Research Institute for Advanced Computing Applications and Technologies Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign On Sep 18, 2012, at 9:31 AM, Jim Dinan wrote: > On 9/18/12 3:02 AM, Torsten Hoefler wrote: >> I think we should add an additional point about >> MPIX_ERR_RMA_WRONG_FLAVOR as well. > > Are we planning to deprecate this name in 3.1 and replace is with MPI_ERR_FLAVOR? If so, an errata would be helpful. > > ~Jim. > _______________________________________________ > mpi3-rma mailing list > mpi3-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma

1 0

Re: [Mpi3-rma] Target displacement in dynamic windows
by Jim Dinan 18 Sep '12

18 Sep '12

On 9/18/12 3:02 AM, Torsten Hoefler wrote: > I think we should add an additional point about > MPIX_ERR_RMA_WRONG_FLAVOR as well. Are we planning to deprecate this name in 3.1 and replace is with MPI_ERR_FLAVOR? If so, an errata would be helpful. ~Jim.

1 0

Re: [Mpi3-rma] Target displacement in dynamic windows
by Hubert Ritzdorf 18 Sep '12

18 Sep '12

Hi, the MPI_Auint has the problem that this doesn't work for the RMA communication interfaces and that you have to create new interfaces for this. I think another possibility would be to change the base address for dynamic windows to MPI_WIN_BASE_DYNAMIC which should be a predefined address and which could be MPI_MAX_AINT + 1, where MPI_MAX_AINT is the maximal positive value of an MPI_Aint.. Advantages: (*) You can address the entire memory range with the MPI_Aint (positive and negative) displacement No differences between Fortran and C. No integer overflow (*) Same value for Fortran and C Currently, you have to differentiate between dynamic windows created by C and Fortran with all the possible problems in mixing C and Fortran routines/functions since the base addresses are different. (*) Improved scalability Fixed predefined value which will be (probably) the same in all processes. In Fortran, the value of MPI_BOTTOM (address of a global variable) can be different for different processes (example: MPMD applications). Address calculations would simple (ptr = ((char *) MPI_WIN_BASE_DYNAMIC) + displacement + 1) and could be placed into a macro for C. A function could work for Fortran. What you need is a single additional function MPI_Get_address_dynamic() which returns the displacement relative to MPI_WIN_BASE_DYNAMIC and not relative to MPI_BOTTOM. Hubert PS: I know its late. In the errata, we could replace MPI_BOTTOM by MPI_WIN_BASE_DYNAMIC and can add the function. MPI_Get_address_dynamic() in addition to the non-negative displacements. The MPI implementers would have the possibility to realize this and we to discuss it further. If we stay with MPI_BOTTOM as base for dynamic windows, it's fixed. ________________________________________ From: mpi3-rma-bounces(a)lists.mpi-forum.org [mpi3-rma-bounces(a)lists.mpi-forum.org] on behalf of Jim Dinan [dinan(a)mcs.anl.gov] Sent: Monday, September 17, 2012 11:12 PM To: MPI 3.0 Remote Memory Access working group Subject: [Mpi3-rma] Target displacement in dynamic windows Hi All, In today's meeting, we arrived at the following approach to addressing the target displacements issue for dynamic windows: The MPI 3.0 errata should include the following two statements: 1. For dynamic windows, the target_disp argument to RMA communication operations is not restricted to non-negative values. 2. Users are cautioned that displacement arithmetic can overflow in variables with MPI_Aint type and result in unexpected values on some platforms. This issue will be addressed in a future version of MPI. In MPI 3.1, we will look at two possible mechanisms for providing safe offset arithmetic: 1. MPI_Get_pointer(MPI_Aint address, void *ptr), which will allow a user to convert between addresses and pointers. 2. MPI_Auint -- an address integer which can represent the full range of MPI_Aint as a positive integer. This could be an unsigned integer, but for Fortran compatibiliy, it may need to be a singed integer that's larger than MPI_Aint. The following conversion functions would be provided: MPI_Get_auint(MPI_Aint addr, MPI_Auint *uaddr) MPI_Get_aint(MPI_Auint uaddr, MPI_Aint *addr) Please send comments. Thanks, ~Jim. _______________________________________________ mpi3-rma mailing list mpi3-rma(a)lists.mpi-forum.org http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma Click https://www.mailcontrol.com/sr/KWlIxjc1E8DTndxI!oX7Ug!f1TOPbjCaDuFAgQbitRSh… to report this email as spam.

1 0

Re: [Mpi3-rma] Target displacement in dynamic windows
by Torsten Hoefler 18 Sep '12

18 Sep '12

Hi Jim, > In today's meeting, we arrived at the following approach to addressing > the target displacements issue for dynamic windows: > > The MPI 3.0 errata should include the following two statements: > > 1. For dynamic windows, the target_disp argument to RMA communication > operations is not restricted to non-negative values. > > 2. Users are cautioned that displacement arithmetic can overflow in > variables with MPI_Aint type and result in unexpected values on some > platforms. This issue will be addressed in a future version of MPI. Sounds good! I think we should add an additional point about MPIX_ERR_RMA_WRONG_FLAVOR as well. > In MPI 3.1, we will look at two possible mechanisms for providing safe > offset arithmetic: > > 1. MPI_Get_pointer(MPI_Aint address, void *ptr), which will allow a user > to convert between addresses and pointers. > > 2. MPI_Auint -- an address integer which can represent the full range of > MPI_Aint as a positive integer. This could be an unsigned integer, but > for Fortran compatibiliy, it may need to be a singed integer that's > larger than MPI_Aint. The following conversion functions would be > provided: > > MPI_Get_auint(MPI_Aint addr, MPI_Auint *uaddr) > MPI_Get_aint(MPI_Auint uaddr, MPI_Aint *addr) This certainly needs more discussion! All the Best, Torsten -- ### qreharg rug ebs fv crryF ------------- http://www.unixer.de/ ----- Torsten Hoefler | Assistant Professor Dept. of Computer Science | ETH Zï¿½rich Universitï¿½tsstrasse 6 | Zurich-8092, Switzerland CAB E 64.1 | Phone: +41 76 309 79 29

1 0

Re: [Mpi3-rma] Target displacement in dynamic windows
by Rajeev Thakur 17 Sep '12

17 Sep '12

The errata will need to identify the page and line at which the exact text should be inserted or changed. Rajeev On Sep 17, 2012, at 4:12 PM, Jim Dinan wrote: > Hi All, > > In today's meeting, we arrived at the following approach to addressing the target displacements issue for dynamic windows: > > The MPI 3.0 errata should include the following two statements: > > 1. For dynamic windows, the target_disp argument to RMA communication operations is not restricted to non-negative values. > > 2. Users are cautioned that displacement arithmetic can overflow in variables with MPI_Aint type and result in unexpected values on some platforms. This issue will be addressed in a future version of MPI. > > In MPI 3.1, we will look at two possible mechanisms for providing safe offset arithmetic: > > 1. MPI_Get_pointer(MPI_Aint address, void *ptr), which will allow a user to convert between addresses and pointers. > > 2. MPI_Auint -- an address integer which can represent the full range of MPI_Aint as a positive integer. This could be an unsigned integer, but for Fortran compatibiliy, it may need to be a singed integer that's larger than MPI_Aint. The following conversion functions would be provided: > > MPI_Get_auint(MPI_Aint addr, MPI_Auint *uaddr) > MPI_Get_aint(MPI_Auint uaddr, MPI_Aint *addr) > > Please send comments. Thanks, > > ~Jim.

1 0

Re: [Mpi3-rma] Target displacement in dynamic windows
by Jeff Hammond 17 Sep '12

17 Sep '12

do the Fortran 2008 bindings support a hook for C unsigned ints? since FORTRAN77 bindings are still optional, maybe we can just accept that they don't work as well as the C bindings, which isn't too different than the existing situation where all sorts of things aren't certain to work. jeff On Mon, Sep 17, 2012 at 9:18 PM, William Gropp <wgropp(a)illinois.edu> wrote: > Thanks - I think the 3.0 errata is a reasonable solution. I'll note that the 3.1 solutions may not be practical in Fortran (e.g., a larger integer may not be available). > > Bill > > William Gropp > Director, Parallel Computing Institute > Deputy Director for Research > Institute for Advanced Computing Applications and Technologies > Paul and Cynthia Saylor Professor of Computer Science > University of Illinois Urbana-Champaign > > > > On Sep 17, 2012, at 4:12 PM, Jim Dinan wrote: > >> Hi All, >> >> In today's meeting, we arrived at the following approach to addressing the target displacements issue for dynamic windows: >> >> The MPI 3.0 errata should include the following two statements: >> >> 1. For dynamic windows, the target_disp argument to RMA communication operations is not restricted to non-negative values. >> >> 2. Users are cautioned that displacement arithmetic can overflow in variables with MPI_Aint type and result in unexpected values on some platforms. This issue will be addressed in a future version of MPI. >> >> In MPI 3.1, we will look at two possible mechanisms for providing safe offset arithmetic: >> >> 1. MPI_Get_pointer(MPI_Aint address, void *ptr), which will allow a user to convert between addresses and pointers. >> >> 2. MPI_Auint -- an address integer which can represent the full range of MPI_Aint as a positive integer. This could be an unsigned integer, but for Fortran compatibiliy, it may need to be a singed integer that's larger than MPI_Aint. The following conversion functions would be provided: >> >> MPI_Get_auint(MPI_Aint addr, MPI_Auint *uaddr) >> MPI_Get_aint(MPI_Auint uaddr, MPI_Aint *addr) >> >> Please send comments. Thanks, >> >> ~Jim. > > > _______________________________________________ > mpi3-rma mailing list > mpi3-rma(a)lists.mpi-forum.org > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma -- Jeff Hammond Argonne Leadership Computing Facility University of Chicago Computation Institute jhammond(a)alcf.anl.gov / (630) 252-5381 http://www.linkedin.com/in/jeffhammond https://wiki.alcf.anl.gov/parts/index.php/User:Jhammond

1 0

Re: [Mpi3-rma] Target displacement in dynamic windows
by William Gropp 17 Sep '12

17 Sep '12

Thanks - I think the 3.0 errata is a reasonable solution. I'll note that the 3.1 solutions may not be practical in Fortran (e.g., a larger integer may not be available). Bill William Gropp Director, Parallel Computing Institute Deputy Director for Research Institute for Advanced Computing Applications and Technologies Paul and Cynthia Saylor Professor of Computer Science University of Illinois Urbana-Champaign On Sep 17, 2012, at 4:12 PM, Jim Dinan wrote: > Hi All, > > In today's meeting, we arrived at the following approach to addressing the target displacements issue for dynamic windows: > > The MPI 3.0 errata should include the following two statements: > > 1. For dynamic windows, the target_disp argument to RMA communication operations is not restricted to non-negative values. > > 2. Users are cautioned that displacement arithmetic can overflow in variables with MPI_Aint type and result in unexpected values on some platforms. This issue will be addressed in a future version of MPI. > > In MPI 3.1, we will look at two possible mechanisms for providing safe offset arithmetic: > > 1. MPI_Get_pointer(MPI_Aint address, void *ptr), which will allow a user to convert between addresses and pointers. > > 2. MPI_Auint -- an address integer which can represent the full range of MPI_Aint as a positive integer. This could be an unsigned integer, but for Fortran compatibiliy, it may need to be a singed integer that's larger than MPI_Aint. The following conversion functions would be provided: > > MPI_Get_auint(MPI_Aint addr, MPI_Auint *uaddr) > MPI_Get_aint(MPI_Auint uaddr, MPI_Aint *addr) > > Please send comments. Thanks, > > ~Jim.

1 0