Re: Floating point assembler code
All Here is another alternative for coding and testing all 3 flavors of IBM mainframe floating point including HFP, BFP, and DFP. The free open source zcobol compiler written in structured z390 portable mainframe assembler supports all the floating point formats using the latest COBOL draft standard for defining each of the 9 different floating point types in the data division. In the procedure division you can code compute statements to perform any floating point calculations you like mixing and matching any of the floating point types along with traditional data types. You can use if statements to test results, and you can use display statements to display results of compute statements. The zcobol compiler generates HLASM compatible assembler source code with data labels and paragraph labels so the generated code can be run using z390 emulator on Windows or Linux or you can take the generated assembler source to mainframe and assemble and execute it using HLASM. The zcobol code generator supports the latest z9, z10, and z196 instructions and uses the very powerful PFPO for conversions between floating point data types as required. When you download and install z390 via www.z390.org, it includes zcobol plus a number of demo and regression test programs such as the following: 1. z390\zcobol\test\TESTHFP1.CBL - test hex HFP data type calculations 2. z390\zcobol\test\TESTBFP1.CBL - test binary BFP data type calculations 3. z390\zcobol\test\TESTDFP1.CBL - test decimal DFP data type caculations Once z390 is installed on Windows or Linux, you can run zcobol demos with the command ZCRT390.BAT from GUI or command line. You can also run all the zcobol regression tests including the above 3 with the command ZCRTTEST.BAT. Note z390 v1.5.03 has new z196 instructions added, but the next release v1.5.04 will add some missing extended opcodes which are documented in SHARE presentation but are not in the latest POP. Don Higgins d...@higgins.net -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Floating point assembler code
I gave up TSO TEST for ASMIDF years ago. Give it try. In article you wrote: > On Tue, 5 Apr 2011 15:36:24 -0400, Tony Harminc > wrote: > > > >I too continue to use TSO TEST for certain things, and indeed it would > >be a very good thing if they would add just a few little updates: > > > >Better floating point support, as you suggest > > > >Access register support > > > >64-bit GPR support > > > >64-bit indirection (the IPCS syntax would be consistent) > > > >TEST (and other TSO components) are extra frustrating because however > >dead-ended TSO may be, there *are* updates going on, but they never > >seem to take the opportunity to make minor enhancements while they > >have a module open for other purposes. Yeah, I know - internal > >politics, time & budget allocation, and so on. But most development > >labs manage to sneak in a few enhancements under the RADAR from time > >to time. > > > Yes, I've discussed some of those with the TSO guy. The basic problem with > all of them seems to be that they hit the TSO TEST register save and restore > processing and that is non-trivial to upgrade plus its hard to do piecemeal. > I've done a fair amount of AMODE(64) testing on TEST and it seems to run > AMODE(64) stuff just fine; its just that you can't see the addresses when > they're in registers... :-( > Writing the code is the easy part; its getting the resources to thoroughly > test > what you've written that stops you dead in your tracks. > W. Kevin Kelley -- IBM POK Lab -- z/OS Core Technical Development > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO > Search the archives at http://bama.ua.edu/archives/ibm-main.html -- Don Poitras - SAS Development - SAS Institute Inc. - SAS Campus Drive sas...@sas.com (919) 531-5637Cary, NC 27513 -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Floating point assembler code
On Tue, 5 Apr 2011 15:36:24 -0400, Tony Harminc wrote: > >I too continue to use TSO TEST for certain things, and indeed it would >be a very good thing if they would add just a few little updates: > >Better floating point support, as you suggest > >Access register support > >64-bit GPR support > >64-bit indirection (the IPCS syntax would be consistent) > >TEST (and other TSO components) are extra frustrating because however >dead-ended TSO may be, there *are* updates going on, but they never >seem to take the opportunity to make minor enhancements while they >have a module open for other purposes. Yeah, I know - internal >politics, time & budget allocation, and so on. But most development >labs manage to sneak in a few enhancements under the RADAR from time >to time. > Yes, I've discussed some of those with the TSO guy. The basic problem with all of them seems to be that they hit the TSO TEST register save and restore processing and that is non-trivial to upgrade plus its hard to do piecemeal. I've done a fair amount of AMODE(64) testing on TEST and it seems to run AMODE(64) stuff just fine; its just that you can't see the addresses when they're in registers... :-( Writing the code is the easy part; its getting the resources to thoroughly test what you've written that stops you dead in your tracks. W. Kevin Kelley -- IBM POK Lab -- z/OS Core Technical Development -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Floating point assembler code
On 4 April 2011 21:53, W. Kevin Kelley wrote: > The floating point instructions in some ways are easier to use than the > integer > instructions because you don't have to worry about scaling and implied > decimal points and all of that. The biggest problem with floating point is > converting numbers to floating point and from floating point (any of the 3 > formats). Until the conversion instructions came along to assist in the > effort, > doing the conversions was pretty arcane. Now its easier, but still work. The > book "IBM Assembler Language Programming" by Sharon Tuggle (long out of > print I imagine) does discuss the hexadecimal floating instructions and does > provide sample programs for converting into and out of hexadecimal floating > point. I was the reviewer for the floating point section. Another implementation of conversion routines that makes an interesting reference is from Cornell University, available in the VMSHARE archives at http://vm.marist.edu/~vmshare/read?fn=CONVASM&ft=NOTE&line=1 > I still find TSO TEST useful for testing short segments of Assembler code, and > one of my frustrations is that it only formats the original 4 floating point > registers (0, 2, 4 and 6). It hasn't a clue that there are now 16 floating > point > registers or that there are now 3 floating point formats. In general, most > testing environments give "lip service" to floating point. I too continue to use TSO TEST for certain things, and indeed it would be a very good thing if they would add just a few little updates: Better floating point support, as you suggest Access register support 64-bit GPR support 64-bit indirection (the IPCS syntax would be consistent) TEST (and other TSO components) are extra frustrating because however dead-ended TSO may be, there *are* updates going on, but they never seem to take the opportunity to make minor enhancements while they have a module open for other purposes. Yeah, I know - internal politics, time & budget allocation, and so on. But most development labs manage to sneak in a few enhancements under the RADAR from time to time. Tony H. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Floating point assembler code
On Mon, 4 Apr 2011 20:53:35 -0500 "W. Kevin Kelley" wrote: :>On Mon, 4 Apr 2011 11:09:29 -0400, O'Brien, David W. (NIH/CIT) [C] :> wrote: :>>Are there any tutorials/samples for writing floating point assembler code. We :>need to work with the z/VM Performance data and it is stored in short floating :>point. Been playing in the Principles of Operation but have a bit of a time :>wrapping my head around floating point manipulation. :>The floating point instructions in some ways are easier to use than the integer :>instructions because you don't have to worry about scaling and implied :>decimal points and all of that. The biggest problem with floating point is :>converting numbers to floating point and from floating point (any of the 3 :>formats). Until the conversion instructions came along to assist in the effort, :>doing the conversions was pretty arcane. Now its easier, but still work. The :>book "IBM Assembler Language Programming" by Sharon Tuggle (long out of :>print I imagine) does discuss the hexadecimal floating instructions and does :>provide sample programs for converting into and out of hexadecimal floating :>point. I was the reviewer for the floating point section. :>I still find TSO TEST useful for testing short segments of Assembler code, and :>one of my frustrations is that it only formats the original 4 floating point :>registers (0, 2, 4 and 6). It hasn't a clue that there are now 16 floating point :>registers or that there are now 3 floating point formats. In general, most :>testing environments give "lip service" to floating point. Why not suggest to IBM that they enhance it? :>W. Kevin Kelley -- IBM POK Lab -- z/OS Core Technical Development Perhaps the office upstairs on the left? -- Binyamin Dissen http://www.dissensoftware.com Director, Dissen Software, Bar & Grill - Israel Should you use the mailblocks package and expect a response from me, you should preauthorize the dissensoftware.com domain. I very rarely bother responding to challenge/response systems, especially those from irresponsible companies. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Floating point assembler code
On Mon, 4 Apr 2011 11:09:29 -0400, O'Brien, David W. (NIH/CIT) [C] wrote: >A colleague asked that the following query be posted: > >Are there any tutorials/samples for writing floating point assembler code. We need to work with the z/VM Performance data and it is stored in short floating point. Been playing in the Principles of Operation but have a bit of a time wrapping my head around floating point manipulation. > Dave, The floating point instructions in some ways are easier to use than the integer instructions because you don't have to worry about scaling and implied decimal points and all of that. The biggest problem with floating point is converting numbers to floating point and from floating point (any of the 3 formats). Until the conversion instructions came along to assist in the effort, doing the conversions was pretty arcane. Now its easier, but still work. The book "IBM Assembler Language Programming" by Sharon Tuggle (long out of print I imagine) does discuss the hexadecimal floating instructions and does provide sample programs for converting into and out of hexadecimal floating point. I was the reviewer for the floating point section. I still find TSO TEST useful for testing short segments of Assembler code, and one of my frustrations is that it only formats the original 4 floating point registers (0, 2, 4 and 6). It hasn't a clue that there are now 16 floating point registers or that there are now 3 floating point formats. In general, most testing environments give "lip service" to floating point. W. Kevin Kelley -- IBM POK Lab -- z/OS Core Technical Development -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Floating point assembler code
On Mon, Apr 4, 2011 at 8:09 AM, O'Brien, David W. (NIH/CIT) [C] wrote: > A colleague asked that the following query be posted: > > Are there any tutorials/samples for writing floating point assembler code. We > need to work with the z/VM Performance data and it is stored in short > floating point. Been playing in the Principles of Operation but have a bit of > a time wrapping my head around floating point manipulation. You may want to try looking at compiler listings for unoptimized (or optimized) COBOL code using COMP-1 (short float) or COMP-2 (long float). By using the various compiler options (to choose between the various types of floating point numbers) and doing arithmetic and/or moving floating point number to Binary and/or Packed number you will be able to generate a quick tutorial for yourself. The same can be done with the C/C++ compiler other other compilers on your system. Sam > > Thank You, > Dave O'Brien > NIH Contractor > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO > Search the archives at http://bama.ua.edu/archives/ibm-main.html > -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Floating point assembler code
On 4/4/2011 9:09 AM, O'Brien, David W. (NIH/CIT) [C] wrote: A colleague asked that the following query be posted: Are there any tutorials/samples for writing floating point assembler code. We need to work with the z/VM Performance data and it is stored in short floating point. Been playing in the Principles of Operation but have a bit of a time wrapping my head around floating point manipulation. Thank You, Dave O'Brien NIH Contractor You might have better luck posting this on the IBM Assembler list: IBM Mainframe Assembler List Although I'm sure there are experienced people here, too. -- Kind regards, -Steve Comstock The Trainer's Friend, Inc. 303-393-8716 http://www.trainersfriend.com * To get a good Return on your Investment, first make an investment! + Training your people is an excellent investment * Try our new tool for calculating your Return On Investment for training dollars at http://www.trainersfriend.com/ROI/roi.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Floating point assembler code
A colleague asked that the following query be posted: Are there any tutorials/samples for writing floating point assembler code. We need to work with the z/VM Performance data and it is stored in short floating point. Been playing in the Principles of Operation but have a bit of a time wrapping my head around floating point manipulation. Thank You, Dave O'Brien NIH Contractor -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html