Re: [Beowulf] Your thoughts on the latest RHEL drama?

[Gerald Henriksen]

On Mon, 26 Jun 2023 14:27:23 -0400, you wrote:

>By now, most of you should have heard about Red Hat's latest to 
>eliminate any competition to RHEL. If not, here's some links:

I think it is safer to say IBM's efforts.

>3. After RH starting contributing funding to GNOME development, the next 
>major version of RHEL didn't install other desktops during the install. 
>I remember RHEL saying this was a bug, but I've always suspected it was 
>a deliberate act to reduce KDE market share and and give RH another area 
>of the Linux ecosystem it could control.

I think that is unlikely given how unsuitable (at least in the Linux
world) long term stable distributions like RHEL are to the desktop
environment.

I always attributed that stuff to internal politics with the internal
Gnome team doing things to preserve their existence, particular when
they went different with Gnome 3.

>4. RH takes over control of CentOS, which at the time was the only 
>competitor to RHEL.

It was more a case of Red Hat rescuing CentOS.

At the time the CentOS project was in trouble as they struggled and
failed to bring out their versions in anything like a timely manner
after the RHEL release.  I suspect the IBM version of Red Hat would
have just let CentOS fail.

>Not long after, RHEL eliminates CentOS as a competitor by 
>changing it to "CentOS  Stream" so it's no longer a competitor to RHEL. 

No.

CentOS as part of Red Hat lasted almost 7 years (taken over by Red Hat
in January 2014, killed in December 2020)

Guesswork, but if IBM doesn't buy Red Hat it's possible CentOS still
exists.

>CentOS Stream is now a development version of sorts for RHEL, but I 
>thought that was exactly what Fedora was for.

Recent discussion on Fedora mailing list has it as major version of
RHEL split off from Fedora but CentOS Stream is used for minor
versions.

How long this remains to be true is debatable.

Red Hat/IBM recently eliminated a paid position dealing with Fedora,
they have killed off LibreOffice in RHEL (by not replacing an
employee) and thus its future in Fedora is dependent on new
maintainers stepping up.  Red Hat/IBM has also been looking at
stretching the Fedora rules for their OpenJDK support in Fedora with a
unsaid threat of we could stop providing OpenJDK packages in Fedora.

Also Fedora went to BTRFS as a default file system even though Red Hat
stopped supporting it.

>With RH (and IBM?) so focused on market dominance/profits, it's not a 
>stretch to think they they'll eventually "say no" to supporting anything 
>other than x86 and POWER processors, since the other processors don't 
>have enough market share to make it profitable, or compete with IBM's 
>offerings.

ARM has a reasonable presence in the cloud providers and probably has
a bigger market share than POWER at this point.

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] Power Cycling Question

[Gerald Henriksen]

On Fri, 16 Jul 2021 15:35:11 -0400, you wrote:

>Reducing power use has become an important topic. One
>of the questions I always wondered about is
>why more cluster do not turn off unused nodes. Slurm
>has hooks to turn nodes off when not in use and
>turn them on when resources are needed.

Given the expense of a cluster (purchase, running, space allocation,
etc) perhaps that is the wrong question.

If you have enough spare capacity that turning off nodes would create
a noticable power saving, then maybe you should be looking at why you
have that capacity?

The goal really should be more about keeping all the nodes busy, so is
there something else that can be done to make sure that the nodes have
work to be done?

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] AMD and AVX512 [EXT]

[Gerald Henriksen]

On Sun, 20 Jun 2021 06:51:58 +0100, you wrote:

>That is a very interesting point! I never thought of that.
>Also mobile drives ARM development - yes I know the CPUs in Isambard and
>Fugaku will not be seen in your mobile phone but the ecosystem is propped
>up by having a diverse market and also the power saving priorities of
>mobile will influence HPC ARM CPUs.

I think the danger is in thinking of ARM (or going forward RISC-V) in
the same way that we have traditionally considered CPU families like
the x86 / x64 / Power families.

One of things hobbling x64 is that is effectively 1 design that Intel
(and to a lesser extent AMD) try to fit into multiple roles - often
without success.  Consider the now abandoned attempts to get Intel
chips into phones and tablets.

ARM has no such contraints - they are quite happy to develop new
designs for specific markets that are entirely unsuitable for their
existing strengths.

Hence, as part of the ARM push into HPC, the new Neoverse V1 - a
design for HPC that probably won't appear in phones.

https://www.arm.com/blogs/blueprint/neoverse-v1

Or consider that the ARM ecosystem has shunned making multiple-bitness
CPUs/SOCs - they essentially made a clean break with 64-bit only chips
that sit alongside the 32-bit only chips - vendors choose the hardware
for their needs and don't carry along legacy stuff that eats up
silicon space and power.

ARM is about taking ARM IP and creating custom designs for specific
markets.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] AMD and AVX512

[Gerald Henriksen]

On Wed, 16 Jun 2021 13:15:40 -0400, you wrote:

>The answer given, and I'm 
>not making this up, is that AMD listens to their users and gives the 
>users what they want, and right now they're not hearing any demand for 
>AVX512.
>
>Personally, I call BS on that one. I can't imagine anyone in the HPC 
>community saying "we'd like processors that offer only 1/2 the floating 
>point performance of Intel processors".

I suspect that is marketing speak, which roughly translates to not
that no one has asked for it, but rather requests haven't reached a
threshold where the requests are viewed as significant enough.

> Sure, AMD can offer more cores, 
>but with only AVX2, you'd need twice as many cores as Intel processors, 
>all other things being equal.

But of course all other things aren't equal.

AVX512 is a mess.

Look at the Wikipedia page(*) and note that AVX512 means different
things depending on the processor implementing it.

So what does the poor software developer target?

Or that it can for heat reasons cause CPU frequency reductions,
meaning real world performance may not match theoritical - thus easier
to just go with GPU's.

The result is that most of the world is quite happily (at least for
now) ignoring AVX512 and going with GPU's as necessary - particularly
given the convenient libraries that Nvidia offers.

> I compared a server with dual AMD EPYC >7H12 processors (128)
> quad Intel Xeon 8268 >processors (96 cores).

> From what I've heard, the AMD processors run much hotter than the Intel 
>processors, too, so I imagine a FLOPS/Watt comparison would be even less 
>favorable to AMD.

Spec sheets would indicate AMD runs hotter, but then again you
benchmarked twice as many Intel processors.

So, per spec sheets for you processors above:

AMD - 280W - 2 processors means system 560W
Intel - 205W - 4 processors means system 820W

(and then you also need to factor in purchase price).

>An argument can be made that for calculations that lend themselves to 
>vectorization should be done on GPUs, instead of the main processors but 
>the last time I checked, GPU jobs are still memory is limited, and 
>moving data in and out of GPU memory can still take time, so I can see 
>situations where for large amounts of data using CPUs would be preferred 
>over GPUs.

AMD's latest chips support PCI 4 while Intel is still stuck on PCI 3,
which may or may not mean a difference.

But what despite all of the above and the other replies, it is AMD who
has been winning the HPC contracts of late, not Intel.

* - https://en.wikipedia.org/wiki/Advanced_Vector_Extensions
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] [External] RIP CentOS 8 [EXT]

[Gerald Henriksen]

On Thu, 10 Dec 2020 13:05:25 +, you wrote:

>A fork is something im thinking about doing in all fairness. Hoping to start 
>soon on it. Need to at this point figure out how to clone the repositories and 
>start my own  testing etc.

Not trying to discourage you, but doing a Linux fork regardless of the
upstream is a lot of work which is why few of the forks survive for
long.

When Red Hat took over CentOS a couple of people started up another
fork, but it never actually got to shipping before they gave up given
the effort involved (can't recall the name they called themselves, and
a quick search doesn't reveal anything)
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] First cluster in 20 years - questions about today

[Gerald Henriksen]

On Tue, 4 Feb 2020 21:27:51 -0500, you wrote:

>Assuming my work and writing is acceptable quality, how likely will I be to
>get published with just a master degree?

Can't answer that, but my understanding is that publishing in academic
style journals costs money so that may also be a consideration for you
even if you create something of interest and can work past the
education/lack of institution.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] First cluster in 20 years - questions about today

[Gerald Henriksen]

On Tue, 4 Feb 2020 21:27:51 -0500, you wrote:

>I'm going to start by trying to learn abinit.  They have experimental, CUDA
>only, GPU support, so I may save up for some used nVidia cards at some
>point, maybe I can find a deal on P106 class cards.

Before looking for used Nvidia cards check your requirements first.
During the mining craze a bunch of cards were released targeted at the
miners, and they came with only PCI Express 1.0 with very limited
bandwidth (compared to the standard PCI Express 3.0).

The mining cards can be fine if your data can fit into the card's
memory and it just sit's there being dealt with, but if you need to
transfer data frequently it could be a problem.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] First cluster in 20 years - questions about today

[Gerald Henriksen]

On Sun, 2 Feb 2020 21:51:21 -0500, you wrote:

>I'm testing FreeBSD right now and will test *SunOS before commiting a full
>deployment.
>
>My initial 3 nodes are socket 940 Opteron based.  I'm planning to acquire
>two nodes of dual cpu hexacore Xeons for the cost of a small road trip next
>weekend (otherwise free).
>
>The more I think about it, maybe I'll have two clusters, one for Opterons
>and one for Xeons.
>
>I'm currently out of work on medical leave and don't have a lot of free
>capital.

Fair enough, without a budget then you need to use what you either
have on hand or can get for free regardless of how slow it may be.

So the question then becomes is this something so you have a challenge
to take up your time all day, or do you actually want to get some
research done?

If it's for a challenge, then by all means try and get Solaris or even
one of the *bsd's working with the appropriate software - it will
likely give you weeks if not months of challenges.

But if you actually want to do research then as others have said then
you really need to choose a version of Linux and move on to doing the
research.

Similarly, why the insistence on ZFS?  Without a budget you are
unlikely to have the storage capacity where ZFS typically kicks in as
being appropriate.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] First cluster in 20 years - questions about today

[Gerald Henriksen]

On Sat, 1 Feb 2020 22:21:09 -0500, you wrote:

>Should I consider Solaris or illumos?  I do plan on using ZFS, especially
>for the data node, but I want as much redundancy as I can get, since I'm
>going to be using used hardware.  Will the fancy Solaris cluster tools be
>useful?

Unless you are absolutely sure the software you want to run works on
Solaris, using Solaris/Illumos is likely asking for trouble.

You can get ZFS on FreeBSD, or if you are willing to overlook the
license issues my understanding Ubuntu works quite well with ZFS.

As for hardware, what do you mean by "used"?

Thanks to AMD the performance of newer hardware is signficantly better
than something 5+ years old, and relatively speaking not necessarily
that expensive, with added benefits of less power required than
anything 10+ years old.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

[Beowulf] Immersive Cooling

[Gerald Henriksen]

Given that is has come up in the past, a podcast and video about a
supercomputer that uses liquid cooling

https://insidehpc.com/2019/09/podcast-extreme-power-and-cooling-efficiency-at-downunder-geosolutions/

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

[Beowulf] arXiv - Securing HPC using Federated Authentication

[Gerald Henriksen]

Perhaps of interest to some

https://arxiv.org/abs/1908.07573
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] Lustre on google cloud

[Gerald Henriksen]

On Wed, 31 Jul 2019 04:10:12 +, you wrote:

>They now have Lustre through FSx or what ever AWS have called it. I am not 
>sure you guys have heard about the capital one data breach but at times im 
>still rather weary of the cloud.

Not sure what the Capital One data breach has to do with the cloud, it
was (yet again?) misconfigured software that allowed the theft.

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] flatpack

[Gerald Henriksen]

On Mon, 22 Jul 2019 22:47:30 -0700, you wrote:

>On 22/7/19 10:40 pm, Jonathan Aquilina wrote:
>
>> So in a nut shell this is taking dockerization/ containerization and
>> making it more for the every day Linux user instead of the HPC user?
>
>I don't think this goes as far as containers with isolation, as I think 
>that's not what they're trying to do. But it does seem they're thinking 
>along those lines.

Flatpack is aimed at the desktop, and as it requires assorted desktop
technologies isn't meant to work on servers (which is in some ways
unfortunate).

As it is meant for desktop apps some of the isolation goals of
something like Docker don't work so well, but the intent is to try and
make things "safer" than just installing a binary and running it.

>> It would be interesting to have a distro built around such a setup.
>
>I think this is targeting cross-distro applications.  With all the 
>duplication of libraries, etc, a distro using it would be quite bulky.

While it is cross-distro, there is a project from Fedora to build a
desktop distribution around it called Silverblue

https://silverblue.fedoraproject.org/
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

[Beowulf] Nvidia bring CUDA to ARM

[Gerald Henriksen]

By the end of the year Nvidia will have CUDA and the associated AI / HPC
stuff available for ARM

https://nvidianews.nvidia.com/news/nvidia-brings-cuda-to-arm-enabling-new-path-to-exascale-supercomputing
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] Containers in HPC

[Gerald Henriksen]

On Fri, 24 May 2019 03:41:11 +, you wrote:

>You mention to move data to storage how is fedora's gnome desktop edition 
>going to achieve that? Wont one need to use some sort of block storage on aws, 
>google cloud, azure or host your own setup in house?

No idea, given that I don't like Gnome these days I haven't followed
closely and Silverblue, like a bunch of other Fedora ideas, seems to
be a project that is taking a while to become usuable.

My superficial understanding is that the desktop version of containers
will allow access to a sandboxed version of the filesystem.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] Containers in HPC

[Gerald Henriksen]

On Thu, 23 May 2019 12:35:13 +, you wrote:

>Thanks for the great explanation and clarification. Another question that 
>stems from the below what mechanisms exist in terms of security for the 
>containers to be as secure as a VM?

I know there have been security concerns about Docker (what most
people think of when they talk about containers these days), though I
am not sure what exactly they are.

They obviously won't be as a secure as a VM as they are sharing the
underlying kernel and perhaps a few system libraries, so if a
different container somehow finds a way to compromise the kernel
(maybe not so theoritical in the current Intel era) then there will be
the possiblity of at least getting at any system calls any other
containers make to the kernel.

And at least Docker containers also have the issue that they typically
don't have permanent storage so you need to move any data you want to
keep out of the container prior to killing the container.

Despite that they have a lot of advantages, and for example Fedora has
a project to create a new version of their Gnome Desktop edition using
containers instead of traditional rpm packages called Silverblue, and
this is partly due to the containers additional security over a
traditionally installed application (for example, the ability to
restrict access to the underlying filesystem).



___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

[Beowulf] Containers in HPC

[Gerald Henriksen]

Paper on arXiv that may be of interest to some as it may be where HPC
is heading even for private clusters:

Evalutation of Docker Containers for Scientific Workloads in the Cloud
https://arxiv.org/abs/1905.08415
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] HPE to acquire Cray

[Gerald Henriksen]

On Mon, 20 May 2019 18:42:31 -0400, you wrote:

>I am curious, do you have some evidence for the demise of CentOS
>other than IBM bought RH?

The important thing to remember about CentOS (and presumably why Red
Hat brought it on board) is that it is not really a RHEL competitor,
rather it is a RHEL marketing tool.

Red Hat, for RHEL, needs to worry about Ubuntu who (as far as I know)
make no difference between the LTS release publicly available to all
and the version they offer paid support for.

Red Hat needs a similiar arrangement, and CentOS does that for them.
They need the "free" version of the OS that start ups and companies
can use legally while building a project / business and then when and
if they decide they want paid support then RHEL is the obvious choice.

If there is no CentOS then that market defaults to Ubuntu, and if you
are running Ubuntu and you now need the insurance that paid support
offers you are unlikely to suddenly change your infrastructure to
RHEL.

My observation / opinion is that IBM isn't the real threat to Red Hat,
rather Red Hat is the biggest threat to Red Hat.

Like any organization that has been around for a while various
"kingdoms" have been established within Red Hat, and those combined
with inertia mean Red Hat hasn't been making the changes or tough
decisions that I suspect the market is moving towards.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] HPE to acquire Cray

[Gerald Henriksen]

On Fri, 17 May 2019 10:10:16 -0400, you wrote:

>On Fri, May 17, 2019 at 10:01 AM Jonathan Aquilina
> wrote:
>>
>> Redhat and IBM im worried as I use Centos big time and then the only thing I 
>> can think of is forking fedora and roll a rolling distro if they decide to 
>> pull the plug. But we have to wait and see.
>
>in regards to centos, that's another ugh moment.  i was looking to see
>when centos was going to drop v8.  to me the the one wiki page they
>put up reads like a big whine about how much work there is and people
>should stop asking.

This happens every time, people expect CentOS to ship at the same time
as RHEL and it has always been a couple of months later is my
recollection, perhaps with at least one case being longer.

While being part of Red Hat helps with some things it is still a large
amount of work, and the move to Fedora infrastructure will be causing
some new issues.

As for the Red Hat / IBM issue, time will tell I guess.  Anything
major though would likely mean a bunch of Red Hat employees suddenly
being available and so it might be worth taking a wait and see
approach.

>and intel just released a linux distro (i didn't read into it much),
>so i wonder if openhpc might move to that and drop centos.  in my view
>centos's future is uncertain, which makes me unhappy :(

Clear Linux has been around for at least a couple of years, the
biggest (and obvious) issue is that it is aimed at Intel processors
and anyone using AMD is likely to have various issues (not to mention
the unlikelyhood of having Clear Linux run on ARM or Power for those
who that matters).

And with yet another Intel security issue released this week that is
causing performance issues jumping onto an Intel focused distribution
may not be an action to take without a lot of thought.

Unless someone (with a realistic plan for the manpower and
infrastructure) forks Fedora / CentOS then realistically I suspect
that most valid althernative to Fedora / CentOS / Red Hat is Debian
(unless one really needs corporate support) given the either Ubuntu or
openSUSE also have the risk of a takeover.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] Frontier Announcement

[Gerald Henriksen]

On Wed, 8 May 2019 14:13:51 -0400, you wrote:

>On Wed, May 8, 2019 at 1:47 PM Jörg Saßmannshausen <
>sassy-w...@sassy.formativ.net> wrote:
>>
>Once upon a time portability, interoperabiilty, standardization, were
>considered good software and hardware attributes.
>Whatever happened to them?

I suspect in a lot of cases they were more ideals and goals than
actual things.

Just look at the struggles the various BSDs have in getting a lot of
software running given the inherent Linuxisms that seem to happen.

In the case of what is relevant to this discussion, CUDA, Nvidia saw
an opportunity (and perhaps also reacted to the threat of not having
their own CPU to counter the integrated GPU market) and invested
heavily into making their GPUs more than simply a 3D graphics device.

As Nvidia built up the libraries and other software to make life
easier for programmers to get the most out of Nvidia hardware AMD and
Intel ignored the threat until it was too late, and partial attempts
at open standards struggled.

And programmers, given struggling with OpenCL or other options vs
going with CUDA with its tools and libraries, went for what gave them
the best performance and easiest implementation (aka a win/win).

Of course then ML came along and suddenly AMD and Intel couldn't
ignore the market anymore, but they are both struggling from a distant
2nd place to try and replicate the CUDA ecosystem...
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] LFortran ... a REPL/Compiler for Fortran

[Gerald Henriksen]

On Mon, 25 Mar 2019 06:07:00 -0700, you wrote:

>Hmm, how does this compare to Flang
>?
>
>On Sun, Mar 24, 2019 at 12:33 PM Joe Landman  wrote:
>
>> See https://docs.lfortran.org/ .   Figured Jeff Layton would like this :D

It appears based on the limited information on the website lfortran is
in its early stages.  It appears to have as a goal the ability to
execute code interactively, and as a negative is creating its own
extension to the Fortran Standard.  Also seems to require Java.

Flang on the other hand is really at this point 2 projects, both
supported/developed by Nvidia.

Flang is the current compiler that is either complete or reasonably
complete from a Fortran standpoint but requires various things to work
(a custom version of clang for example I believe).

Also included under the Flang umbrella is f18, a new Fortran compiler
being written from scratch with the goal of it being a proper member
of the LLVM community, and they have interacted with the existing
LLVM/clang community to find the best way to proceed.

In both cases (Flang/f18) they are traditional style compilers that
don't offer an interactive mode I believe.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] Introduction and question

[Gerald Henriksen]

On Thu, 21 Mar 2019 16:31:38 -0400, you wrote:

>Thanks for sharing this. I was recently asked for my input in a job 
>description for a new position. They wanted to make the education 
>requirements a minimum of a BS in Math, Physics, Engineering, or CS. I 
>recommended that they DO NOT list any education requirements for this 
>position, because most of the skills they were looking for (git, make 
>files, GNU autoconf, CMake, etc.), are not taught in any college 
>curriculum I know of, so a formal education is no guarantee of those 
>skills. and some of the best sys admins and programmers I ever met  had 
>no formal education in STEM, or at all.
>
>I was overruled.

Certainly all true, but I suspect you will find in almost anying
hiring event these days the goal isn't necessarily to find the best
person for the position, but to find someone close enough with the
minimal effort.

Thus things like requiring certain degrees or certifications is an
easy way to make the effort easier as you can simply have HR/the
software being used filter out all those people applying who don't
meet your keywords...
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] Large amounts of data to store and process

[Gerald Henriksen]

On Fri, 15 Mar 2019 05:28:42 +, you wrote:

>I think what I was getting at is why not include the current HPC practices to 
>every day desktops in the sense since we are reaching certain limits and have 
>to write code to take advantage of more and more cores. Why not use MPI and 
>the like to help distribute the software side of things to the cores?

I suspect the simple answer is a combination of most of the software
running on desktops doesn't really need all those extra cores we are
now getting, and some of the more common desktop applications don't
really lend themselves to parallel processing anyway.

It's great that you can go out and get 32 core Threadripper and sit it
on your desk for an affordable price, but for 90% or more of the
market at least 28 of those cores would spend most of their life idle.



___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
https://beowulf.org/cgi-bin/mailman/listinfo/beowulf

Re: [Beowulf] 2 starting questions on how I should proceed for a correct first micro-cluster (2-nodes) building

[Gerald Henriksen]

On Sun, 3 Mar 2019 09:28:03 +, you wrote:

>ps. If you are interested in parallelism...  there is Julia.
>https://docs.julialang.org/en/v1/manual/parallel-computing/index.html

Also Chapel https://chapel-lang.org/
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] HPC workflows

[Gerald Henriksen]

On Fri, 7 Dec 2018 16:19:30 +0100, you wrote:

>Perhaps for another thread:
>Actually I went t the AWS USer Group in the UK on Wednesday. Ver
>impressive, and there are the new Lustre filesystems and MPI networking.
>I guess the HPC World will see the same philosophy of building your setup
>using the AWS toolkit as Uber etc. etc. do today.
>Also a lot of noise is being made at the moment about the convergence of
>HPC and Machine Learning workloads.
>Are we going to see the MAchine Learning folks adapting their workflows to
>run on HPC on-premise bare metal clusters?
>Or are we going to see them go off and use AWS (Azure, Google ?)

I suspect that ML will not go for on-premise for a number of reasons.

First, ignoring cost, companies like Google, Amazon and Microsoft are
very good at ML because not only are they driving the research but
they need it for their business.  So they have the in house expertise
not only to implement cloud systems that are ideal for ML, but to
implement custom hardware - see Google's Tensor Processor Unit.

Second, setting up a new cluster isn't going to be easy.  Finding
physical space, making sure enough utilities can be supplied to
support the hardware, staffing up, etc.  are not only going to be
difficult but inherently takes time when instead you can simply sign
up to a cloud provider and have the project running within 24 hours.
Would HPC exist today as we know it if the ability to instantly turn
on a cluster existed at the beginning?

Third, albeit this is very speculative.  I suspect ML learning is
heading towards using custom hardware.  It has had a very good run
using GPU's, and a GPU will likely always be the entry point for
desktop ML, but unless Nvidia is holding back due to a lack of
competition is does appear the GPU is reaching and end to its
development much like CPUs have.  The latest hardware from Nvidia is
getting lacklustre reviews, and the bolting on of additional things
like raytracing is perhaps an indication that there are limits to how
much further the GPU architecture can be pushed.  The question then is
the ML market big enough to have that custom hardware as a OEM product
like a GPU or will it remain restricted to places like Google who can
afford to build it without the necessary overheads of a consumer
product.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] HPC workflows

[Gerald Henriksen]

On Wed, 5 Dec 2018 09:35:07 -0800, you wrote:

>Certainly the inability of distros to find the person-hours to package
>everything plays a role as well, your cause and effect chain there is
>pretty accurate. Where I begin to branch is at the idea of software that is
>unable to be packaged in an rpm/deb.

In some convenient timing, the following was posted by overtmind on
Reddit discussing why Atom hasn't been packaged for Fedora(*):

---
"This means, for every nodejs dependency Electron needs - and there
are a metric #$%# ton - since you can't use npm as an
installer/package manager - you need to also package all of those and
make sure they're in fedora and up-to-date, and then you also need to
package all of the non-nodejs dependencies that come along with
Electron apps, such as electron itself, and THEN you need to extract
and remove all of the vendor'd libraries and binaries that essentially
make Electron work, and THEN you need to make sure that there's no
side-car'd non-free or questionable software that is forbidden in
fedora also, like ffmpeg. G'head and look at the Chromium SPEC, it's a
living nightmare (Spot godbless your heart)"
---

Now obviously you could do what for example Java does with a jar file,
and simply throw everything into a single rpm/deb and ignore the
packaging guidelines, but then you are back to in essence creating a
container and just hiding it behind a massive rpm/deb.

>The thing we can never measure and thus can only speculate about forever
>is:  if all the person-hours poured into containers (and pypi/pip and cran
>and cpan and maven and scons and ...) had been poured into rpm/deb
>packaging would we just be simply apt/yum/dnf installing what we needed
>today? (I'm ignoring other OS/packaging tools, but you get the idea.)

I (theoretically) could write a new library in
Python/Perl/Javascript/Go/etc. and with very minimal effort can place
that library in the repository for that language with minimal effort.
My library is now available to everyone using that language regardless
of what OS they are using.

Alternately, I could spend many, many hours perhaps even days learning
multiple different packaging systems, joining multiple different
mailing lists / bugzillas / build systems, so that I can make my
library easily available to people on Windows, macOS, Fedora, RHEL,
openSUSE, Debian, Ubuntu, ...  - or alternately hope that someone will
not only take the time to package my library for all those different
platforms, but also commit the future time to keep it up to date.

Option 2 worked 20 years ago when we only cared about 2 or 3
distributions of Linux and had a lot less open source / free software.
But, unfortunately, it does not scale and so for that reason (and a
few others) the effort to create Docker / npm, maven, etc. is the
lesser of the options.

* - https://www.reddit.com/r/Fedora/comments/a3q1a2/atom_editoride/
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] HPC workflows

[Gerald Henriksen]

On Mon, 3 Dec 2018 10:12:10 -0800, you wrote:

> And then I realized that I was seeing
>software which was "easier to containerize" and that "easier to
>containerize" really meant "written by people who can't figure out
>'./configure; make; make install' and who build on a sand-like foundation
>of fragile dependencies to the extent that it only runs on their Ubuntu
>laptop so you have to put their Ubuntu laptop in a container."

The problem is that essentially nobody knows how autotools works, so
that those C/Fortran codes that use it have usually copy/pasted
something until it seems to work.

So 2 things happened.

First, all the non-traditional languages created their own build
systems, and more importantly their own package management systems.
This developed because most development was happening on non-Linux
systems, because Linux still struggles on laptops and laptops have
taken over the non-server computer world.  It also happened because
those developers using Linux, or at least aware of deploying on Linux,
rebelled at the limitations of the Linux ecosystem (namely
libraries/components that hadn't been natively packaged, or the normal
conflict of the "wrong" version being packaged).

A side effect of all these package management systems is that they are
frequently hostile to the "Linux way", and create software that is
essentially unable to be packaged into RPM or deb format.

The other issue of course is that open source won, and the explosion
of open source means the distributions no longer have the person-power
not just to package everything, but for those packages to do much of
the heavy lifting in keeping the software up to date.

As for autotools, it to is now being abandoned with the 2 leading
contenders being cmake and meson, but it being C++ the chaos wouldn't
be complete with multiple competing package management solutions...

> Then I
>started asking myself "do I want to trust software of that quality?" And
>after that, "do I want to trust the tools written to support that type of
>poor-quality software?"

On the other hand can you really trust the software built in more
traditional ways? see OpenSSL / Heartbleed.

>From the perspective of the software being containerized, I'm even more
>skeptical. In my world (bioinformatics) I install a lot of crappy software.
>We're talking stuff resulting from "I read the first three days of 'learn
>python in 21 days' and now I'm an expert, just run this after installing
>these 17 things from pypi...and trust the output" I'm good friends with
>crappy software, we hang out together a lot. To me it just doesn't feel
>like making crappy software more portable is the *right* thing to do. When
>I walk my dog, I follow him with a bag and "containerize" what drops out.
>It makes it easier to carry around, but doesn't change what it is. As of
>today I see the biggest benefit of containers as that they force a
>developer to actually document the install procedure somewhere in a way
>that actually has to work so we can see firsthand how ridiculous it is
>(*cough* tensorflow *cough*).

All very true.  To paraphrase, containers are the best of a bunch of
bad options.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Fortran is Awesome

[Gerald Henriksen]

On Tue, 4 Dec 2018 11:25:01 -0500, you wrote:

> From my experience, CS professors like to teach in Java because it's a 
>"pure" object-oriented programming language, unlike C++ which is just C 
>with object-oriented extensions,and still allows you to use C syntax.

Java took over in part because at one point it was the "in" language,
but more importantly it was a) free and b) available where needed in a
single version - on Solaris or Linux for the computer labs, Macs and
Windows for the students at home.

The other options when the switch happened in the late 90s either
required money to buy tools or weren't standardized very well across
platforms.

Of course other languages like Python would later follow, but by then
Java was established and difficult to dethrone though Oracle might be
making it easy to abandon at this point.

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] HPC workflows

[Gerald Henriksen]

On Sat, 1 Dec 2018 06:43:05 +0100, you wrote:

>My own thoughts on HPC for a tightly coupled, on premise setup is that we
>need a lightweight OS on the nodes, which does the bare minimum. No general
>purpose utilities, no GUIS, nothing but network and storage. And container
>support.

One of the latest attempts at this is Fedora CoreOS, the merger of
Fedora Atomic and CoreOS (which Red Hat bought).

https://coreos.fedoraproject.org/

>The cluster will have the normal login nodes of course but will present
>itself as a 'black box' to run containers.
>But - given my herd analogy above - will we see that? Or will we see
>private Openstack setups?

Maybe, Red Hat appears to be moving in that direction as well with a
Red Hat CoreOS offering with OpenShift though how it all ends up is
yet to be seen I suspect.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] HPC workflows

[Gerald Henriksen]

On Wed, 28 Nov 2018 13:51:05 +0100, you wrote:

>Now I am all for connecting divers and flexible workflows to true HPC systems 
>and grids that feel different if not experienced
>with (otherwise what is the use of a computer if there are no users making use 
>of it?), but do not make the mistake of thinking
>everything is cloud or will be cloud soon that fast. 

The "cloud" is a massive business that is currently growing fast.

Will it take over everything or continue its growth forever, of course
not.

But dismissing it is equally a dangerous thing to do, particularly if
your job relies on something not being in the cloud.

>So, one could say bare metal cloud have arisen mostly because of this but they 
>also do come with expenses. Somehow I find that a
>simple rule always seems to apply; if more people in a scheme need to be paid, 
>the scheme is probably more expensive than
>alternatives, if available. Or state differently; If you can do things 
>yourself, it is always a cheaper option than let some
>others do things (under normal 'open market' rules and excluding the option of 
>slavery :)).

But this is one area where the cloud can often win - the scale of the
Azure/Google/AWS operations means that you get 24/7/365 coverage with
essentially the lowest possible labour overhead.

And the fact is that while much of society insists on making decisions
purely based on cost - see airfares for example - there are a lot of
cases where people are willing to pay a premium for a service/product
that "just works".

>One has to note that in academia one often is in the situation that grants are 
>obtained to buy hardware and that running costs
>(i.e. electricity and rack space) are matched by the university making the 
>case of spending the grant money on paying amazone or
>google to do your 'compute' not so sensible if you can do things yourself.

Currently.

If on premise HPC doesn't reflect the ease of use that can be found
elsewhere, combined with some lobbying by the existing or specialized
cloud providers, and those grants could become a lot more flexible.

And given that many/most/all universities are often short on space and
they may well welcome an opportunity to be able to repurpose an
existing cluster space...

>There is also another aspect when for example dealing with sensitive data you 
>are to be helt responsible for. The Cloud model is
>not so friendly under those circumstances either. Again your data is put "on 
>someone else's computer". Thinking of GDPR and
>such.

I don't think this is so clear an advantage to on premise as some
think.

I think the fact that we are all on this mailing list in order to
learn and discuss issues puts us as an outlier - there are very few
people participating on this list, and even allowing for discussions
happening on other sites I (sadly) suspect you will find that the
majority of people running HPC aren't as informed as they should be.

Who do you trust more to keep your data safe - to keep systems
patched, to keep firewalls up to date, to properly configure
everything, etc.?  Is it your local HPC, where maybe they are
struggling to hire staff, or can't afford to offer a "good enough"
salary, or simply can't justify hiring a security specialist?  Or
perhaps you go with Google or Microsoft, who have entire departments
of staff dealing with these issues, who monitor their networks full
time looking for flaws?

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] HPC workflows

[Gerald Henriksen]

On Tue, 27 Nov 2018 07:51:06 -0500, you wrote:

>On Mon, Nov 26, 2018 at 9:50 PM Gerald Henriksen  wrote:
>> On Mon, 26 Nov 2018 16:26:42 +0100, you wrote:
>> If on premise HPC doesn't change to reflect the way the software is
>> developed today then the users will in the future prefer cloud HPC.
>>
>> I guess it is a brave new world for on premise HPC as far as that the
>> users now, and likely more in the future, will have alternatives thus
>> forcing the on premise HPC to "compete" in order to survive.
>
>this seems a bit too stringent of a statement for me.  i don't dismiss
>or disagree with your premise, but i don't entirely agree that HPC
>"must" change in order to compete.  We've all heard this kind of stuff
>in the past if x doesn't change y will take over the world!

HPC, like most things, exists to get something done.

If HPC doesn't change to reflect the changes in society and the way
the software is developed (*) then the users will look for more modern
ways to replace traditional HPC.  As noted the software is no longer
developed on workstations that are connected to the lab/company
network but rather on laptops that stay with the user wherever they
go.

This in turn is at least in part what has driven to the rise of
distributed version control, git in particular.

If HPC doesn't make it easy for these users to transfer their workflow
to the cluster, and the cloud providers do, then the users will move
to using the cloud even if it costs them 10%, 20% more because at the
end of the day it is about getting the job done and not about spending
time to work to antiquated methods of putting jobs in a cluster.

And of course if the users would rather spend their department budgets
with Amazon, Azure, Google, or others then the next upgrade cycle
their won't be any money for the in house cluster...


* - note the HPC isn't unique in this regard.  The Linux distributions
are facing their own version of this, where much of the software is no
longer packagable in the traditional sense as it instead relies on
language specific packaging systems and languages that don't lend
themselves to the older rpm/deb style system.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] HPC workflows

[Gerald Henriksen]

On Mon, 26 Nov 2018 16:26:42 +0100, you wrote:

>This leads me to ask - shoudl we be presenting HPC services as a 'cloud'
>service, no matter that it is a non-virtualised on-premise setup?
>In which case the way to deploy software would be via downloading from
>Repos.
>I guess this is actually more common nowadays.

Simple answer yes.

If on premise HPC doesn't change to reflect the way the software is
developed today then the users will in the future prefer cloud HPC.

I guess it is a brave new world for on premise HPC as far as that the
users now, and likely more in the future, will have alternatives thus
forcing the on premise HPC to "compete" in order to survive.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Oh.. IBM eats Red Hat

[Gerald Henriksen]

On Tue, 06 Nov 2018 14:21:20 -0800, you wrote:

>KDE is being dropped and guess who supports Gnome3...IBM!

Coincidence.

The timing of the announcement is such that the IBM purchase could
have nothing to do with it, certain headlines notwithstanding.

Given how bad KDE is on RHEL (RHEL 7 is still on KDE4) having the
Fedora KDE sig provide an up to date version for RHEL 8 via EPEL is
likely a better option anyway.

>Centos support will probably be dropped completely maybe even extinguished.

Time will tell, though a) it would be a bad move and b) unless CentOS
was on the verge of collapse prior to being absorbed by RH it would
likely just rebirth itself as an independent project again though
perhaps needing a new name.

>I though Fedora was killed of a couple years ago after version 7.

Red Hat has never (publicly at least) pondered killing Fedora and
still continues to invest significantly into it.

Of the 2 Fedora is likely more vulnerable because its utility to a
server version of RHEL isn't as clear and it is unlikely that it would
survive without the funding (direct and indirect) of Red Hat.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Oh.. IBM eats Red Hat

[Gerald Henriksen]

On Sun, 04 Nov 2018 11:28:31 +, you wrote:

>There is one thing that is going to be super interesting to see. Red hat 
>fairly recently absorbed Centos Dev's and added them to the pay roll.
>
>Two questions yet to be answered are

Well, anything is obviously possible and there is certainly examples
of companies ruining / running into the ground aquisitions.

But with the deal not being completed for another 6 months or so there
will be a lot of ananswered questions for a while.

>1) whats the future of Centos

This one is easy (at least in a sensible world) - CentOS needs to
continue.

The biggest threat to RHEL isn't lost sales to CentOS but losing
customers and mindshare to Ubuntu (which certainly appears to have
been an issue the last number of years based on the number of software
projects that support Ubuntu but not Red Hat).

Anything that helps maintain / stop the erosion of minshare /
marketshare needs to be supported and to me CentOS delivers that (at
least to the extent of what Red Hat is willing to do).

>2) whats the future of Fedora.

More difficult to answer, and in some ways goes beyond Fedora.

There are some uncomfortable (from an open source purity standpoint)
questions that need to be faced, and Fedora in some ways complicates
them given it is officially community run and governed.

I my opinion it is a question Red Hat should have discussing at the
executive level if if the IBM deal hadn't come along.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] More about those underwater data centers

[Gerald Henriksen]

On Sat, 3 Nov 2018 18:27:05 +, you wrote:

> I’m not sure there’s a huge population of Xcloud-Xbox gamers in Orkney.  
> There's not much daylight this time of year, of course, so maybe that's what 
> those Orcadians are up to.

Likely just a convenient place for a second test unit.

In a way this is just an extension of the idea/product Sun came up wth
where they put a datacentre in a shipping container with the idea that
you could quickly get the datacentre where it was needed.

While I wouldn't say this won't fail, I think there is a lot of
attraction to the concept given not just the time lag do build a
traditional data centre (mentioned in the article), but even the cost
of real estate in many/most places people live these days.  Do you,
for one example, want to pay NYC rents or just throw a bunch of pods
in the Hudson?

I guess once you accept the idea that we no longer maintain these
datacentres in the traditional way - we now just let hardware fail in
place and ignore it until it's time to replace all the hardware -
moving to smaller sealed units doesn't seem to strange.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Oh.. IBM eats Red Hat

[Gerald Henriksen]

On Wed, 31 Oct 2018 09:22:00 -0400 (EDT), you wrote:

>But your observations certainly have some validity, which is one of many
>reasons that there is room in the universe for competition here.  I
>really have wanted to be able to go both ways -- run Android apps on my
>laptops (including paid apps, games, and more) and run "real" linux apps
>on my android devices (tablets with large screens and keyboards ARE
>"laptops", or could be), or variants thereof.  IBM, or anybody else who
>understands the OSS universe and doesn't try to over-monetize it, has
>plenty of opportunity here as the COMMUNITY would (IMO) build a true
>open/linux tiering that would run, within reason and the constraints of
>the device(s) on things from phones through server rooms.  

The community tried, it was known as Maemo, and it came out 2 years
prior to the iPhone and Nokia eventually offered it on hardware.

But in true OSS fashion they couldn't make decisions, or stick to the
decision they did make.

It started out GTK based, and then they added Qt, and then other stuff
was thrown in, resulting in a mess with no consistent look and feel,
no constent APIs, and contradictory message to developers.

Needless to say it failed.

>> Say want one wants about Apple, but they have had the best ARM
>> processors for mobile for a while now and noone appears to be close to
>> catching up anytime soon.
>
>You could be right, but it would be a shame if you were.  As I said,
>while Apple had the first "personal" computer, IBM came in and changed
>processors, changed OS philosophy and design, changed the software
>market, and blew Apple away to the point where (hard as it is to

Not how I remember it.

My recollection is IBM cobbled together a PC, expected it to be a
failure, and as a result of the cobbling made some "mistakes" that
allowed the clones to arrive.  It was the clones, let by Compaq, that
led to the demise of Apple and the rise of Microsoft.

IBM fought the clones, tried to return things to the "one true IBM
way" with the proprietary Micro-Channel Architecture but the market
ignored them and the decline of IBM started.

Or maybe you meant that wonder of IBM prowess known as PCjr.

>remember) they almost went away at their low point before their
>NON-computer offerings brought them back from the brink of extinction.

Actually, Microsoft saved Apple with a $150 million dollar investment.

>IBM absolutely could do this again.

Given that they didn't do it the first time, doubtful.

>acquiring Red Hat, while still owning and running an extensive hardware
>business, IBM will be the first company in years that has both.  I'm not
>by any means certain that they WILL, but they COULD.  They COULD
>actually build and sell the so-far mythical pre-installed linux laptop
>and desktop that does NOT come with M$ as even an option,

You mean like System76?

Also, IBM has no current experience in the mass market and doesn't
even own its own fabs anymore.

> and do so at
>competitive prices and with IBM's quality assurance and support, and
>could generate an instant market for software developers for said
>preinstalled hardware.  If I think way, way back, they actually flirted
>briefly with this back in the days of the ordinary PC, but M$ at that
>time had the Intel marketplace in a monopoly hammerlock.

IBM was never competitivly priced, hence the rise of the clones.

>No more.  Apple sells Intel-based boxes running Unix with the Apple
>special sauce in their graphical interface and management layer.  I buy
>Intel boxes and install Unix (in the form of linux) with the Linux
>special sauce and a choice of graphical interfaces and management tools.

Which is why Linux will never succeed on the Desktop - the market in
general wants 1 Desktop not 20.

>The only difference between the end result is that I have an enormous --
>truly mind-boggling -- set of free software I can then install without
>getting up from my seat or spending a nickel of additional money (on the
>good side) 

True also for Windows and macOS.

>IBM could pick an archtecture -- like the Thinkpad I'm typing this onto
>under Fedora 28, assuming not unreasonably that they still are tight
>with Lenovo

In 2018 partnering with a Chinese company isn't a wise move.

>support for all of the system's devices and with its OS cleaned up just
>a hair to match the functionality of Apple's best efforts to date.  They
>could dump some love in the general direction of game companies to win
>young hearts and minds. 

You mean like Steam, who actually has the connections with game
companies, who actually has put money into getting parts of Linux up
to scratch, and has about 1% Linux sales?

> They've got the sales force and marketing
>relationships to be able to walk into any fortune 500 company and sell a
>top to bottom IBM solution, with IBM support, and YET with that huge
>mountain of software that they have to pay for for any M$ OR Apple based
>solution.

No, IBM no longer has the sales force or 

Re: [Beowulf] Oh.. IBM eats Red Hat

[Gerald Henriksen]

On Wed, 31 Oct 2018 06:44:38 +, you wrote:

>MArk Shuttleworth's statement here
>https://blog.ubuntu.com/2018/10/30/statement-on-ibm-acquisition-of-red-hat
>I make no comment.

Sounds like a for sale listing.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Oh.. IBM eats Red Hat

[Gerald Henriksen]

On Tue, 30 Oct 2018 13:49:05 -0400 (EDT), you wrote:

>> "Will Amazon, Google, and Microsoft now run out and buy SUSE, Ubuntu,
>> Apache, etc? Yes.
>> 
>> "Will there be a mad rush to create new Linux distros? No. I think that boat
>> has already sailed and further Linux branding won?t happen, at least not for
>> traditional business reasons.
>
>Google already has one.  It's called "Android".  Microsoft has been
>flirting with Linux for the first time in forever as every one of their
>efforts to compete with Android and IOS has underwhelmed, if not
>flopped.  At this point Linux actually owns a substantial chunk of the
>desktop, in the form of Android on tablets that have largely replaced or
>augmented actual computers,

Not even close.  Android on tablets is essentially dead, they only
thing the remaining tablets are used for is media viewers.

The Android app developers never developed tablet versions of the
apps, and Google dropped the ball yet again.

Google's latest attempt, announced several weeks back, is an attempt
at a ChromeOS tablet...

And when that fails, perhaps we will finally see Fuchsia.

> and is just under Apple in the phone market

Depends on how you measure, some ways Android is ahead.

But the big problem is that few people are making money from Android,
in part because few Android users buy apps.

>with M$ a joke down near the bottom in both domains.  

Microsoft existed the phone business several years ago, and at least
in terms of usefulness for doing anything other than watching a movie
their tablet offerings are far ahead of anything Android.  It's not a
coincidence that the latest iPad Pros announced today look like
Surface Pros.

>But Android is
>vulnerable -- lots of people dislike it and dislike the play store and
>all that goes with it and with iOS. 

The only people who hate the iOS store are certain techies, and
certain players big enough that they dislike the percentage.

The users love it because it is hassle free and has none of the
virus/malware/etc issues that they experience elsewhere.

Google Play is vulnerable because Google refuses to invest in it and
the developers increasingly leary of getting locked out with no way to
appeal the decision.

> IBM has the resources to actually
>make an OPEN tablet/phone OS if they choose to and are at least as
>likely as M$ is to be able to step into the market and steal away
>mindshare from Android and iOS -- if they couple it to a slick AI
>component, maybe semi-proprietary, they might even jump to the head of
>the line as Alexa and Siri etc leave a great deal to be desired.

Any attempt at an open tablet/phone would require signficant money and
there is no way IBM is going to invest the money in hardware and
software to try it.

Say want one wants about Apple, but they have had the best ARM
processors for mobile for a while now and noone appears to be close to
catching up anytime soon.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] New tools from FB and Uber

[Gerald Henriksen]

On Wed, 31 Oct 2018 01:27:23 +0800, you wrote:

>FB has open sourced some interestimg kernel tools, including 'cgroups2' and
>btfrs (!! Wasn't that already floss?)

Btrfs has been around for a while, and has even been in the Linux
kernel for a while.

I'm guessing that Facebook has had their own development branch which
they are now making public as that link differs from the development
branch that is shown on the kernel Btrfs page:

https://btrfs.wiki.kernel.org/index.php/Btrfs_source_repositories

Currently I believe only OpenSUSE ships it in a supported form, Red
Hat announced they are removing support and it won't be in future
releases.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Oh.. IBM eats Red Hat

[Gerald Henriksen]

On Mon, 29 Oct 2018 11:58:18 -0400, you wrote:

>The downside to (most) of the stable distros are the aging compilers, 
>languages, and libraries.  RH ships with 4.9.x, Debian 9.x ships with 
>6.3.x.  You can easily install gcc7 and gcc8 in debian.  Its a little 
>harder for pre-built rpms in RH (and its never a good idea to replace 
>distro required packages with updated ones ... always use a separate 
>tree, or a container).

Red Hat provides a parallel installable set of compilers that are
reasonably up to date, and the collection even includes languages that
weren't around (or barely around) when Red Hat 7 was released like Go
and Rust.

https://developers.redhat.com/blog/2018/04/06/new-red-hat-compilers-in-beta-clang-llvm-gcc-go-rust/

>Python 2.x is dead, 3.x should be used/shipped everywhere.

Even Fedora is still working to try and kill Python 2.x, blame Python
of mishandling the transition and allowing Python 2.x to last for so
long that is failed to encourage devlopers to port their code to
Python 3.x

But the bigger problem is that for whatever reason Red Hat is late in
getting Red Hat 8 out the door - it was just under 4 years between Red
Hat 6.0 and Red Hat 7.0 

It is now 4.5 years since the release of RHEL7 and we still don't even
have a beta of 8, and no indication that a beta will be anytime soon.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Oh.. IBM eats Red Hat

[Gerald Henriksen]

On Mon, 29 Oct 2018 16:49:29 +, you wrote:

>> oh, but RH's function is so much more nowadays than just a paid for
>distribution.
>> hence the acquisition which is not about that. but the whole ecosystem
>can suffer as a result.
>Well said.  coreutils   gccllvm….  what of them?
>Redhat does a lot for the core stack.
>
>And systemd  (arrrghhh  - I said it again)

And Gnome, Java, X/Wayland, etc.  It's not just the core stack but a
lot of the open source Linux ecosystem relies on paid Red Hat
developers.  A decision by IBM, in say 2 or 3 years, to focus entirely
on cloud/server could have a detrimental effect on a lot of Linux
stuff.

But not LLVM, that is an entirely independent project (though IBM does
sort of support the Power version of LLVM based on demand from IBM's
customers).
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Oh.. IBM eats Red Hat

[Gerald Henriksen]

On Mon, 29 Oct 2018 15:45:32 -0400, you wrote:

>How well has Linux been supporting IBM's POWER processors? I would 
>imagine pretty well, since the Linux community always seems eager to run 
>on new hardware.

Quite well, but it really isn't community driven.

Fedora has both PPC64 and PPC64LE available to download, though the
PPC64 is being phased out I believe.

The bigger problem is that while the OS is reasonable about all you
can really say about the rest of the software is that it compiles -
the lack of affordable hardware seriously restricts the testing and
development that occurs.

>Could it be that the Linux community hasn't been quick enough in 
>accepting IBM's contributions to Linux to support POWER, so IBM decided 
>to buy the largest member of the community to fix that?
>
>Or, could it be that IBM didn't have enough Linux expertise in-house to 
>support POWER on Linux, so they decided to buy a lot of expertise?

This is purely a cloud play, and that is one of the reasons for
concern.

IBM has been in decline for a while, and the stock markets aren't
entirely happy with the declining revenue for a number of years now.

So this is an attempt to reverse things at IBM by getting into the
Amazon AWS/Google Cloud/Microsoft Azure business which is a growing
market.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] C++ compilers and assembly

[Gerald Henriksen]

On Mon, 10 Sep 2018 04:50:24 +0100, you wrote:

>Lastly Jason mentioned IncludeOS http://www.includeos.org/
>these gusy are implementing something I have tossed about on this list
>in the past - running applications in a lightweight OS without all the
>overhead of a multitasking system.
>I dont know that much about it - is anyone from that company on the list?

The founder/one of the founders participated in a pocast 2 years ago:

http://cppcast.com/2016/07/alfred-bratterud/
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] If I were specifying a new custer...

[Gerald Henriksen]

On Wed, 17 Oct 2018 16:35:52 +1100, you wrote:

>On Saturday, 13 October 2018 12:38:15 AM AEDT Gerald Henriksen wrote:
>
>> If ARM, or Power, want to move from their current positions in the
>> market they really need to provide affordable developer machines,
>
>Not sure if this comes in at a price point that makes sense for this, but 
>there is now an ATX Power9 mainboard available.
>
>https://raptorcs.com/TALOSIILITE/
>
>They claim:
>
>https://twitter.com/RaptorCompSys/status/1020371675316215809
>
># TalosIILite in stock and ready to ship! #POWER9 mainboard + CPU + RAM + HSF
># for under $2,000 USD, what's not to like? Supports all of our Sforza CPU
># options, from 4 core to the high end 22 core CPUs.

Not really.

While there obviously is a lot of corporate funded work in the open
source community I would guess little of it is interested in anything
but the traditional AMD/Intel systems, and maybe ARM.

To give a new / minor platform traction you really need to have
something priced where it can be a personal purchase, normally as a
secondary machine as few people will move to ARM or Power (at this
point) as an only machine.

They do however have another platform coming, no prices yet, but
should hopefully be more affordable in the Blackbird:

https://wiki.raptorcs.com/wiki/Blackbird

And to point out what is obvious to many, the reason these cheaper
systems are needed is to get all that open source software working and
tested, Raptor has had to do work just to get Chrome working on Power:

https://www.phoronix.com/scan.php?page=news_item=Raptor-Chrome-JIT-PPC64LE-Work
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] If I were specifying a new custer...

[Gerald Henriksen]

On Fri, 12 Oct 2018 09:24:18 +0100, you wrote:

>Doug,
>I think the ARM/Cavium Thunder is going to see a lot of attention.
>I saw a report recently from the Bristol/Cray Brunel cluster - they are
>offering a range of chemistry codes and OpenFOAM,
>compiled up for ARM.

Perhaps this paper?
http://uob-hpc.github.io/assets/cug-2018.pdf

Also interesting is another entrant into the ARM server market is
Ampere, who have had some hardware tested by Phoronix:

https://www.phoronix.com/scan.php?page=article=ampere-emag-osprey=1

Perhaps more importantly though is they appear to be a bit more
developer / open source community friendly with an employee commenting
in the Phoronix discussion thread and they even have a developer
website:

https://developer.amperecomputing.com/
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] If I were specifying a new custer...

[Gerald Henriksen]

On Fri, 12 Oct 2018 09:24:18 +0100, you wrote:

>I think the ARM/Cavium Thunder is going to see a lot of attention.
>I saw a report recently from the Bristol/Cray Brunel cluster - they are
>offering a range of chemistry codes and OpenFOAM,
>compiled up for ARM.
>Poke me and I will search for the report - I saw it on a twitter feed.

ARM essentially has 2 problems.

One, you go from cheap/limited SOC boards like the Raspberry Pi and
jump straight into the expensive Cavium line.  There is no affordable
ARM option for developers to use painfree to write code, port code,
and test code.

If ARM, or Power, want to move from their current positions in the
market they really need to provide affordable developer machines,
machines priced at a point where the open source community can justify
buying it as a secondary machine to work on making the Linux
distributions more solid and the associated software.

Two, for whatever reason the only company that seems to be able to
produce ARM processors with decent peformance is Apple, and they
aren't giving up their advantage.  Microsoft would like to move into
the ARM based notebook market but so far Qualcomm doesn't seem to be
able to provide a chip with adequate performance.  The ARM ecosystem
needs to do much better at converting the ARM designs into actual
hardware.

>Regarding the Intel Fab issues, what is the public evidence for this?
>I guess there may be a lot of industry scuttlebutt around.

Earning reports, stock downgrades, prices.

Intel is having serious issues moving from 14nm to 10nm, is  3 years
late and now saying sometime in 2019, and is having supply issues
given they have fabs unable to supply product as the conversion
continues to be extended.

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] OT, X11 editor which works well for very remote systems?

[Gerald Henriksen]

On Wed, 06 Jun 2018 16:04:07 -0700, you wrote:

>On 06-Jun-2018 15:28, Fred Youhanaie wrote:
>> Does enabling ssh compression with -C help?
>
>The local side is over putty from a Windows machine.  Enabled its ssh 
>compression option, which I think is the same thing.  It didn't make a 
>noticeable difference.

Haven't used it but Visual Studio Code can apparently remote edit
using an extension:

https://spin.atomicobject.com/2017/12/18/remote-vscode-file-editing/


___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] [upgrade strategy] Intel CPU design bug & security flaw - kernel fix imposes performance penalty

[Gerald Henriksen]

On Fri, 5 Jan 2018 17:32:19 -0500, you wrote:

>According to several articles I read today:
>
>Meltdown (1 exploit)is Intel-specific
>Spectre  (2 different exploits) affects just about every processor on 
>the planet.

For anyone interested this is AMD's response:

https://www.amd.com/en/corporate/speculative-execution


The Google Project Zero number/titles being used correspond to:

1 & 2 - Spectre

3 - Meltdown

Also an explanation in less technical terms from Red Hat:

https://www.redhat.com/en/blog/what-are-meltdown-and-spectre-here%E2%80%99s-what-you-need-know?sc_cid=701600127NJAAY
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] nVidia revealed as evil

[Gerald Henriksen]

On Fri, 5 Jan 2018 22:52:19 +, you wrote:

>There has been a conversation going on on the AMBER mailing list for some 
>time, related to this and specifically to the Volta card in some way, since 
>AMBER performs best on the consumer grade stuff and doesn’t require the 
>enterprise class features (I guess the reason for the performance difference 
>is that the next gen consumer cards come out first?).

Actually, the consumer cards (at least for the last generation or so)
have been released last.

Volta has not yet been released in a consumer form - the cheapest
available so far is the Titan V (direct from Nvidia only) at $3k I
believe.  Speculation is that consumer Volta may come out this year,
but there is no pressure on Nvidia given AMD's troubles in the GPU
market.

The big difference is that of course most of the work is being done on
consumer hardware because that's what the developers and researchers
can afford.  Secondarily, the new feature Volta offers is still to
recent to be properly supported (and some software may get no benefit
from it) because affordable hardware isn't yet available to allow
developer access to the new Tensor cores that Volta offers.

>Anyhow, I spoke to an NVIDIA rep about it at SC17 and he kind of said “we are 
>happy you’re buying whatever chip of ours). I said, sure, maybe, but I 
>understand you’re putting the screws to the systems vendors so how are we 
>supposed to buy them. Didn’t get a real concrete answer.

I would think if Nvidia isn't careful this could provide an opening
for AMD.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] [upgrade strategy] Intel CPU design bug & security flaw - kernel fix imposes performance penalty

[Gerald Henriksen]

On Fri, 5 Jan 2018 17:32:19 -0500, you wrote:

>According to several articles I read today:
>
>Meltdown (1 exploit)is Intel-specific
>Spectre  (2 different exploits) affects just about every processor on 
>the planet.

This is correct, and the other key difference is that so far there is
only a solution to Meltdown.

The below linked page has this to say about Spectre:

"As it is not easy to fix, it will haunt us for quite some time"

Fedora has linked to this page for more info on the two:

https://spectreattack.com/

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Intel kills Knights Hill, Xeon Phi line "being revised"

[Gerald Henriksen]

On Wed, Dec 20, 2017 at 7:02 AM, Mikhail Kuzminsky  wrote:
>
> I partially disagree with "confusion". It's simple because KNM has minimal
> microarchitecture changes vs KNL, and does not focus on normal
> DP-precision. KNM focuses on SP etc, and is oriented to Deep Learning, AI
> etc.
>
>
>
By confusion I meant the product names - Hill and Mill will be easily
confused and mixed up in discussions and people's minds.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Intel kills Knights Hill, Xeon Phi line "being revised"

[Gerald Henriksen]

On Thu, 16 Nov 2017 10:45:40 +1100, you wrote:

>Interesting times (via a colleague on the Australian HPC Slack).
>
>https://www.top500.org/news/intel-dumps-knights-hill-future-of-xeon-phi-product-line-uncertain/
>
>Looks like fallout from the delayed Aurora system.
>
>Rumours flying that the Xeon Phi family is in jeopardy, but the
>article has an addendum to say:
>
># [Update: Intel denies they are dropping the Xeon Phi line,
># saying only that it has "been revised based on recent
># customer and overall market needs."]

This should cause some confusion.

While Knights Hill was cancelled, Intel has quietly put information
about Knights Mill online as the next Phi product line:

https://www.anandtech.com/show/12172/intel-lists-knights-mill-xeon-phi-on-ark-up-to-72-cores-at-320w-with-qfma-and-vnni
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Intel kills Knights Hill, Xeon Phi line "being revised"

[Gerald Henriksen]

On Thu, 16 Nov 2017 10:45:40 +1100, you wrote:

>Interesting times (via a colleague on the Australian HPC Slack).
>
>https://www.top500.org/news/intel-dumps-knights-hill-future-of-xeon-phi-product-line-uncertain/
>
>Looks like fallout from the delayed Aurora system.
>
>Rumours flying that the Xeon Phi family is in jeopardy, but the
>article has an addendum to say:
>
># [Update: Intel denies they are dropping the Xeon Phi line,
># saying only that it has "been revised based on recent
># customer and overall market needs."]

Maybe worth pointing out that Intel has big changes in store, which
may or may not be a factor in the Xeon Phi future:

https://www.theverge.com/2017/11/9/16627470/intel-raja-koduri-graphics-amd-radeon-business-hire

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

[Beowulf] Your Raspberry Pi dream comes true

[Gerald Henriksen]

 750 node cluster in use at LANL for development purposes:

https://www.anandtech.com/show/12037/cheap-supercomputers-lanl-has-750node-raspberry-pi-development-clusters

https://www.servethehome.com/bitscope-raspberry-pi-cluster-3000-cores-30u/

http://cluster.bitscope.com/solutions
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

[Beowulf] ARM aims at HPC with new instruction set

[Gerald Henriksen]

ARM V8-A adds scalable vector extensions:

http://www.anandtech.com/show/10586/arm-announces-arm-v8a-with-scalable-vector-extensions-aiming-for-hpc-and-data-center
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

[Beowulf] Lastest Xeon Phi

[Gerald Henriksen]

In a 2U chasis you can get 4 Xeon Phi systems (each with 72 cores):

http://www.anandtech.com/show/10553/asrock-rack-launches-2u4nf-x-200-knights-landing-xeon-phi-cpu
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] recommendations on ARM distro?

[Gerald Henriksen]

Forgot in my reply to include a link to the $300 board:

http://www.lenovator.com/product/103.html

There is also supposed to be something called a HuskyBoard but no
indication of when/if it will ship:

http://www.96boards.org/products/ee/

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] recommendations on ARM distro?

[Gerald Henriksen]

On Mon, 16 May 2016 00:27:21 +0800, you wrote:

>Very nice reply Gerald, not meaning to nit, but for certain workloads
>I'd emphasize that accelerators make more sense than Intel.

Yes, and more below.

>Power and ARM have some uphill battles ahead of them, but I'm
>optimistic that in the next 2-4 years we're going to see Intel go
>against increasingly interesting products.

The good news for Power and ARM is that Intel is in a little bit of
disarray, and while their new emphasis on Server and IoT is likely a
good thing it both depends on what they mean by that and how much of a
disruption the layoffs cause amongs the employees that stay.

The big advantage the ARM could have is the ability of someone to come
along and take the open design of ARM and create some unique hardware.
Quite unlikely, but how about a SoC designed for accelerators, with
say Infiniband in the SoC and support for say 5 accelerators on one
motherboard?

But, to get back to the question that started this thread, the biggest
problem that both Power and ARM have is getting the software working
correctly.  It is clearly unacceptable that getting something as basic
as LDAP working is a time consuming, trial and error task.

In my likely flawed view neither ARM nor Power will escape from their
current niche markets until they get "affordable" hardware to the
public.

The number of programming languages currently in use, the number of
frameworks, the number of applications available is simply beyond the
ability at this point of any distribution, Linux or *bsd, to test
everything.  Instead it is at best a if it compiles, it ships, with
only the core portions of the distribution getting any testing, and
then really only on the hardware that anyone has access to.

If Power had a motherboard/CPU combination in the $300 to $500 range,
and ARM in the $200 range, then all the people who work on the open
source software would be able to at least consider getting a second
system to play with, to test their software on, etc. and as a result
fix things or at least file bugs.  But as long as Power systems are
into the thousands of dollars, and ARM (or more specifically one of
their partners) is offering a 2, almost 3 year old CPU that is one to
two generations old and the unrealistic price of $300 (*) that aren't
going to get the public to buy their hardware, and thus they won't get
the testing/developing/etc of the software ecosystem.

It's a shame, because while I couldn't afford even my optimistically
priced examples, I think we do need a competitor to Intel to keep
things moving, but I just don't see that Power really wants to be that
and ARM is simply failing to execute.


* the cheapest price announced for an AArch64 board with proper
hardware (ie. RAM slots, SATA support, UEFI boot system) is $300 and
offers an AMD A1100 SoC that was announced just over 2 years ago. This
board is not yet shipping, and the reviews from January indicate that
the A1100 isn't really competitive (in performance/power
consumption/price) against Intel Atom offerings.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] recommendations on ARM distro?

[Gerald Henriksen]

On Fri, 13 May 2016 11:38:33 +0800, you wrote:

>My issue in general with aarch64 OS has been the pita install process. If 
>that's now insert USB drive.. click click done I'd be really happy. Trying to 
>deal with it via remote hands who may or may not be familiar with Linux is 
>painful

Red Hat has been pushing hard for aarch64 to support industry standard
BIOS/UEFI for booting the hardware, and for ACPI to control it, making
it much easier for the OS to boot and run.

The hardware vendors have been slow, but there is now some hardware
that supports those standards, though I have no experience with any of
it.

So it should be as easy as your USB drive soon.

For those interested Jon Masters (Red Hat - Chief ARM Architect) did a
talk earlier this year on the effort to get the ARM vendors to support
standards, including a brief demo at the end of a Qualcomm 24 core
system running Red Hat Enterprise 7.2:

https://www.youtube.com/watch?v=IVK0LqxULZc

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] recommendations on ARM distro?

[Gerald Henriksen]

On Sun, 15 May 2016 09:39:50 +0200, you wrote:

>Was actually thinking of a clustered server setup using 64bit arm board
>if possible

First problem is that while the RPi3 has a 64bit cpu, there is only
support for 32bit operating systems (you need binary blobs to get
Linux running on these ARM boards, and for RPi3 they only support
32bit).  

[the Pi people chose the chip for its speed while maintaining
backwards compatibility, thus they don't support or care about 64bit]

Your next problem is that there is only 1GB RAM, and most of the I/O
goes through a single USB2 port that has reports of being buggy.

The RPi3 (or any of the Pi models) can certainly be used for
educational purposes regarding a cluster, but unlikely to be actually
useful.

If you really want a cheap 64bit ARM board there are better options
available that actually run 64bit Linux though you may need to pay a
bit more.  Best advice is to find the ARM community for your preferred
version of Linux and see what boards they support given the
problematic nature of the cheap ARM boards.


Is there hope for the future of 64bit ARM?

Maybe.

Red Hat is pushing hard for the the ARM vendors to implement the
standard BIOS/UEFI boot process for AAarch64 boards which should
remove most of the issues booting Linux on AAarch64, but the ARM
hardware side has been very slow in coming to market and the cheapest
board so far supporting a BIOS/UEFI is in the $300 range I believe.


But for now, until the AAarch64 board makers can get their act
together, for getting work done (like the 3D rendering you mention in
another message) you are better off in the Intel/AMD world.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] recommendations on ARM distro?

[Gerald Henriksen]

On Fri, 13 May 2016 03:09:26 +0800, you wrote:

>2) Fedora has put quite a bit of effort into their AArch64 distro

The web based information about AArch64 Fedora is not the greatest,
but despite that Red Hat is putting a lot of effort into AArch64 and
making sure everything runs correctly.

The Fedora mailing list for ARM is:

https://lists.fedoraproject.org/admin/lists/arm.lists.fedoraproject.org/

which does have a several Red Hat people participating who are working
on the Fedora ARM ports.

And from the list the announcement of Fedora 24 beta for AArch64 with
a link to the download site:

https://lists.fedoraproject.org/archives/list/a...@lists.fedoraproject.org/thread/VWPJDPPFUUHYZMMP4RROPQXA5R4BQOVZ/

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

[Beowulf] New Tesla P100 from NVIDIA

[Gerald Henriksen]

Of course no price given, however:

http://venturebeat.com/2016/04/05/nvidia-creates-a-15b-transistor-chip-for-deep-learning/

http://anandtech.com/show/10222/nvidia-announces-tesla-p100-accelerator-pascal-power-for-hpc

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] RAID question

[Gerald Henriksen]

On Wed, 18 Mar 2015 14:00:15 -0700, you wrote:

The disk errors were a red herring. The system had a Seagate USB disk 
plugged into it which I was not aware of.  (It was less not obvious 
because of the rats nest of cables behind it.)  This disk's partition 
table was marked bootable - even though there was nothing on that disk 
which would have supported a boot.  This was the disk that was showing 
up as /dev/sdb.  When CentOS booted normally it was automatically 
mounting this disk, which is why there was no mention of it in 
/etc/fstab.  However, nothing was using this disk.  It looks like at 30 
minute intervals the OS pinged the device to see if it was still 
there, and the enclosure/disk did not fully support whatever command was 
being used for this operation, resulting in the sense error messages in 
the log files. When the rescue DVD was
used it saw this device, created /dev/sda for it (yes, device names were 
exchanged in the two environments) and didn't mount it.

Linux does not guarantee device names to remain the same, which is why
partitions are usually mounted via a unique partition ID in /etc/fstab

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Xeon D systems? (and 10G in general)

[Gerald Henriksen]

On Thu, 12 Mar 2015 10:49:48 -0300, you wrote:

I'm not sure of what to make of this new Xeon, especially because it 
cuts right through the E3-1200 series, as you can see in the link that 
John provided, but may I speculate a little with two possibilities:

- Intel is phasing out the E3-1200 series
( Unlikely, IMHO )

- Intel is beefing up its [def|off]ensive options against ARMv8, 
especially after Cray and Lenovo announced tests with Cavium's Thunder-X 
chip
   
 http://www.eweek.com/servers/cray-to-evaluate-arm-chips-in-its-supercomputers.html
   http://www.theplatform.net/2015/02/27/prototype-arm-clusters-muscle-hpc/



Any other bet ?

Don't know enough about Hadoop, but it almost seems designed for it.
The 10G network would help with both getting data onto the drives, and
then with the distributed file system, and the systems will be mostly
I/O bound so you don't need the greatest CPU.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] IBM paying GlobalFoundries to take their CPU business

[Gerald Henriksen]

On Tue, 21 Oct 2014 09:30:30 +1100, you wrote:

On 21/10/14 08:11, Douglas Eadline wrote:

 Which means to me, $1.5bn is far less than they would spend over
 the next 5-10 years running a FAB. And, I'm sure they
 get a special parking spot out front for when they need
 to make wafers. (i.e. there is more to the deal than the $1.5bn dowry)

My guesswork:

The $1.5 billion (which is payable over 3 years) keeps the current fab
running producing the processors IBM needs for the next 3 years while
they redesign their chips to work with a newer fab process.

Running fabs has become a big $ cost business, and more specifically
running a fab capable of producing server (or more generally non-ARM)
class chips.

This gets IBM out of it chicken-egg situation of designing for an
obsolete fab, and running a fab for its designs, without having to put
a lot more money into updating its fab.

In 3 years IBM can then have their designs being manufactured on
modern processes in someone elses fab.

Another guess, someone will specialize in making non-ARM chips and get
the business from Oracle (Sparc), IBM (Power), and AMD (x64) and
perhaps the AMD and NVIDIA GPU business, with the hope that the
production from the 3(5) of them will be viable given their needs are
the most similar.
 
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Gentoo in the HPC environment

[Gerald Henriksen]

On Sat, 28 Jun 2014 15:38:22 +0200, you wrote:

My arguments for gentoo is more along the lines of optimization of the
source code for ones hardware.

Every now and then someone gets on the Fedora list
suggesting/demanding that they change some compiler flags or other
optimizations that the person has used on a distribution like Gentoo
and seen significant performance gains.

The standard response is for the person to do some benchmarks that
prove the claimed improvement is real, and then it will be considered.

So far, no one has come back with proof of improvements.

For most use cases the OS is not a limitation, and expending effort to
improve things will not gain any significant improvements and may
possibly make things worse.

You mention CentOS and Red hat how optimized is the heart of the system
for the hard ware its running on?

What is there to optimize?

Yes, the kernel and a few key libraries have internal code paths that
can optimize based on processor but the 64bit world is a fairly
uniform instruction set (unlike the 32bit Intel world) with the
exception being the SSE and AVX type extensions which aren't of much
use to the OS.

I for instance currently even though nothing is running on it have a vps
with 2gb of ram and with no websites or db on there yet just the servers
ram wise i use about 62mb. I look at the low memory foot print as an
advantage in the sense that the rest of the ram can be used for more
important things.

That isn't so much optimizing for hardware as it is making sure you
are only running services/daemons/programs that you need for your
requirements, which is easy to do with any distribution.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Gentoo in the HPC environment

[Gerald Henriksen]

On Sat, 28 Jun 2014 16:54:34 +0200, you wrote:

Rapidly changing distros is mentioned in the response. What would classify
a rapidly changing distro. 

Rapidly changing would be Fedora/Mint/Ubuntu with their 6 month
release schedules, as opposed to Red Hat or the long term release
version of Mint/Ubuntu, or Debian.

The 6 month cycle is as short as you can get and still have any sort
of realistic amount of testing.

Take ubuntu is there six month release cycle
quick enough and even then they still wont have the latest versions of
software.

But they will have new enough versions of languages and libraries so
that you can easily compile almost anything else (that is either not
in the distribution, or not new enough).

New versions of software are being released daily and please correct if im
wrong but most distros do not release anything newer shortly there after
it coming out.

Actually, in most cases there is about a 2 month prior to release cut
off to allow for testing and bug fixing prior to release.

But in most cases this is not an issue.  The issue is with Red Hat or
any other LTS type release where the languages (either compiler or
interpreter) and libraries are several generations out of date.

It comes down to what you want/need.  There are parts of the software
industry where you don't want change, once you get something working
you want it kept that way.  For these people Red Hat and its
competitors are ideal, and they provide Red Hat with a very good
revenue stream.

For others they need newer compilers or libraries, so they need to put
up with the short support lifecyles of a Fedora/Mint/Ubuntu in order
to get those features (or luck into and freeze on a newly released Red
Hat/LTS).

Or, more likely in the HPC case, the people writing the software don't
necessarily need the newer stuff but happen to be using it because
what they do need is the newer features of the desktop environment (or
hardware support for laptop/etc)  and drag in a newer
Perl/gcc/gfortran/etc. as a side effect of that need.

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

[Beowulf] Intel MXC

[Gerald Henriksen]

Intel will introduce MXC, which they are calling next generation
optical connector with speeds up to 1.6 Terabits per second at Intel
Developer Forum (Sept. 10 to 12).

Protocols supported include Infiniband, Ethernet and PCI-Express.

http://www.pcworld.com/article/2046676/intel-proposes-new-standard-to-light-up-data-transfers.html#tk.rss_all

https://intel.activeevents.com/sf13/connect/search.ww#loadSearch-searchPhrase=MXCsearchType=sessiontc=0sortBy=abbreviationSortp=

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

[Beowulf] IBM announcement

[Gerald Henriksen]

Not much in the way of actual information yet but IBM, NVIDIA, and
others have formed the OpenPOWER consortium:

http://www-03.ibm.com/press/us/en/pressrelease/41684.wss

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Bolts of Thunder and Upgraded desktop interconnect silicon....

[Gerald Henriksen]

On Mon, 10 Jun 2013 18:57:29 -0400, you wrote:

So a company based out of Cupertino mentioned using this silicon in a
revamp of their MacPro line today...

http://blogs.intel.com/technology/2013/06/video-creation-bolts-ahead-%E2%80%93-intel%E2%80%99s-thunderbolt%E2%84%A2-2-doubles-bandwidth-enabling-4k-video-transfer-display-2/

we appear to have a second version of a 20GB/s consumer connection
(latency unknown), and yet this search:

https://www.google.com/search?q=linux+thunderbolt+interconnect

does not really go anywhere cool like a github or kernel.org repo

Any qualified folks know where this thunderbolt stuff is all heading
and are able to talk in public?

Thunderbolt so far is a very niche product that almost everybody
appears to be ignoring.

The cables are expensive - starting in the $40 range for a 1 m cable.

Actual devices are hard to come by, and tend towards the expensive -
raid boxes, and a couple of docks in the $300 range.

This might change later this year when the new Mac Pro is released as
the closed box nature of it will force/encourage any expansion to be
external.

But for now it looks like another Firewire in terms of market
acceptance.

Until Intel (with or without the help of Apple) can get the cable
prices down and some more common affordable peripherals available -
like a single drive enclosure - the market will likely continue to
ignore Thunderbolt.
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Oracle sells Lustre to Xyratex

[Gerald Henriksen]

On Wed, 20 Feb 2013 15:42:39 +0100, you wrote:


kraut 
http://www.heise.de/newsticker/meldung/Oracle-verkauft-Cluster-Dateisystem-Lustre-an-Xyratex-1806924.html
 /kraut

In English from the purchaser:

http://www.xyratex.com/news/blog/xyratex-assumes-ownership-lustre%C2%AE-and-related-assets
___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Xeon Phi out as well

[Gerald Henriksen]

Tom's Hardware has a report on the Intel presentation:

http://www.tomshardware.com/reviews/xeon-phi-larrabee-stampede-hpc,3342.html

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] Digital Image Processing via HPC/Cluster/Beowulf - Basics

[Gerald Henriksen]

On Tue, 6 Nov 2012 02:16:30 +0100, you wrote:

Jim as someone who produced games, this is not how it works for most  
movies/animations/commercials where graphics work is needed.

Note that most movies get editted as well in the same way (each  
nation usually has different requirements).

First of all - the budget is rather tiny. So that means using  
hardware you would probably yourself not consider using at NASA.

Not quite.

Avator - 10,000 square feet, 4,000 machines, 40,000 processors just
for the renderfarm at Weta:

http://www.datacenterknowledge.com/archives/2009/12/22/the-data-crunching-powerhouse-behind-avatar/

Then add in all the machines to support the artists...

You can also get some ideas of the non-renderfarm requirements at
http://en.wikipedia.org/wiki/Avatar_(2009_film)#Visual_effects


Secondly the calculation power required to produce movies is rather  
limited. Even a graphics card from years ago has more
capabilities than any artist single handed can design for. The real  
limitation is the amount of graphics a single person can produce.

Producing a head requires the work of an artist specialized in that,  
for a full month. Rendering that is a manner of minutes at
a single core. We speak of a CPU.

A whole animation in full blown HD gets rendered just as quick.

Not quite.  Articles typically mention that the time to render a frame
of film has remainted pretty constant over the last decade, any speed
increases in the hardware being used to execute more complex
algorithms.  The move from 24fps to 48fps (ie The Hobbitt) isn't going
to help matters either.

You can render the next episodes of the upcoming starwars movies deep  
at a single socket i7 pretty quick, which includes
testrun after testrun.

I am sure Disney would like to speak to you if you could accomplish
this, it would save them a significant amount of money.

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] cluster building advice?

[Gerald Henriksen]

On Wed, 26 Sep 2012 23:59:13 +0200, you wrote:

Yes easily.

Google for what linus posted there and what i posted there in code  
around 2007 already.

Where i showed how f'ed up GCC was, where it basically modified some  
simple code sample
to something ugly slow, instead of creating a CMOV instruction.

Basically it had the habit to do a comparision, if above zero then  
jump to the end of the function, execute 2 instructions there and  
jump bad.
Real real slow... ...getting the full stall of the pipeline when  
mispredicted.
Other compilers were on that code sample factor 30 faster of course.

Even my simple 32 bits C code it lost 35% there or so. And at C++  
codes it was factor 2+ slower for the bitboard chess engines
as some authors reported.

Linus posted something like: there is no excuse to not generate CMOV  
instructions now that also intel core2 can execute them fast.

It was not a gcc issue.

The distributions deliberately disabled CMOV in their i386 based
releases because there were still processors in common use that did
not support the CMOV instruction.

Fedora moved to CMOV only processor support with Fedora 12 (i686.rpm
packages), and RHEL moved to CMOV only support with RHEL 6.

Note that this was with gcc 4.4.2

That's why i wonder about all those distro's still just supporting  
gcc 4.4 and/or 4.5

Because the target markets specifically don't want change.  For most
people/companies the hardware is more powerful than is needed, or the
CPU isn't the bottleneck, so needlessly going through updates that
introduce new bugs or require updates to source code because the
compiler is stricter/re-interprets the standard is not something to be
done every 6 to 12 months.

And even for clusters, where CPU performance does matter, most people
using clusters will likely be using one of the commercial compilers
that generate better code than gcc anyway.

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Re: [Beowulf] cluster building advice?

[Gerald Henriksen]

On Wed, 26 Sep 2012 10:54:01 +0200, you wrote:

There's nearly nothing there at that link.

Just a handful of SRPMS.

All of the SRPMS for SL are there.

My point of openfabrics is: most people build a cluster in order to  
be have more performance
than a single machine can give. Not seldom that's also with latest  
hardware (not in my case
but i expect most beowulf aren't using old hardware).

To get performance you want latest compilers and pretty much latest  
stable kernel therefore
as a minimum requirement.

Not necessarily.  You want something that is stable and well tested,
which will complete the work you give it.  This is not necessarily the
latest and greatest.

But let's start with the performance issue.

Even trying to compile GCC 4.7.0 both 32 and 64 bits in SL was a  
problem though it had been
released for months.

Not difficult.

Googled RHEL gcc 4.7 and discovered that Red Hat has something called
Red Hat Developer Toolset which is gcc 4.7 for RHEL 5 and 6.

Went to ftp.redhat.com and found:

ftp://ftp.redhat.com/redhat/linux/enterprise/6Workstation/en/RHDevToolset/SRPMS/

Downloaded and installed the SRPMS, installed the dependencies via
yum, and built the RPMS. 

___
Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf