Hi Patrik,
When a large volume of sequential data is being processed the difference between a CPU bound workload and an I/O bound workload is determined by the performance and throughput of the supporting I/O subsystem. Actual throughput becomes a function of how fast input buffers can be replenished and output buffers emptied. While more available memory can be used to greatly increase buffers to smooth out the CPU processing ¨C I/O ?overlap sooner or later if there are no buffers available then throughput is determined by the I/O subsystem performance.
Of course, additional memory can be allocated to use larger data blocksizes and DASD full track blocking. In my work with the OS/360 Sort/Merge program I compared the use of 2341 DASD with a data block size and track size of approximately 7K to a half track and a data blocksize of approximately 28K on 3390 DASD. The I/O count dropped by a factor of 4 and so did the elapsed time. With the 2314 example approximately 84K was allocated to buffers. With the half track 3390 example 340K was allocated to buffers.
For non sequential processing additional memory can be utilized to hold an increased number and levels of indexes reducing index I/O operations. Pretty much any program can be changed from being I/O bound to CPU bound by the use of an SSD for data storage. In summary, additional region size is only of benefit if the processing program is able to effectively use it to reduce I/O operations.
Regards
Tom
|
Re: Real programmers don't write in FORTRAN. FORTRAN is for pipe stress freaks and crystallography weenies. SAYS MY Tk4- COOKIE JAR
On Sun, 29 Nov 2020 15:07:50 -0000, "Dave Wade" <dave.g4ugm@...> wrote: << let me fix this>>
Folks,
Well for a long time we bashed, bent stretched and squished FORTRAN
into various distortions because we had no other language. Even the
folks behind ¡°C¡± and Unix recognized this and produced RATFOR for
those of us stuck with FORTRAN. I use to think that the best programming language is, first of all, the one that is available. If more than one is available, the better is the one that: 1- Gets the work done in an useful way. 2- I am more familiar with. Once I was in a field operation (199x, no net, no telephony, and hours flight away from any civilization) and had the need to have a program just to compare two CSV files. About 15 columns and 200 lines. The only available language was the ms-basic available in a Windows <something>. In 15 minutes I had a first program that were doing in two seconds what people was doing in almost one hour. And making mistakes doing that. Cheers! Roxo -- ---------------- Non luctari, ludare -------------------+ WYSIWYG Fernando M. Roxo da Motta <mvs@...> | Editor? Except where explicitly stated I speak on my own behalf.| VI !! PU5RXO | I see text, ------------ Quis custodiet ipsos custodes?-------------+ I get text!
|
I think this is probably a continuation ?card of a previous comment card.
Joe
On Sun, Nov 29, 2020 at 5:03 PM Bill Lewis < wrljet@...> wrote: I recently purchased an old punch card cabinet from the estate sale of a local hobby
machinist club member.
There was one single punch card in the bottom of the cabinet.? And I'm wondering, is this a typo?
???? // *
"slash slash space star"
Bill
 
|
I recently purchased an old punch card cabinet from the estate sale of a local hobby machinist club member. There was one single punch card in the bottom of the cabinet.? And I'm wondering, is this a typo?
???? // *
"slash slash space star" Bill  
|
Re: IBM Assembler History
Marco, I think the list here is pretty complete, but its missing a timeline ?
? I also think its missing the ¡°SLAC¡± modifications to Assembler ¡°H¡± ?
? and it says XF was an upgrade to ¡°F¡± but I think it was a total re-write. ? You won¡¯t find Assembler H in the wild it always was a chargeable product. ? These days there is the also the Z390 assembler which is written in JAVA so should run anywhere JAVA does. ? Dave G4UGM ? ? ?
toggle quoted message
Show quoted text
From: [email protected] <[email protected]> On Behalf Of marcoxa@...
Sent: 29 November 2020 14:37
To: [email protected]
Subject: [H390-MVS] IBM Assembler History? Hi
one question that I have for the knowledgeable folks here.? I gathered, by lurking, that there has been a number of different assemblers in MVS and z/OS (and, I presume, VM, and possibly MTS and MUSIC/SP).? I have read about IFOX, ASSIST, etc etc
Is there a link anywhere with a timeline of these tools?? Or maybe somebody here would care to comment?
I am particularly interested in understanding what is currently available on MVS 3.8j/TK4- and z/OS.
Thanks
--
Marco Antoniotti
Somewhere over the rainbow
|
Re: IBM Assembler History
Yes, thank you for that. I forgot to post the link to that section, which is? long way down the page :-)
Roops
Thank you Rupert.? Following from the link you posted you get to
which contains more information.
--
Marco Antoniotti
Somewhere over the rainbow
|
Re: IBM Assembler History
Thank you Rupert.? Following from the link you posted you get to
which contains more information.
--
Marco Antoniotti
Somewhere over the rainbow
|
Re: IBM Assembler History
I found this. Any comments on accuracy?
Roops
[Edited Message Follows]
Yep.
It's be nice to have a table with
??? Assembler name, IBM native Y/N, Year released, Program name, AKA/Nicknames/Aliases, Features
With "Features" obviously being a kitchen sink.
All the best
--
Marco Antoniotti
Somewhere over the rainbow
|
Re: IBM Assembler History
Yep.
It's be nice to have a table with
??? Assembler name, IBM native Y/N, Year released, Program name, AKA/Nicknames/Aliases, Features
With "Features" obviously being a kitchen sink.
All the best
--
Marco Antoniotti
Somewhere over the rainbow
|
Re: IBM Assembler History
I keep forgetting this, too.?
I have IFOX00 (Assembler XF) in TK4- and there was one (IEUASM--Assembler F, which came from MVT, was it?) which might have difficulty with certain macros or conditional assembly.
After that, the next I was aware of was IEV90 (Assembler H), the one I did most of my paid work with.
Finding out which you have isn't always obvious, as often one name is an alias in the load library :-)
Roops
Hi
one question that I have for the knowledgeable folks here.? I gathered, by lurking, that there has been a number of different assemblers in MVS and z/OS (and, I presume, VM, and possibly MTS and MUSIC/SP).? I have read about IFOX, ASSIST, etc etc
Is there a link anywhere with a timeline of these tools?? Or maybe somebody here would care to comment?
I am particularly interested in understanding what is currently available on MVS 3.8j/TK4- and z/OS.
Thanks
--
Marco Antoniotti
Somewhere over the rainbow
|
Re: revisiting SYSGEN in Z world
OK ... for the last time, you CANNOT SYSGEN on any OS since MVS/ESA.
In MVS/ESA, IBM replaced the SYSGEN process with?MVSCP/IOCP. The reason was simple, as explained:
So, IBM changed the game to MVSCP/IOCP in MVS/ESA:
Subsequently, in z/OS, they migrated to a new tool, called HCD:
So, SYSGEN does not exist anymore. It has been replaced. Unlike with SYSGEN'd systems, there is now only one copy of the nucleus that?supports multiple I/O configurations.
Joe
On Sun, Nov 29, 2020 at 6:55 AM Rahim Azizarab via <rahimazizarab= [email protected]> wrote: That is what the myth is;? and that is exactly what I was trying to get around of by copying the DASD into an old MVS 38 system and doing the file allocations under my newly created master catalog then after allocations I can bring it back in to populate the datasets.
Zos does not allow the use of STEPCAT which trows a monkey wrench into the sysgen process.
On Saturday, November 28, 2020, 8:39:21 PM CST, Joe Monk < joemonk64@...> wrote:
Once again, you cannot sysgen on z/os, in fact I think since OS/390.?
Joe On Sat, Nov 28, 2020 at 7:49 PM Greg Price < procegrog@...> wrote: On 2020-11-29 11:05 AM, Rahim Azizarab via wrote:
> but when it came time to allocate my system datasets (SYS1) under my
> newly created master catalog Zos does not allow STEPCAT or JOBCAT to
> catalog my system files under my newly created catalog
It seems a problem has been needlessly invented, yes?
STEPCAT/JOBCAT is not needed under Z.? The CATALOG operand of IDCAMS
commands allows the target catalog to be specified.? Non-VSAM data sets
can be allocated and loaded without being cataloged.? Data sets on the
IPL volume should not be cataloged "by the system" (such as with a DISP
of CATLG) but should be cataloged indirectly such as by
DEF NVSAM (NAME(dsname) DEVT(3390) VOL(&SYSR1)) CAT(mynewcatalog/mynewpswd)
or similar, to allow for swapping IPL volumes using the same catalog and
paging configuration.
It's not as if you can port a catalog between z/OS and MVS 3.8 in either
direction...
Cheers,
Greg
|
Re: Real programmers don't write in FORTRAN. FORTRAN is for pipe stress freaks and crystallography weenies. SAYS MY Tk4- COOKIE JAR
<< let me fix this>> ? Folks, Well for a long time we bashed, bent stretched and squished FORTRAN into various distortions because we had no other language. Even the folks behind ¡°C¡± and Unix recognized this and produced RATFOR for those of us stuck with FORTRAN. ? When I worked on Networking software for Salford University for UK Universities, we wrote much of it in Fortran 77. It was not an ideal language, BUT FORTRAN 77 was the one language every UK University had. It would probably have been much nicer in ¡°C¡± but even if the Universities owned ¡°C¡± it was probably a different flavour in each University. The University had already written a Fortran 77 compiler in Fortran 77 so we re-used some of the techniques to build X25 Networking code which ran on MVS, VM, PRIMOS plus a few others. ? It was a challenge, so we had a macro processor that allowed us to create our main program as a basic message switch. Then because we had no dynamic storage, we had a vector of 256 byte strings that was used for storing ¡°control information¡±. So for example for a file transfer there would be one character string containing information about the source file, one for the destination file. The locations within each string were defined by constants read by include file. There was a matching array of integers which allowed each the strings to be chained together, and the type of each string to be defined¡ ? As for Fortran 77, generally on a particular platform they all had the Vendors FORTRAN 77. So DEC VAX users had DEC Fortran, IBM users VS Fortran, Honeywell users, Honeywell Fortran. The exception was PR1ME. Salford had PR1MES and when they were benchmarked the Fortran77 compiler was not available, so the benchmarks were don with Fortran 66. When PR1ME¡¯s Fortran 77 was delivered, it was, to put it mildly a ¡°right dog¡¯s dinner¡±. Salford¡¯s had written a FORTRAN 77 compiler for the ICL1900. In those days UK University computers were supposed to last 10 years, so although ICL replaced the ICL1900 hardware with a 2900 emulating the 1900, they made them keep the old software so only Fortran 66. So they wrote a Fortran 77 compiler as a stop gap. When PR1ME FORTRAN 77 was found to be terrible it was ported to PR1MOS. Later it was ported to MSDOS as FTN77. Today its still available, upgraded Fortran 95 as SilverFrost FTN85¡. ? Dave G4UGM ?
toggle quoted message
Show quoted text
From: [email protected] <[email protected]> On Behalf Of sfbusfbu
Sent: 29 November 2020 01:04
To: [email protected]
Subject: Re: [H390-MVS] Real programmers don't write in FORTRAN. FORTRAN is for pipe stress freaks and crystallography weenies. SAYS MY Tk4- COOKIE JAR? I still find myself thinking of the underlying assembly, even when I'm using a compiled language :-)
All programmers should think about what their higher level constructs and statements are doing to the hw.
|
Hi
one question that I have for the knowledgeable folks here.? I gathered, by lurking, that there has been a number of different assemblers in MVS and z/OS (and, I presume, VM, and possibly MTS and MUSIC/SP).? I have read about IFOX, ASSIST, etc etc
Is there a link anywhere with a timeline of these tools?? Or maybe somebody here would care to comment?
I am particularly interested in understanding what is currently available on MVS 3.8j/TK4- and z/OS.
Thanks
--
Marco Antoniotti
Somewhere over the rainbow
|
Re: Real programmers don't write in FORTRAN. FORTRAN is for pipe stress freaks and crystallography weenies. SAYS MY Tk4- COOKIE JAR
Folks, Well for a long time we bashed, bent stretched and squished FORTRAN into various distortions because we had no other language. ? When I worked on Networking software for Salford University for UK Universities, we wrote much of it in Fortran 77. It was not an ideal language, BUT FORTRAN 77 was the one language every University had. It would probably have been much nicer in ¡°C¡± but even if the Universities owned ¡°C¡± it was probably a different flavour in each University. Generally on a particular platform they all had the Vendors FORTRAN 77. So DEC VAX users had DEC Fortran, IBM users VS Fortran, Honeywell users, Honeywell Fortran. The exception was PR1ME. Salford had PR1MES and when they were benchmarked the Fortran77 compiler was not available, so the benchmarks were don with Fortran 66. When PR1ME¡¯s Fortran 77 was delivered, it was, to put it mildly a ¡°right dogs dinner¡±. ? Oddly Salford had written a FORTRAN 77 compiler. ?
toggle quoted message
Show quoted text
From: [email protected] <[email protected]> On Behalf Of sfbusfbu
Sent: 29 November 2020 01:04
To: [email protected]
Subject: Re: [H390-MVS] Real programmers don't write in FORTRAN. FORTRAN is for pipe stress freaks and crystallography weenies. SAYS MY Tk4- COOKIE JAR? I still find myself thinking of the underlying assembly, even when I'm using a compiled language :-)
All programmers should think about what their higher level constructs and statements are doing to the hw.
|
Re: revisiting SYSGEN in Z world
On 2020-11-29 11:55 PM, Rahim Azizarab via groups.io wrote: That is what the myth is;? and that is exactly what I was trying to get around of by copying the DASD into an old MVS 38 system and doing the file allocations under my newly created master catalog then after allocations I can bring it back in to populate the datasets. Um, is this what a circular argument is? Razim: Z world does not do STEPCAT (which is useful in some sort of SYSGEN context) but I can use MVS 3.8 to advantage to bypass this limitation. Joe: SYSGEN does not exist in Z. Greg: You don't need STEPCAT in Z to do anything you want to do in Z. Joe: SYSGEN does not exist in Z. Razim: That's exactly the problem I can bypass by some weird (and what I (Greg) say is totally unnecessary) cross-generational procedure. I wonder when I can retire... :) I expect I've missed some pivotal point which would make sense of it all.? I apologise for my shortcomings. Cheers, Greg
|
Re: Hercules Optimization and Tuning
ThanX.
As we now these days cpu has 2 thread and seen double core as real core in OS.
In hercules MAXCPU is max thread i think or max core and not max real cpu core( i dont know) and not max cpu slot(for server may have 2 or 4 slot cpu).
First what is your meaning by cpu in your reply? (core or thread)
Second if i use command like settask? should i have better performance?
|
Re: revisiting SYSGEN in Z world
That is what the myth is;? and that is exactly what I was trying to get around of by copying the DASD into an old MVS 38 system and doing the file allocations under my newly created master catalog then after allocations I can bring it back in to populate the datasets.
Zos does not allow the use of STEPCAT which trows a monkey wrench into the sysgen process.
On Saturday, November 28, 2020, 8:39:21 PM CST, Joe Monk <joemonk64@...> wrote:
Once again, you cannot sysgen on z/os, in fact I think since OS/390.?
Joe
On Sat, Nov 28, 2020 at 7:49 PM Greg Price < procegrog@...> wrote: On 2020-11-29 11:05 AM, Rahim Azizarab via wrote:
> but when it came time to allocate my system datasets (SYS1) under my
> newly created master catalog Zos does not allow STEPCAT or JOBCAT to
> catalog my system files under my newly created catalog
It seems a problem has been needlessly invented, yes?
STEPCAT/JOBCAT is not needed under Z.? The CATALOG operand of IDCAMS
commands allows the target catalog to be specified.? Non-VSAM data sets
can be allocated and loaded without being cataloged.? Data sets on the
IPL volume should not be cataloged "by the system" (such as with a DISP
of CATLG) but should be cataloged indirectly such as by
DEF NVSAM (NAME(dsname) DEVT(3390) VOL(&SYSR1)) CAT(mynewcatalog/mynewpswd)
or similar, to allow for swapping IPL volumes using the same catalog and
paging configuration.
It's not as if you can port a catalog between z/OS and MVS 3.8 in either
direction...
Cheers,
Greg
|
Re: Real programmers don't write in FORTRAN. FORTRAN is for pipe stress freaks and crystallography weenies. SAYS MY Tk4- COOKIE JAR
I still find myself thinking of the underlying assembly, even when I'm using a compiled language :-)
All programmers should think about what their higher level constructs and statements are doing to the hw.
|
yes, you are right, one thing is cpu clock, and other is peripheral troughput. Anyway I was amazed by the speed in sort and sequential string seek , in 32 k data sets I saw in my PC.
At the time I was a student, all the students records ( about 50.000, with about 5000 active? ) in Engineering School, at Buenos Aires University, were handled with an IBM 360 model 40, with a couple of Wichester disks, and some tapes. they used a lot PL/I SortA, and I imagine the master databases were ISAM. And this was really fast.
May be I can do some comparation with a differential equation solver, We ran it at IBM, and took 14 minutes of pure calculations on an IBM 370, to solve a grid of heath transfer.
?Regards
|
Hello Tom, Am 28.11.2020 um 23:52 schrieb tomarmstrong255@...: Providing additional memory, above and beyond their stated memory specifications generally delivers little or no performance improvement as they are not designed to exploit the additional memory provided by MVS. Your statement is obviously true for CPU bound workloads, but is it still true when the program has to gnaw through datasets of manifold size compared to available memory region the task is running within? :wq! PoC
|
Marco Antoniotti
PD
carlos feldman