https://www.jiscmail.ac.uk/cgi-bin/webadmin?RSS&L=COMP-FORTRAN-90&v=1.0COMP-FORTRAN-90 List
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A0=COMP-FORTRAN-90
COMP-FORTRAN-90 List Archives2016-06-21T15:04:56ZPowered by L-Soft's LISTSERV mailing list manager
http://www.lsoft.com/products/listserv-powered.asp
http://www.lsoft.com/images/listserv_small.gifRe: coarray objects in unformatted stream IO
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;cd9880e4.1606
On Jun 21, 2016, at 9:05 AM, Lionel, Steve <steve.lionel@INTEL.COM> wrote:<br><br>> -----Original Message-----<br>> From: Fortran 90 List [mailto:COMP-FORTRAN-90@JISCMAIL.AC.UK] On Behalf Of Anton Shterenlikht<br>> Sent: Tuesday, June 21, 2016 4:58 AM<br>> To: COMP-FORTRAN-90@JISCMAIL.AC.UK<br>> Subject: coarray objects in unformatted stream IO<br>><br>>> This example appeared in comp.lang.fortran.<br>>> Three compilers show 3 different behaviours.<br>> [snip]<br>><br>> I think the program is fine. [...]
2016-06-21T15:04:48+00:00Bill Longhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;cd9880e4.1606Re: coarray objects in unformatted stream IO
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;4409b28f.1606
>From owner-comp-fortran-90@jiscmail.ac.uk Tue Jun 21 15:48:42 2016<br>><br>>> This example appeared in comp.lang.fortran.<br>>>Three compilers show 3 different behaviours.<br>>[snip]<br>><br>>I think the program is fine. Intel Fortran produces the "bad results" and I will send that on to the developers. I don't think the I/O type (unformatted stream) is relevant - my experiments show it doesn't matter. For a while I wondered if there needed to be at least a SYNC MEMORY executed on all images between the write loop and the read loop, but even if I add that it doesn't change the behavior and there [...]
2016-06-21T15:52:42+01:00Anton Shterenlikhthttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;4409b28f.1606Re: coarray objects in unformatted stream IO
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;de149b5f.1606
-----Original Message-----<br>From: Fortran 90 List [mailto:COMP-FORTRAN-90@JISCMAIL.AC.UK] On Behalf Of Anton Shterenlikht<br>Sent: Tuesday, June 21, 2016 4:58 AM<br>To: COMP-FORTRAN-90@JISCMAIL.AC.UK<br>Subject: coarray objects in unformatted stream IO<br><br>> This example appeared in comp.lang.fortran.<br>>Three compilers show 3 different behaviours.<br>[snip]<br><br>I think the program is fine. Intel Fortran produces the "bad results" and I will send that on to the developers. I don't think the I/O type (unformatted stream) is relevant - my experiments show it doesn't matter. For a while I wondered if there needed to be at least a SYNC MEMORY executed on all images between the [...]
2016-06-21T14:05:37+00:00Lionel, Stevehttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;de149b5f.1606coarray objects in unformatted stream IO
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;d3cec496.1606
This example appeared in comp.lang.fortran.<br>Three compilers show 3 different behaviours.<br><br>program z<br>implicit none<br>real, allocatable :: y(:)[:], y2(:)[:], x(:)<br>integer nim, i, funit<br><br>nim = num_images()<br>allocate( y(3)[*], y2(3)[*] )<br>allocate( x(3) )<br>y = this_image()<br>y2 = 0.0<br><br>sync all<br><br>if( this_image() .eq. 1 ) then<br>print *, NIM<br>open( newunit = funit, file="ustream.bin", access="stream", &<br>form="unformatted", status="replace" )<br>do i = 1, nim<br>write( funit ) y(:)[i]<br>end do<br>close( funit ) [...]
2016-06-21T09:57:47+01:00Anton Shterenlikhthttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;d3cec496.1606nag compiler news
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;fc895fe5.1605
As some of you will be aware<br>Nag have released version 6.1<br>Of their compiler.<br><br>What you may not be aware of is that the Mac<br>Version now has the Fortran Builder front end.<br><br>Visit<br><br>http://www.nag.com/nagware/NP/fortranbuilder.asp<br><br>these are some of the features<br><br>On 64-bit Windows systems, produces either 32- or 64-bit code<br>GUI debugger<br>Integrated help system<br>"Pretty Printer" - also converts fixed to free format<br>Call graph generator<br>Interface block builder<br>Quad precision real<br>Informative runtime error messages<br>Precision unifier [...]
2016-05-24T09:29:06+01:00Ian Chivershttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;fc895fe5.1605New resource on object oriented programming
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;61fc1025.1605
I was contacted this week about some dead links<br>On object oriented in Fortran in some of our Fortran Resources,<br>with details of a link to a new site dedicated to oop.<br><br>Here is the site address.<br><br>Guide to Object-Oriented Programming<br><br>http://wiht.link/oo-programming<br><br>Thanks to Sophia Kahn<br>for the link.<br><br>I've added a section in our Fortran Resource file<br>On OOP in Fortran with a link to this site. [...]
2016-05-14T16:50:49+01:00Ian Chivershttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;61fc1025.1605Re: reading from input_unit intrinsic assignment?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;3c340d1d.1605
>> Or shall I submit a paper for the<br>>> J3 meeting suggesing a new edit?<br>><br>>A J3 paper would be needed to generate a fix.<br>>I’ve added it to my list of issues for the next meeting.<br><br>I'm happy with that.<br><br>Thank you<br><br>Anton
2016-05-03T15:08:35+01:00Anton Shterenlikhthttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;3c340d1d.1605Re: reading from input_unit intrinsic assignment?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;2fb904b4.1605
On May 3, 2016, at 3:47 AM, Anton Shterenlikht <mexas@BRIS.AC.UK> wrote:<br><br>>> In the section on character input formatting for list-directed input:<br>>><br>>> "Let len be the length of the next eective item, and let w be the length of the character sequence. If len is less than or equal to w, the leftmost len characters of the sequence are transmitted to the next eective item. If len is greater than w, the sequence is transmitted to the leftmost w characters of the next eective item and the remaining len≠w characters of the next eective item are filled with [...]
2016-05-03T12:18:11+00:00Bill Longhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;2fb904b4.1605Re: reading from input_unit intrinsic assignment?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;cffa34b8.1605
>In the section on character input formatting for list-directed input:<br>><br>>"Let len be the length of the next eective item, and let w be the length of the character sequence. If len is less than or equal to w, the leftmost len characters of the sequence are transmitted to the next eective item. If len is greater than w, the sequence is transmitted to the leftmost w characters of the next eective item and the remaining len≠w characters of the next eective item are filled with blanks. The effect is as though the sequence were assigned to the next [...]
2016-05-03T09:47:47+01:00Anton Shterenlikhthttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;cffa34b8.1605Fwd: (CALL FOR PAPERS) PAW16: PGAS Applications Workshop
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;dacf5a48.1604
Hi,<br><br>I’m forwarding this for Damian Rouson and Karla Morris.<br><br>Begin forwarded message:<br><br>On Apr 26, 2016, at 9:33 AM, Morris, Karla NMN <knmorri@sandia.gov <mailto:knmorri@sandia.gov>> wrote:<br><br>CALL FOR PAPERS<br><br>PAW16: PGAS Applications Workshop<br>http://sourceryinstitute.github.io/PAW/ <http://sourceryinstitute.github.io/PAW/><br><br>November 14, 2016 - Salt Lake City, Utah<br><br>Held in conjunction with SC16<br>http://sc16.supercomputing.org/ <http://sc16.supercomputing.org/><br>In cooperation with SIGHPC<br><br>SUMMARY<br><br>The race towards Exascale computing is on, and a lot of stress<br>is put on researchers to break the boundaries of productivity<br>and efficiency imposed by traditional programming models.<br>Partitioned Global Address Space (PGAS) languages are an<br>effective alternative, and the most promising path towards<br> [...]
2016-04-27T14:57:28-06:00Dan Naglehttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;dacf5a48.1604unsubscribe
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;a05a154.1604
Sabino Chavez-Cerda<br>INAOE
2016-04-26T18:48:40-05:00<>https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;a05a154.1604Re: How to use BOZ-literal-constants in an integer array constructor?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;3693d85c.1604
On Mon, 2016-04-25 at 15:20 +0000, Bill Long wrote:<br>> Correct. There was a proposal for Fortran 2008 to include a “BITS”<br>> data type that would have enabled something like this. However, in an<br>> attempt to reduce the size of the changes to the standard, it was<br>> dropped. While there were no actual technical problems with the<br>> proposal, it was seen as too much effort for the benefit received.<br>> Weighing against it was the difficulty of implementing any change to<br>> the typing system. [...]
2016-04-25T13:20:51-07:00Van Snyderhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;3693d85c.1604Re: How to use BOZ-literal-constants in an integer array constructor?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;acbc8776.1604
Indeed, if I invoke standards checking, I see that as well. This does seem to be an unnecessary irregularity in the standard - allowing a scalar of something as an argument to an elemental function while prohibiting an array of the same things.<br><br>Cheers,<br>Bill<br><br>On Apr 25, 2016, at 1:11 PM, Vipul Parekh <parekhvs@GMAIL.COM> wrote: [...]
2016-04-25T18:31:18+00:00Bill Longhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;acbc8776.1604unsubscribe
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;3bb2909f.1604
J. Clark Powers<br>MS Technical Communication<br>NC State University<br>LinkedIn: J. Clark Powers <http://www.linkedin.com/in/jclarkpowers><br>@JClarkPowers <http://twitter.com/JClarkPowers>
2016-04-25T14:14:21-04:00Jason Powershttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;3bb2909f.1604Re: How to use BOZ-literal-constants in an integer array constructor?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;82da01d.1604
Bill,<br><br>Thanks much for your response.<br><br>You advise, "I do not see any reason for this to fail. INT is an elemental<br>function. INT ( [a,b,c] ) has an array result with the value [ INT(a),<br>INT(b), INT(c) ]. If you have a compiler that is rejecting this, you might<br>contact the vendor." in the context of my statement, "integer(kind=xx),<br>parameter :: bar(*) = int( [ z'0' ], kind=kind(bar) ) !.. Not OK". [...]
2016-04-25T14:11:01-04:00Vipul Parekhhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;82da01d.1604Re: How to use BOZ-literal-constants in an integer array constructor?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;bae8b856.1604
On Apr 22, 2016, at 7:41 PM, Vipul Parekh <parekhvs@GMAIL.COM> wrote:<br><br>> Hi,<br>><br>> I need to construct an integer array of a named constant using BOZ-literal-constants as in<br>><br>> integer(kind=xx), parameter :: SOME_DATA(*) = [ z'..', z'..', .. ]<br>><br>> but the current standard doesn't seem to allow this.<br>><br><br>Correct. There was a proposal for Fortran 2008 to include a “BITS” data type that would have enabled something like this. However, in an attempt to reduce the size of the changes to the standard, it was dropped. While there were no actual technical problems with [...]
2016-04-25T15:20:05+00:00Bill Longhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;bae8b856.1604Re: Missing Fortran Forum issues
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;63b8dcb7.1604
Hi Loren<br><br>I just went to the ACM site<br><br>http://dl.acm.org/citation.cfm?id=J286<br><http://dl.acm.org/citation.cfm?id=J286&picked=prox&cfid=775080374&cftoken=5<br>1756888> &picked=prox&cfid=775080374&cftoken=51756888<br><br>They go back to 1982. Worst case scenario would be to 'print'<br><br>The missing ones.<br><br>Cheers<br><br>Ian Chivers<br><br>Current FF editor.<br><br>From: Fortran 90 List [mailto:COMP-FORTRAN-90@JISCMAIL.AC.UK] On Behalf Of<br>Loren P Meissner<br>Sent: 23 April 2016 02:36<br>To: COMP-FORTRAN-90@JISCMAIL.AC.UK<br>Subject: Missing Fortran Forum issues<br><br>I have been looking at the stuff on my library shelf that I am no longer<br>using. [...]
2016-04-23T09:31:41+01:00Ian Chivershttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;63b8dcb7.1604Re: Missing Fortran Forum issues
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;f072f80c.1604
Loren:<br><br>Nice to hear from you.<br><br>I tried to convince Jeanne Martin to give her collection of J3 papers to<br>CHM (I think she said 12 shelf-feet). I don't remember why that worked<br>out. If you still have contact with her, maybe you can convince her to<br>do it. From 07-007:<br><br>Jeanne T. Martin<br>11 Manti Terrace<br>Alamo, CA 94507 [...]
2016-04-22T19:00:34-07:00Van Snyderhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;f072f80c.1604Missing Fortran Forum issues
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;6e7222f4.1604
I have been looking at the stuff on my library shelf that I am no longer using.<br><br>I haven't used Fortran for a long time, but I have a lot of documents from "the olden days."<br><br>I have begun discussions with Computer History Museum at Mountain View CA (not far from San Jose where I live), toward contributing some documents. [...]
2016-04-23T01:35:38+00:00Loren P Meissnerhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;6e7222f4.1604How to use BOZ-literal-constants in an integer array constructor?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;b88fc76.1604
Hi,<br><br>I need to construct an integer array of a named constant<br>using BOZ-literal-constants as in<br><br>integer(kind=xx), parameter :: SOME_DATA(*) = [ z'..', z'..', .. ]<br><br>but the current standard doesn't seem to allow this.<br><br>The standard states:<br><br>C4102 (R463) A boz-literal-constant shall appear only as a<br>data-stmt-constant in a DATA statement, or where<br>explicitly allowed in subclause 13.7 as an actual argument of an intrinsic<br>procedure [...]
2016-04-22T20:41:55-04:00Vipul Parekhhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;b88fc76.1604Re: reading from input_unit intrinsic assignment?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;9d6ba149.1604
For READ and WRITE statements.<br><br>"If an input item or an output item is allocatable, it shall be allocated.”<br><br>which would contradict the idea of reallocation, since it also applies to unallocated objects.<br><br>In the section on character input formatting for list-directed input:<br><br>"Let len be the length of the next eective item, and let w be the length of the character sequence. If len is less than or equal to w, the leftmost len characters of the sequence are transmitted to the next eective item. If len is greater than w, the sequence is transmitted to the leftmost w [...]
2016-04-13T12:16:59+00:00Bill Longhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;9d6ba149.1604Re: reading from input_unit intrinsic assignment?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;4ba19b40.1604
>I think you provided an allocated variable to be read into and<br>>because it was zero length no data was transferred.<br><br>I expected the variable to be reallocated<br>with the length equal to the length of the<br>input string.<br><br>>> I cannot see from 7.2.1.2 or 7.2.1.3<br>>> whether reading character variable<br>>> from input_unit is an intrinsic assignment or not?<br>><br>>Seems quite clear to me (quote is not from the version you quoted)<br>> "An intrinsic assignment statement is an assignment statement..." [...]
2016-04-13T12:09:22+01:00Anton Shterenlikhthttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;4ba19b40.1604Re: reading from input_unit intrinsic assignment?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;6fca7e48.1604
I think you provided an allocated variable to be read into and<br>because it was zero length no data was transferred.<br><br>> I cannot see from 7.2.1.2 or 7.2.1.3<br>> whether reading character variable<br>> from input_unit is an intrinsic assignment or not?<br><br>Seems quite clear to me (quote is not from the version you quoted)<br>"An intrinsic assignment statement is an assignment statement..." [...]
2016-04-13T10:41:38+00:00Harvey Richardsonhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;6fca7e48.1604reading from input_unit intrinsic assignment?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;b8cca2b6.1604
character(:), allocatable :: word<br>word = ""<br>read (*,*) word<br>write (*,*) word, len(word)<br>end<br><br>It seems "word" is not reallocated<br>when READ statement is executed.<br>len(word) always returns 0.<br><br>16-007 in 6.7.3.2p7 says "When an intrinsic assignment<br>statement (7.2.1.3) is executed, any noncoarray<br>allocated allocatable subobject is deallocated."<br><br>I cannot see from 7.2.1.2 or 7.2.1.3<br>whether reading character variable<br>from input_unit is an intrinsic assignment or not? [...]
2016-04-13T11:08:11+01:00Anton Shterenlikhthttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;b8cca2b6.1604oracle beta release
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;c6ce7988.1604
There was a post in comp.lang.fortran by<br>Robert Corbett about the latest beta<br>Release of Oracle Solaris Studio<br><br>Here is an extract<br><br>On 2016-04-08, robert.corbett@oracle.com <robert.corbett@oracle.com> wrote:<br><br>> The beta version of the 12.5 release of Oracle Solaris Studio Fortran gets<br>the expected result. The beta is available as a free download from<br>><br>><br>http://www.oracle.com/technetwork/server-storage/solarisstudio/overview/sola<br>ris-studio-beta-program-2798438.html<br>><br>> Robert Corbett
2016-04-09T11:42:02+01:00Ian Chivershttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;c6ce7988.1604Re: What are the possible hurdles to enhanced "enum" functionality in Fortran?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;313297c0.1604
On Fri, 2016-04-01 at 12:02 +0900, Cohen Malcolm wrote:<br>> Or we could actually go back and do<br>> typealias properly (some people might agree with you that "type alias" is a<br>> bad idea, but a very large number of people would strongly disagree).<br><br>When we were debating typealias versus a new type, I had concluded that<br>there was nothing that could be done with typealias that could not be<br>done with a new type, but the opposite is not true. For example,<br>typealias is useless for generic resolution, or even more simply, just<br>for getting the processor to [...]
2016-04-01T11:01:03-07:00Van Snyderhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;313297c0.1604Re: What are the possible hurdles to enhanced "enum" functionality in Fortran?
https://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;8de3b1a1.1604
Right, that is precisely a proper version of the second kind I mentioned.<br>The "new type" requirement being associated with the *first kind*.<br><br>That is, "first kind" ~= "like the Ada facility", you get a new type;<br>"second kind" ~= "like the C facility", you get an alias for an intrinsic<br>integer type+kind.<br><br>Both kinds have some advantages, but many (most?) use cases can use either.<br>A few use cases require a specific one: some the "first kind" (new type),<br>some the "second kind" (integer type). [...]
2016-04-01T12:02:59+09:00Cohen Malcolmhttps://www.jiscmail.ac.uk:443/cgi-bin/webadmin?A2=COMP-FORTRAN-90;8de3b1a1.1604