Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
> On 7 Aug 2023, at 2:46 pm, Timothy Sipples wrote: > > David Crayford wrote: >> Maybe wait until there is actually some tangible AI libraries such as >> TensorFlow, PyTorch and SnapML before blowing trumpets. > > Huh? You *can* run these libraries on z/OS, on zIIPs even. They run on the > z/OS Container Extensions (zCX) or on OpenShift for z/OS, as you prefer. IBM > documents this deployment pattern here (TensorFlow and SnapML examples): > > https://ibm.github.io/ai-on-z-101/tensorflow/ > https://ibm.github.io/ai-on-z-101/snapml/ > Absolutely! The issue is that zCX is not a mature technology. zCX on OpenShift has serious performance issues. It current hogs 5 zIIPS running idle (only running OCP). Of course, it will improve just like DB2, Java etc but it’s not ready for prime time yet. > Are you asking specifically for z/OS UNIX System Services-based > implementations? Of course. Considering the zDDN library is available on z/OS I expected the Python libraries to be available at the same time as Linux on Z. > If so, have you asked IBM in an official way? Yes. It’s in the pipeline but we can’t give you a date :) A bit like Java 17. And that’s a bigger problem because Spring Boot moves to a Java 17 as the baseline in December 2023 which is making a lot of vendors slightly nervous. Although as CICS uses SB I have high hopes. > > — > Timothy Sipples > Senior Architect > Digital Assets, Industry Solutions, and Cybersecurity > IBM zSystems/LinuxONE, Asia-Pacific > sipp...@sg.ibm.com > > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
David Crayford wrote: >Maybe wait until there is actually some tangible AI libraries such as >TensorFlow, PyTorch and SnapML before blowing trumpets. Huh? You *can* run these libraries on z/OS, on zIIPs even. They run on the z/OS Container Extensions (zCX) or on OpenShift for z/OS, as you prefer. IBM documents this deployment pattern here (TensorFlow and SnapML examples): https://ibm.github.io/ai-on-z-101/tensorflow/ https://ibm.github.io/ai-on-z-101/snapml/ Are you asking specifically for z/OS UNIX System Services-based implementations? If so, have you asked IBM in an official way? — Timothy Sipples Senior Architect Digital Assets, Industry Solutions, and Cybersecurity IBM zSystems/LinuxONE, Asia-Pacific sipp...@sg.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Channelized I/O WAS
> From Parwez: My mistake, the 370/195 had 2 MB, this customer's 360/75 had 1 MB In those ancient days an MB of memory was $$expensive$$ and fairly rare. In the very early 70s I worked in an installation that had two 360/75s, each with 3 MB (1 MB normal memory and 2 MB LCS). The second 75 was just for backup! I was impressed -- the largest systems I had touched before that were two 360/65s with large (256KB) memory and even these were impressive. Times have changed. (My primary laptop has 64 GB memory.) Bill Ogden -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
Also ADP. From: IBM Mainframe Discussion List on behalf of Bob Bridges Sent: Friday, August 4, 2023 8:53 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives Right, I think it was "EDP" (electronic data processing) when I started. Or maybe even that wasn't the first one I was aware of; it's been a long time now. --- Bob Bridges, robhbrid...@gmail.com, cell 336 382-7313 /* Neither irony nor sarcasm is argument. -Samuel Butler */ -Original Message- From: IBM Mainframe Discussion List On Behalf Of Seymour J Metz Sent: Friday, August 4, 2023 07:14 Also 'Data Processing'; I vaguely recall that there were a few more terms. From: P H [04843e86df79-dmarc-requ...@listserv.ua.edu] Sent: Friday, August 4, 2023 5:23 AM > BTW: When I started my career during the early 70s, IT didn't exist. It was > 'computers' or 'computing'. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
Right, I think it was "EDP" (electronic data processing) when I started. Or maybe even that wasn't the first one I was aware of; it's been a long time now. --- Bob Bridges, robhbrid...@gmail.com, cell 336 382-7313 /* Neither irony nor sarcasm is argument. -Samuel Butler */ -Original Message- From: IBM Mainframe Discussion List On Behalf Of Seymour J Metz Sent: Friday, August 4, 2023 07:14 Also 'Data Processing'; I vaguely recall that there were a few more terms. From: P H [04843e86df79-dmarc-requ...@listserv.ua.edu] Sent: Friday, August 4, 2023 5:23 AM > BTW: When I started my career during the early 70s, IT didn't exist. It was > 'computers' or 'computing'. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
Data Processing, most probably from DP division of IBM of that time. Sent from Outlook for Android<https://aka.ms/AAb9ysg> From: IBM Mainframe Discussion List on behalf of Seymour J Metz Sent: Friday, August 4, 2023 12:14:07 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives > The one I worked on at a sister (can I say this or should it be 'person' > organisation of CERN) had a grand total of 1 MB main memory! That sounds more appropriate for a 65 than a 195. > BTW: When I started my career during the early 70s, IT didn't exist. It was > 'computers' or 'computing'. Also 'Data Processing'; I vaguely recall that there were a few more terms. -- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 From: IBM Mainframe Discussion List [IBM-MAIN@LISTSERV.UA.EDU] on behalf of P H [04843e86df79-dmarc-requ...@listserv.ua.edu] Sent: Friday, August 4, 2023 5:23 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives In response to your comments and some made by others, my 2 cents worth. This discussion started talking about mainframes and 'split' into sub-threads questioning/focusing on IBM z e.g., z/Architecture, what has z ever done, any uniqueness/special features of z etc. In my response, I tried to answer some Qs and correcting some of the numbers which were being quoted. I did end by saying 'horses for courses'. No single system/platform is perfect. All have their uniqueness, strengths and weakness. Today, other platforms may have similar functions/features as z and some may be even better, the point is during the evolution of z under it's different marketing names (S/360, S/370, S/390, eServer, System z etc) z has evolved, adapted and embraced technologies which businesses require for modern workloads. z continues to evolve! z can't be everything to everyone. There are alternatives to 'mainframes' both from IBM (POWER) and others. Talking about weakness, as an example I did mention that my x86 also has limitations. I don't need a better machine, especially an overpriced Mac. Just like some customers think they don't need an overpriced z. With today’s mind set of 'good enough' computing i.e. if it doesn't work reboot, if what you have meets your needs, so be it. Just like your iPhone, my smartphone has most probably more power/memory than S370/195 of early 70s. The one I worked on at a sister (can I say this or should it be 'person' organisation of CERN) had a grand total of 1 MB main memory! However, I doubt our smartphones could process tons of data generated by accelerators causing collisions of energetic particles to investigate the structure of the atomic nucleus. Even during the 70's S370/195 did it very successfully i.e., process large amounts of data (strength of the I/O subsystem ??). Yes, there are other suppliers of MF like systems. Someone else on this thread mentioned that they saw a demo of one which had similar RAS capabilities as of z. I am familiar with that system. Great demo that lots customers rushed to introduce these into their IT infrastructure. One customer I know of, who soon after Y2K were encouraged by their 'consultants' to ditch their centralised IBM MF installed lots of these systems at all their distributed sites, 3 systems at each site (Dev/Test/Prod). In case of this customer the hype of the demo turned sour as the systems were more 'down' then 'up'. To overcome the RAS deficiencies the solution was to have 'spare' systems on site. No pun intended, needless to say these systems were sunset and almost 20 years later the customer is still using z. I am sure amongst this list, others will have examples of customers ditching other platforms including z for all sorts of reasons. We can debate for ever which system is better etc. At the end of day just put your money where your mouth is into what best suits your IT needs🙂 BTW: When I started my career during the early 70s, IT didn't exist. It was 'computers' or 'computing'. From: IBM Mainframe Discussion List on behalf of David Crayford Sent: 04 August 2023 00:42 To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives > On 3 Aug 2023, at 2:26 am, P H > <04843e86df79-dmarc-requ...@listserv.ua.edu> wrote: > > The numbers quoted by Tom: > > So I pointed out there's only 12 I/O drawers max on a z16 which is 12 x >
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
My mistake, the 370/195 had 2 MB, this customer's 360/75 had 1 MB Sent from Outlook for Android<https://aka.ms/AAb9ysg> From: IBM Mainframe Discussion List on behalf of Seymour J Metz Sent: Friday, August 4, 2023 12:14:07 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives > The one I worked on at a sister (can I say this or should it be 'person' > organisation of CERN) had a grand total of 1 MB main memory! That sounds more appropriate for a 65 than a 195. > BTW: When I started my career during the early 70s, IT didn't exist. It was > 'computers' or 'computing'. Also 'Data Processing'; I vaguely recall that there were a few more terms. -- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 From: IBM Mainframe Discussion List [IBM-MAIN@LISTSERV.UA.EDU] on behalf of P H [04843e86df79-dmarc-requ...@listserv.ua.edu] Sent: Friday, August 4, 2023 5:23 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives In response to your comments and some made by others, my 2 cents worth. This discussion started talking about mainframes and 'split' into sub-threads questioning/focusing on IBM z e.g., z/Architecture, what has z ever done, any uniqueness/special features of z etc. In my response, I tried to answer some Qs and correcting some of the numbers which were being quoted. I did end by saying 'horses for courses'. No single system/platform is perfect. All have their uniqueness, strengths and weakness. Today, other platforms may have similar functions/features as z and some may be even better, the point is during the evolution of z under it's different marketing names (S/360, S/370, S/390, eServer, System z etc) z has evolved, adapted and embraced technologies which businesses require for modern workloads. z continues to evolve! z can't be everything to everyone. There are alternatives to 'mainframes' both from IBM (POWER) and others. Talking about weakness, as an example I did mention that my x86 also has limitations. I don't need a better machine, especially an overpriced Mac. Just like some customers think they don't need an overpriced z. With today’s mind set of 'good enough' computing i.e. if it doesn't work reboot, if what you have meets your needs, so be it. Just like your iPhone, my smartphone has most probably more power/memory than S370/195 of early 70s. The one I worked on at a sister (can I say this or should it be 'person' organisation of CERN) had a grand total of 1 MB main memory! However, I doubt our smartphones could process tons of data generated by accelerators causing collisions of energetic particles to investigate the structure of the atomic nucleus. Even during the 70's S370/195 did it very successfully i.e., process large amounts of data (strength of the I/O subsystem ??). Yes, there are other suppliers of MF like systems. Someone else on this thread mentioned that they saw a demo of one which had similar RAS capabilities as of z. I am familiar with that system. Great demo that lots customers rushed to introduce these into their IT infrastructure. One customer I know of, who soon after Y2K were encouraged by their 'consultants' to ditch their centralised IBM MF installed lots of these systems at all their distributed sites, 3 systems at each site (Dev/Test/Prod). In case of this customer the hype of the demo turned sour as the systems were more 'down' then 'up'. To overcome the RAS deficiencies the solution was to have 'spare' systems on site. No pun intended, needless to say these systems were sunset and almost 20 years later the customer is still using z. I am sure amongst this list, others will have examples of customers ditching other platforms including z for all sorts of reasons. We can debate for ever which system is better etc. At the end of day just put your money where your mouth is into what best suits your IT needs🙂 BTW: When I started my career during the early 70s, IT didn't exist. It was 'computers' or 'computing'. From: IBM Mainframe Discussion List on behalf of David Crayford Sent: 04 August 2023 00:42 To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives > On 3 Aug 2023, at 2:26 am, P H > <04843e86df79-dmarc-requ...@listserv.ua.edu> wrote: > > The numbers quoted by Tom: > > So I pointed out there's only 12 I/O drawers max on a z16 which is 12 x
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
> The one I worked on at a sister (can I say this or should it be 'person' > organisation of CERN) had a grand total of 1 MB main memory! That sounds more appropriate for a 65 than a 195. > BTW: When I started my career during the early 70s, IT didn't exist. It was > 'computers' or 'computing'. Also 'Data Processing'; I vaguely recall that there were a few more terms. -- Shmuel (Seymour J.) Metz http://mason.gmu.edu/~smetz3 From: IBM Mainframe Discussion List [IBM-MAIN@LISTSERV.UA.EDU] on behalf of P H [04843e86df79-dmarc-requ...@listserv.ua.edu] Sent: Friday, August 4, 2023 5:23 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives In response to your comments and some made by others, my 2 cents worth. This discussion started talking about mainframes and 'split' into sub-threads questioning/focusing on IBM z e.g., z/Architecture, what has z ever done, any uniqueness/special features of z etc. In my response, I tried to answer some Qs and correcting some of the numbers which were being quoted. I did end by saying 'horses for courses'. No single system/platform is perfect. All have their uniqueness, strengths and weakness. Today, other platforms may have similar functions/features as z and some may be even better, the point is during the evolution of z under it's different marketing names (S/360, S/370, S/390, eServer, System z etc) z has evolved, adapted and embraced technologies which businesses require for modern workloads. z continues to evolve! z can't be everything to everyone. There are alternatives to 'mainframes' both from IBM (POWER) and others. Talking about weakness, as an example I did mention that my x86 also has limitations. I don't need a better machine, especially an overpriced Mac. Just like some customers think they don't need an overpriced z. With today’s mind set of 'good enough' computing i.e. if it doesn't work reboot, if what you have meets your needs, so be it. Just like your iPhone, my smartphone has most probably more power/memory than S370/195 of early 70s. The one I worked on at a sister (can I say this or should it be 'person' organisation of CERN) had a grand total of 1 MB main memory! However, I doubt our smartphones could process tons of data generated by accelerators causing collisions of energetic particles to investigate the structure of the atomic nucleus. Even during the 70's S370/195 did it very successfully i.e., process large amounts of data (strength of the I/O subsystem ??). Yes, there are other suppliers of MF like systems. Someone else on this thread mentioned that they saw a demo of one which had similar RAS capabilities as of z. I am familiar with that system. Great demo that lots customers rushed to introduce these into their IT infrastructure. One customer I know of, who soon after Y2K were encouraged by their 'consultants' to ditch their centralised IBM MF installed lots of these systems at all their distributed sites, 3 systems at each site (Dev/Test/Prod). In case of this customer the hype of the demo turned sour as the systems were more 'down' then 'up'. To overcome the RAS deficiencies the solution was to have 'spare' systems on site. No pun intended, needless to say these systems were sunset and almost 20 years later the customer is still using z. I am sure amongst this list, others will have examples of customers ditching other platforms including z for all sorts of reasons. We can debate for ever which system is better etc. At the end of day just put your money where your mouth is into what best suits your IT needs🙂 BTW: When I started my career during the early 70s, IT didn't exist. It was 'computers' or 'computing'. From: IBM Mainframe Discussion List on behalf of David Crayford Sent: 04 August 2023 00:42 To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives > On 3 Aug 2023, at 2:26 am, P H > <04843e86df79-dmarc-requ...@listserv.ua.edu> wrote: > > The numbers quoted by Tom: > > So I pointed out there's only 12 I/O drawers max on a z16 which is 12 x > 16 = 192 slots or 384 ports max. He replied, but didn't seem to fully > accept that answer. > > are 100% correct. These numbers are the MAXIMUM. Depending on the > configuration, these could be a lot less e.g. the number of coupling links > could reduce the numbers. If z16 is ordered with BPA power supplies, the MAX > I/O drawers go down from 12 to 10. > > I have already mentioned things
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
In response to your comments and some made by others, my 2 cents worth. This discussion started talking about mainframes and 'split' into sub-threads questioning/focusing on IBM z e.g., z/Architecture, what has z ever done, any uniqueness/special features of z etc. In my response, I tried to answer some Qs and correcting some of the numbers which were being quoted. I did end by saying 'horses for courses'. No single system/platform is perfect. All have their uniqueness, strengths and weakness. Today, other platforms may have similar functions/features as z and some may be even better, the point is during the evolution of z under it's different marketing names (S/360, S/370, S/390, eServer, System z etc) z has evolved, adapted and embraced technologies which businesses require for modern workloads. z continues to evolve! z can't be everything to everyone. There are alternatives to 'mainframes' both from IBM (POWER) and others. Talking about weakness, as an example I did mention that my x86 also has limitations. I don't need a better machine, especially an overpriced Mac. Just like some customers think they don't need an overpriced z. With today’s mind set of 'good enough' computing i.e. if it doesn't work reboot, if what you have meets your needs, so be it. Just like your iPhone, my smartphone has most probably more power/memory than S370/195 of early 70s. The one I worked on at a sister (can I say this or should it be 'person' organisation of CERN) had a grand total of 1 MB main memory! However, I doubt our smartphones could process tons of data generated by accelerators causing collisions of energetic particles to investigate the structure of the atomic nucleus. Even during the 70's S370/195 did it very successfully i.e., process large amounts of data (strength of the I/O subsystem ??). Yes, there are other suppliers of MF like systems. Someone else on this thread mentioned that they saw a demo of one which had similar RAS capabilities as of z. I am familiar with that system. Great demo that lots customers rushed to introduce these into their IT infrastructure. One customer I know of, who soon after Y2K were encouraged by their 'consultants' to ditch their centralised IBM MF installed lots of these systems at all their distributed sites, 3 systems at each site (Dev/Test/Prod). In case of this customer the hype of the demo turned sour as the systems were more 'down' then 'up'. To overcome the RAS deficiencies the solution was to have 'spare' systems on site. No pun intended, needless to say these systems were sunset and almost 20 years later the customer is still using z. I am sure amongst this list, others will have examples of customers ditching other platforms including z for all sorts of reasons. We can debate for ever which system is better etc. At the end of day just put your money where your mouth is into what best suits your IT needs🙂 BTW: When I started my career during the early 70s, IT didn't exist. It was 'computers' or 'computing'. ____ From: IBM Mainframe Discussion List on behalf of David Crayford Sent: 04 August 2023 00:42 To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives > On 3 Aug 2023, at 2:26 am, P H > <04843e86df79-dmarc-requ...@listserv.ua.edu> wrote: > > The numbers quoted by Tom: > > So I pointed out there's only 12 I/O drawers max on a z16 which is 12 x > 16 = 192 slots or 384 ports max. He replied, but didn't seem to fully > accept that answer. > > are 100% correct. These numbers are the MAXIMUM. Depending on the > configuration, these could be a lot less e.g. the number of coupling links > could reduce the numbers. If z16 is ordered with BPA power supplies, the MAX > I/O drawers go down from 12 to 10. > > I have already mentioned things like cache, memory, I/O Subsystem, on chip > data compression/Crypto (z has been a leader for this)/Sort/AI capabilities. Maybe for crypto but certainly not for AI. My iPhone has a more sophisticated AI engine than the z16. Other platforms have integrated AI engines, AMD ZenDNN, Intel oneDNN etc. Both ship with open source libraries and toolkits sadly lacking for z/OS. I noticed that IBM have shipped patched Python packages for TensorFlow and SnapML that exploit Telum for Linux on Z. I suppose like everything, we’ll have to wait a while for z/OS. Java 11 still does not utilise zEDC compression on z/OS. Talking about compression and crypto, Intel have hardware accelerators as part of QAT, either PCIe cards or on-die. You could argue that the compression tech is better than zEDC as it supports more formats then just gzip. https://www.in
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
> On 4 Aug 2023, at 1:01 pm, Timothy Sipples wrote: > > David Crayford wrote: >> Other platforms have integrated AI engines, AMD ZenDNN, >> Intel oneDNN etc. Both ship with open source libraries and >> toolkits sadly lacking for z/OS. > > Did you miss zDNN? > Nope, I’m aware. Not quite as open source as the other toolkits but it’s good that IBM have embraced Github. > https://github.com/IBM/zDNN > https://www.ibm.com/docs/en/zos/2.5.0?topic=consider-z-deep-neural-network-library-zdnn > >> I noticed that IBM have shipped patched Python packages for >> TensorFlow and SnapML that exploit Telum for Linux on Z. >> I suppose like everything, we’ll have to wait a while for z/OS. > > Missed this one too? > > https://community.ibm.com/community/user/ibmz-and-linuxone/blogs/evan-rivera/2023/02/24/python-ai-toolkit-for-ibm-zos I suspect you are not familiar with the Python AI Toolkit for z/OS. Apart from SciPy, NumPy and some stuff for Jupyter notebooks there is nothing related to AI in that bundle. It’s nothing more than a curated set of Python packages. Please correct me if I’m wrong. > > Quoting from the IBM Redpaper: > > "The Python AI Toolkit for IBM z/OS also benefits from the IBM zSystems > hardware investments that are lower in the stack. Acceleration from the IBM > Integrated Accelerator for AI provides benefits when running AI workloads > that are built on top of the Python AI Toolkit for IBM z/OS. With this > workload execution acceleration, enterprises can meet successfully some of > the most stringent service-level agreements (SLAs) when integrating AI into > business-critical workloads." > > https://www.redbooks.ibm.com/abstracts/redp5709.html Maybe wait until there is actually some tangible AI libraries such as TensorFlow, PyTorch and SnapML before blowing trumpets. > > — > Timothy Sipples > Senior Architect > Digital Assets, Industry Solutions, and Cybersecurity > IBM zSystems/LinuxONE, Asia-Pacific > sipp...@sg.ibm.com > > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
David Crayford wrote: >Other platforms have integrated AI engines, AMD ZenDNN, >Intel oneDNN etc. Both ship with open source libraries and >toolkits sadly lacking for z/OS. Did you miss zDNN? https://github.com/IBM/zDNN https://www.ibm.com/docs/en/zos/2.5.0?topic=consider-z-deep-neural-network-library-zdnn >I noticed that IBM have shipped patched Python packages for >TensorFlow and SnapML that exploit Telum for Linux on Z. >I suppose like everything, we’ll have to wait a while for z/OS. Missed this one too? https://community.ibm.com/community/user/ibmz-and-linuxone/blogs/evan-rivera/2023/02/24/python-ai-toolkit-for-ibm-zos Quoting from the IBM Redpaper: "The Python AI Toolkit for IBM z/OS also benefits from the IBM zSystems hardware investments that are lower in the stack. Acceleration from the IBM Integrated Accelerator for AI provides benefits when running AI workloads that are built on top of the Python AI Toolkit for IBM z/OS. With this workload execution acceleration, enterprises can meet successfully some of the most stringent service-level agreements (SLAs) when integrating AI into business-critical workloads." https://www.redbooks.ibm.com/abstracts/redp5709.html — Timothy Sipples Senior Architect Digital Assets, Industry Solutions, and Cybersecurity IBM zSystems/LinuxONE, Asia-Pacific sipp...@sg.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
> On 3 Aug 2023, at 2:26 am, P H > <04843e86df79-dmarc-requ...@listserv.ua.edu> wrote: > > The numbers quoted by Tom: > > So I pointed out there's only 12 I/O drawers max on a z16 which is 12 x > 16 = 192 slots or 384 ports max. He replied, but didn't seem to fully > accept that answer. > > are 100% correct. These numbers are the MAXIMUM. Depending on the > configuration, these could be a lot less e.g. the number of coupling links > could reduce the numbers. If z16 is ordered with BPA power supplies, the MAX > I/O drawers go down from 12 to 10. > > I have already mentioned things like cache, memory, I/O Subsystem, on chip > data compression/Crypto (z has been a leader for this)/Sort/AI capabilities. Maybe for crypto but certainly not for AI. My iPhone has a more sophisticated AI engine than the z16. Other platforms have integrated AI engines, AMD ZenDNN, Intel oneDNN etc. Both ship with open source libraries and toolkits sadly lacking for z/OS. I noticed that IBM have shipped patched Python packages for TensorFlow and SnapML that exploit Telum for Linux on Z. I suppose like everything, we’ll have to wait a while for z/OS. Java 11 still does not utilise zEDC compression on z/OS. Talking about compression and crypto, Intel have hardware accelerators as part of QAT, either PCIe cards or on-die. You could argue that the compression tech is better than zEDC as it supports more formats then just gzip. https://www.intel.com/content/www/us/en/architecture-and-technology/intel-quick-assist-technology-overview.html > > Talking about the I/O Subsystem, this is a key strength when it comes to > handling large number of I/Os. Unlike x86, the I/O Subsystem handles this > very well and lets the CP get on with what it's mean to do. What no one has > mentioned is the 'processing' power of z. In addition to the main CPs (up to > 200 for z16 Models A01 and L01), the I/O Subsystem has up eo 192 POWER > processors. These are in a N+1 config making a total of 384 in he sub-system > alone. > > Impressive numbers. What do all these prove? Taken out of context, these are > meaningless. As I stated previously, one has to consisder the whole system. > This is where z has strengths. It has a 'balanced system design'. This > morning I decided to do a full virus scan on my 2 year old latop with an > Intel i5 chip. While the scan was running, I couldn't even open a 10 MB > Powerpoint presentation 🙁 (before the smartones give me their 2 cents worth, > I know I could have run the scan as a background task). > Get yourself a better machine. My Mac runs clusters of Linux systems on Kubernetes running stacks like Kafka, ELK at a full pelt without breaking a sweat and I can watch YouTube in 4K at the same time. For a more apples to apples comparison to x86 it would be more interesting to compare a z box to an HP Superdome kitted out with all the fruit. There are only three large systems remaining since Oracle killed off the SPARCs. Z, POWER and the super domes. The server market is dominated by single socket rack servers running distributed systems. > Talking about numbers, the Airbus A380 plane has been designed to have up to > 840 passengers. Are there any airlines with A380s which carry such numbers! > > Horses for courses!! > > ____________ > From: IBM Mainframe Discussion List on behalf of > Tom Brennan > Sent: 02 August 2023 17:34 > To: IBM-MAIN@LISTSERV.UA.EDU > Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The > IBM mainframe: How it runs and why it survives > >> I’ve missed this thread. > > He first said 1536 ports (not slots, not lanes) on a full z16. I asked > where he got that number. Response was there are 12 fanout slots on a > CEC drawer (true), so with 4 CEC drawers that's 48 fanout slots (true) > which means the 4 CEC drawers could address 48 I/O drawers with 16 cards > each and 2 ports per card = 1536 ports. > > So I pointed out there's only 12 I/O drawers max on a z16 which is 12 x > 16 = 192 slots or 384 ports max. He replied, but didn't seem to fully > accept that answer. > > Later he said there are 1600 slots (not ports, not lanes) on a z16 so I > asked where he got that new number. He said he meant 1536 slots (not > ports, not lanes) so the number doubled from last time. I replied same > as I did previously. > > Below, he said 1536 slots again. 1536 cards on a single z16 could be > over 3000 cables! I've had to untangle some 150+ cable rats nests, but > for that one I'd just say, Naw... I'm going home :) > > On 8/2/2023 1:53 AM, David Crayford wrote: >>> On 2 Aug 2023, at 12:15 pm, Tom Brennan wrote: >>&
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
The YouTube is excellent in promoting key strengths of z in a light hearted manner. With numerours z systems on the test floor during development, testing and product and stress testing the patch panel is key to enable 'any to any' configurations. From: IBM Mainframe Discussion List on behalf of Jon Perryman Sent: 03 August 2023 03:56 To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives > On Wednesday, August 2, 2023 at 09:34:34 AM PDT, Tom Brennan wrote: > So I pointed out there's only 12 I/O drawers max on a z16 Sorry Tom and all. I don't recall anyone saying max of 12 I/O drawers otherwise it would have been obvious my number was wrong. Yahoo mail does strange things with tab, backspace, space and other keys. > which is 12 x 16 = 192 slots or 384 ports max. Thanks to youtube, the first IBM z I've seen was the z16 tour at https://youtu.be/ZDtaanCENbc. You say 192 slots or 384 ports. I understand slots being PCIe but was is ports? Is this fiber optic cables or does it somehow split a PCIe slot? > I've had to untangle some 150+ cable rats nests, but> for that one I'd just > say, Naw... I'm going home :) Towards the end of the video, they show the patch panels which are true rats nests. Looks like some of the network rooms I've seen. Some people don't mind dealing with a mess. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
On 8/2/2023 7:56 PM, Jon Perryman wrote: You say 192 slots or 384 ports. Not me, it's IBM doc along with Parwez Hamid, top IBM tech person, redbook author, conference speaker, etc. etc. (retired now from IBM I believe). I understand slots being PCIe but was is ports? Is this fiber optic cables or does it somehow split a PCIe slot? IBM I/O cards almost all come with two plug ports, or SFP's, or whatever you want to call the spot where you plug in a cable. The only cards I can think of that only have one plug port are the 10G OSA cards. So that's why I said 384 ports MAX. And crypto cards have no plug ports. So if a machine has any of those, down goes that 384 max count. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
> On Wednesday, August 2, 2023 at 09:34:34 AM PDT, Tom Brennan wrote: > So I pointed out there's only 12 I/O drawers max on a z16 Sorry Tom and all. I don't recall anyone saying max of 12 I/O drawers otherwise it would have been obvious my number was wrong. Yahoo mail does strange things with tab, backspace, space and other keys. > which is 12 x 16 = 192 slots or 384 ports max. Thanks to youtube, the first IBM z I've seen was the z16 tour at https://youtu.be/ZDtaanCENbc. You say 192 slots or 384 ports. I understand slots being PCIe but was is ports? Is this fiber optic cables or does it somehow split a PCIe slot? > I've had to untangle some 150+ cable rats nests, but> for that one I'd just > say, Naw... I'm going home :) Towards the end of the video, they show the patch panels which are true rats nests. Looks like some of the network rooms I've seen. Some people don't mind dealing with a mess. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
615 passengers on a few planes. https://en.m.wikipedia.org/wiki/Seat_configurations_of_Airbus_A380 On Wed, Aug 2, 2023, 13:26 P H < 04843e86df79-dmarc-requ...@listserv.ua.edu> wrote: > The numbers quoted by Tom: > > So I pointed out there's only 12 I/O drawers max on a z16 which is 12 x > 16 = 192 slots or 384 ports max. He replied, but didn't seem to fully > accept that answer. > > are 100% correct. These numbers are the MAXIMUM. Depending on the > configuration, these could be a lot less e.g. the number of coupling links > could reduce the numbers. If z16 is ordered with BPA power supplies, the > MAX I/O drawers go down from 12 to 10. > > I have already mentioned things like cache, memory, I/O Subsystem, on chip > data compression/Crypto (z has been a leader for this)/Sort/AI capabilities. > > Talking about the I/O Subsystem, this is a key strength when it comes to > handling large number of I/Os. Unlike x86, the I/O Subsystem handles this > very well and lets the CP get on with what it's mean to do. What no one has > mentioned is the 'processing' power of z. In addition to the main CPs (up > to 200 for z16 Models A01 and L01), the I/O Subsystem has up eo 192 POWER > processors. These are in a N+1 config making a total of 384 in he > sub-system alone. > > Impressive numbers. What do all these prove? Taken out of context, these > are meaningless. As I stated previously, one has to consisder the whole > system. This is where z has strengths. It has a 'balanced system design'. > This morning I decided to do a full virus scan on my 2 year old latop with > an Intel i5 chip. While the scan was running, I couldn't even open a 10 MB > Powerpoint presentation 🙁 (before the smartones give me their 2 cents > worth, I know I could have run the scan as a background task). > > Talking about numbers, the Airbus A380 plane has been designed to have up > to 840 passengers. Are there any airlines with A380s which carry such > numbers! > > Horses for courses!! > > ________ > From: IBM Mainframe Discussion List on behalf > of Tom Brennan > Sent: 02 August 2023 17:34 > To: IBM-MAIN@LISTSERV.UA.EDU > Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: > The IBM mainframe: How it runs and why it survives > > > I’ve missed this thread. > > He first said 1536 ports (not slots, not lanes) on a full z16. I asked > where he got that number. Response was there are 12 fanout slots on a > CEC drawer (true), so with 4 CEC drawers that's 48 fanout slots (true) > which means the 4 CEC drawers could address 48 I/O drawers with 16 cards > each and 2 ports per card = 1536 ports. > > So I pointed out there's only 12 I/O drawers max on a z16 which is 12 x > 16 = 192 slots or 384 ports max. He replied, but didn't seem to fully > accept that answer. > > Later he said there are 1600 slots (not ports, not lanes) on a z16 so I > asked where he got that new number. He said he meant 1536 slots (not > ports, not lanes) so the number doubled from last time. I replied same > as I did previously. > > Below, he said 1536 slots again. 1536 cards on a single z16 could be > over 3000 cables! I've had to untangle some 150+ cable rats nests, but > for that one I'd just say, Naw... I'm going home :) > > On 8/2/2023 1:53 AM, David Crayford wrote: > >> On 2 Aug 2023, at 12:15 pm, Tom Brennan > wrote: > >> > >>> The IBM z16 can have up to 1,536 PCIe+ slots > >> > >> I'm gonna quit explaining this and just say, "WRONG" every time you say > this as if it's a fact :) > > > > I’ve missed this thread. By 1,536 PCIe slots, that’s slots not lanes > right? Even if it were lanes that would be a ludicrous suggestions! That’s > so far fetched it’s laughable. The Redbook [1] is quite clear about I/O > configurations. What I find interesting is that the z16 seems to use PCIe > gen 3 and not gen 4 which doubles the transfer rate per lane. There must be > a good technical reason for this. > > > > [1] https://www.redbooks.ibm.com/redbooks/pdfs/sg248951.pdf > > > >> > >> On 8/1/2023 8:01 PM, Jon Perryman wrote: > >>> > On Tuesday, August 1, 2023 at 05:20:33 PM PDT, David Crayford < > dcrayf...@gmail.com> wrote: > >>>> What’s the difference between between channelized I/O and a rack of > >>>> x86 servers connected to a SAN using fibre channel driven by high > speed HBAs? > >>> PCIe was created specifically for PCs and IBM z16 chose to use that as > their only channel technology. Channelized I/O for PC has been available > for several
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
The numbers quoted by Tom: So I pointed out there's only 12 I/O drawers max on a z16 which is 12 x 16 = 192 slots or 384 ports max. He replied, but didn't seem to fully accept that answer. are 100% correct. These numbers are the MAXIMUM. Depending on the configuration, these could be a lot less e.g. the number of coupling links could reduce the numbers. If z16 is ordered with BPA power supplies, the MAX I/O drawers go down from 12 to 10. I have already mentioned things like cache, memory, I/O Subsystem, on chip data compression/Crypto (z has been a leader for this)/Sort/AI capabilities. Talking about the I/O Subsystem, this is a key strength when it comes to handling large number of I/Os. Unlike x86, the I/O Subsystem handles this very well and lets the CP get on with what it's mean to do. What no one has mentioned is the 'processing' power of z. In addition to the main CPs (up to 200 for z16 Models A01 and L01), the I/O Subsystem has up eo 192 POWER processors. These are in a N+1 config making a total of 384 in he sub-system alone. Impressive numbers. What do all these prove? Taken out of context, these are meaningless. As I stated previously, one has to consisder the whole system. This is where z has strengths. It has a 'balanced system design'. This morning I decided to do a full virus scan on my 2 year old latop with an Intel i5 chip. While the scan was running, I couldn't even open a 10 MB Powerpoint presentation 🙁 (before the smartones give me their 2 cents worth, I know I could have run the scan as a background task). Talking about numbers, the Airbus A380 plane has been designed to have up to 840 passengers. Are there any airlines with A380s which carry such numbers! Horses for courses!! From: IBM Mainframe Discussion List on behalf of Tom Brennan Sent: 02 August 2023 17:34 To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Channelized I/O WAS: Mainframe Makers WAS: Ars Technica: The IBM mainframe: How it runs and why it survives > I’ve missed this thread. He first said 1536 ports (not slots, not lanes) on a full z16. I asked where he got that number. Response was there are 12 fanout slots on a CEC drawer (true), so with 4 CEC drawers that's 48 fanout slots (true) which means the 4 CEC drawers could address 48 I/O drawers with 16 cards each and 2 ports per card = 1536 ports. So I pointed out there's only 12 I/O drawers max on a z16 which is 12 x 16 = 192 slots or 384 ports max. He replied, but didn't seem to fully accept that answer. Later he said there are 1600 slots (not ports, not lanes) on a z16 so I asked where he got that new number. He said he meant 1536 slots (not ports, not lanes) so the number doubled from last time. I replied same as I did previously. Below, he said 1536 slots again. 1536 cards on a single z16 could be over 3000 cables! I've had to untangle some 150+ cable rats nests, but for that one I'd just say, Naw... I'm going home :) On 8/2/2023 1:53 AM, David Crayford wrote: >> On 2 Aug 2023, at 12:15 pm, Tom Brennan wrote: >> >>> The IBM z16 can have up to 1,536 PCIe+ slots >> >> I'm gonna quit explaining this and just say, "WRONG" every time you say this >> as if it's a fact :) > > I’ve missed this thread. By 1,536 PCIe slots, that’s slots not lanes right? > Even if it were lanes that would be a ludicrous suggestions! That’s so far > fetched it’s laughable. The Redbook [1] is quite clear about I/O > configurations. What I find interesting is that the z16 seems to use PCIe gen > 3 and not gen 4 which doubles the transfer rate per lane. There must be a > good technical reason for this. > > [1] https://www.redbooks.ibm.com/redbooks/pdfs/sg248951.pdf > >> >> On 8/1/2023 8:01 PM, Jon Perryman wrote: >>> > On Tuesday, August 1, 2023 at 05:20:33 PM PDT, David Crayford >>> wrote: >>>> What’s the difference between between channelized I/O and a rack of >>>> x86 servers connected to a SAN using fibre channel driven by high speed >>>> HBAs? >>> PCIe was created specifically for PCs and IBM z16 chose to use that as >>> their only channel technology. Channelized I/O for PC has been available >>> for several decades and is not limited to PCIe. The IBM z16 can have up to >>> 1,536 PCIe+ slots. >>> As for x86 fiber channel connection to a PC, PCIe is only one possibility. >>> -- >>> For IBM-MAIN subscribe / signoff / archive access instructions, >>> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN >> >> -- >> For IBM-MAIN subscribe / si
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
> I’ve missed this thread. He first said 1536 ports (not slots, not lanes) on a full z16. I asked where he got that number. Response was there are 12 fanout slots on a CEC drawer (true), so with 4 CEC drawers that's 48 fanout slots (true) which means the 4 CEC drawers could address 48 I/O drawers with 16 cards each and 2 ports per card = 1536 ports. So I pointed out there's only 12 I/O drawers max on a z16 which is 12 x 16 = 192 slots or 384 ports max. He replied, but didn't seem to fully accept that answer. Later he said there are 1600 slots (not ports, not lanes) on a z16 so I asked where he got that new number. He said he meant 1536 slots (not ports, not lanes) so the number doubled from last time. I replied same as I did previously. Below, he said 1536 slots again. 1536 cards on a single z16 could be over 3000 cables! I've had to untangle some 150+ cable rats nests, but for that one I'd just say, Naw... I'm going home :) On 8/2/2023 1:53 AM, David Crayford wrote: On 2 Aug 2023, at 12:15 pm, Tom Brennan wrote: The IBM z16 can have up to 1,536 PCIe+ slots I'm gonna quit explaining this and just say, "WRONG" every time you say this as if it's a fact :) I’ve missed this thread. By 1,536 PCIe slots, that’s slots not lanes right? Even if it were lanes that would be a ludicrous suggestions! That’s so far fetched it’s laughable. The Redbook [1] is quite clear about I/O configurations. What I find interesting is that the z16 seems to use PCIe gen 3 and not gen 4 which doubles the transfer rate per lane. There must be a good technical reason for this. [1] https://www.redbooks.ibm.com/redbooks/pdfs/sg248951.pdf On 8/1/2023 8:01 PM, Jon Perryman wrote: > On Tuesday, August 1, 2023 at 05:20:33 PM PDT, David Crayford wrote: What’s the difference between between channelized I/O and a rack of x86 servers connected to a SAN using fibre channel driven by high speed HBAs? PCIe was created specifically for PCs and IBM z16 chose to use that as their only channel technology. Channelized I/O for PC has been available for several decades and is not limited to PCIe. The IBM z16 can have up to 1,536 PCIe+ slots. As for x86 fiber channel connection to a PC, PCIe is only one possibility. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
At the risk of being "WRONG" ;¬)) several times, I offer the following. The Processor Units (GPs, CPU, etc.) are PCIe Gen 4, but the 16 slots in the I/O drawer hold Gen 3 cards, up to 16 of them at 16GBps. Each card can support a max of 32 lanes which can be multiplexed. The max theoretical transmission per lane with encoding is 984.6 MBps, and per link is 15.75 GBpsm. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
> On 2 Aug 2023, at 12:15 pm, Tom Brennan wrote: > > > The IBM z16 can have up to 1,536 PCIe+ slots > > I'm gonna quit explaining this and just say, "WRONG" every time you say this > as if it's a fact :) I’ve missed this thread. By 1,536 PCIe slots, that’s slots not lanes right? Even if it were lanes that would be a ludicrous suggestions! That’s so far fetched it’s laughable. The Redbook [1] is quite clear about I/O configurations. What I find interesting is that the z16 seems to use PCIe gen 3 and not gen 4 which doubles the transfer rate per lane. There must be a good technical reason for this. [1] https://www.redbooks.ibm.com/redbooks/pdfs/sg248951.pdf > > On 8/1/2023 8:01 PM, Jon Perryman wrote: >> > On Tuesday, August 1, 2023 at 05:20:33 PM PDT, David Crayford >> wrote: >>> What’s the difference between between channelized I/O and a rack of >>> x86 servers connected to a SAN using fibre channel driven by high speed >>> HBAs? >> PCIe was created specifically for PCs and IBM z16 chose to use that as their >> only channel technology. Channelized I/O for PC has been available for >> several decades and is not limited to PCIe. The IBM z16 can have up to 1,536 >> PCIe+ slots. >> As for x86 fiber channel connection to a PC, PCIe is only one possibility. >> -- >> For IBM-MAIN subscribe / signoff / archive access instructions, >> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
> The IBM z16 can have up to 1,536 PCIe+ slots I'm gonna quit explaining this and just say, "WRONG" every time you say this as if it's a fact :) On 8/1/2023 8:01 PM, Jon Perryman wrote: > On Tuesday, August 1, 2023 at 05:20:33 PM PDT, David Crayford wrote: What’s the difference between between channelized I/O and a rack of x86 servers connected to a SAN using fibre channel driven by high speed HBAs? PCIe was created specifically for PCs and IBM z16 chose to use that as their only channel technology. Channelized I/O for PC has been available for several decades and is not limited to PCIe. The IBM z16 can have up to 1,536 PCIe+ slots. As for x86 fiber channel connection to a PC, PCIe is only one possibility. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Channelized I/O WAS: Mainframe Makers.... WAS: Ars Technica: The IBM mainframe: How it runs and why it survives
> On Tuesday, August 1, 2023 at 05:20:33 PM PDT, David Crayford > wrote: > What’s the difference between between channelized I/O and a rack of > x86 servers connected to a SAN using fibre channel driven by high speed HBAs? PCIe was created specifically for PCs and IBM z16 chose to use that as their only channel technology. Channelized I/O for PC has been available for several decades and is not limited to PCIe. The IBM z16 can have up to 1,536 PCIe+ slots. As for x86 fiber channel connection to a PC, PCIe is only one possibility. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN