[PLUG] HP G62 AMD Dual Core laptop...

Michael Christopher Robinson michael at robinson-west.com
Tue May 28 00:09:26 UTC 2019


On Mon, 2019-05-27 at 14:04 -0700, Ben Koenig wrote:
> First, you can't upgrade the CPU on a laptop the way you expect. It
> is
> NOT a standard AM3+ socket. Check HP's documentation for that, I
> can't
> speak for their upgradeability.
> 
> Second, you are grossly overthinking your question.
> 
> I don't know what Labview is. So I take the following steps:
> 
> Q: Does it support Linux?
> A: search duckduckgo for the terms "labview linux"
> 
> found the following pages on the National Instruments website:
> https://knowledge.ni.com/KnowledgeArticleDetails?id=kA00Z000000kG1MSAU&l=en-US
> https://www.ni.com/download/labview-development-system-2017/6963/en/
> 
> Conclusion: Yes, what you are looking to do is possible in CentOS 7.
> 
> Don't expect me to judge the quality of their installation process or
> provide any further advice, I'm not a labview user.
> 
> -Ben

Thank you Ben.  A dual core AMD processor should be more than adequate
as a Pentium III seems to be sufficient looking at the National
Instruments how do I install Labview 2017 Linux documentation.  
I don't think hardware acceleration is necessary for the graphics and
only 16 bit color is required (also in the National Instruments install
documentation).  

I'm still curious if any vendor sells a laptop or rackmount server made
for Linux and CentOS 7 with Labview 2017 Linux installed.  It is going
to be expensive, Labview Linux is a $6k program apparently, but network
analyzers which labview works with aren't exactly cheap anyways.  A
modern network analyzer from Keytronics is going to cost between $20k-
$60k running Windows 10 embedded.  That seems atrocious, but the
product lines these analyzers support are generally significant money
makers.  Amortized out over 4-5 years, an analyzer is part of the cost
of making these products.  For making MRI equipment that plugs into MRI
machines, it has to be tuned properly.  To give you an idea, keeping an
analyzer properly calibrated so that you can do RF tuning with it costs
$1000/year which is another cost of doing business.  If a suite is a
$16k proposition for the company paying us to manufacture it, and we do
3-4 hardware suites a quarter, the cost of doing business starts to
look more bearable.  There are employees to pay as well as equipment to
buy in that $16k a suite.

Spending $7k on the Linux laptop that runs Labview seems reasonable to
me.  Still, I prefer to have a vendor put the laptop and Labview
together if it costs that much so that I can be confident the
laptop/rackmount server is well engineered for the task.  It's the
hardware choices that really make a difference and how the hardware is
put together all the way down to how the processor is cooled. 
Reliability and maintainability since this laptop/rackmount server is
going to be part of a production line means everything.  Another thing
that matters, this needs to be a tightly controlled computer.  This
computer should go into the production line to gather data from the
analyzer and stay in the production line without failing
for at least 1 year to 5 years.  With Linux, you can be a server too
and provide a VPN gateway so that the network connecting the analyzer
to the computer can be completely isolated most of the time (this is
a controlled manufacturing environment).  A dhcp server for the network
between analyzer and server/labview computer seems reasonable, but in
general the analyzers and server could be statically configured into an
RFC 1918 private subnet.  

Labview by the way is for automation of instruments and it is put out
by National Instruments.  Think, the external labview equipped computer
can actually remote control an analyzer, gathering data from that
analyzer is one thing Labview should be able to do, but data gathering
is not everything Labview can do.  Maybe Libreoffice Calc can pull said
data from a database into a spreadsheet that has the company logo on it
and then you print this form out and initial the readings taken for
traceability reasons.  MRI paddles are medical grade equipment, so
there are special documentation requirements when you do the
manufacturing and other requirements as well.  I'm trying to improve
the data gathering by having a computer collect the numbers instead of
having a physical human being try to read them off of a screen and hand
write them on the form.  Do I need full blown labview to accomplish
this?  I don't know for sure.  The tuning is critical for this
equipment or it simply won't work where failed builds are an expensive
proposition that hurt my company's bottom line.  The numbers that get
recorded need to be accurately recorded, no inverted digits or misread
digits.

An AMD Athlon II by today's standards is obsolete and an HP G62 is
arguably junk at this point in time.  I have a better laptop, but I
need to keep it for myself at home until I certify as a Javascript,
HTML5, and CSS3 coder and my personal data is on it.  This G62 is my
old laptop from my college days qith zero personal data on it.  My
employer should provide equipment, but I'm fairly new to the company
where the company is in the black, but barely in the black.  I
don't want to rock the boat by asking for new equipment let alone Linux
equipment until I have the proper business case for it.  I'll
experiment with Linux on Junk for now and hopefully figure out a case
for good Linux equipment including what good Linux equipment is. 
Overkill is out of the question, so I say rackmount server very
lightly.  Yes there is a rack and a server room, no that is not clearly
the direction we should go.  Oh, and the old cloud is implemented on a
Linux server that is in the server room which nobody knows how to
maintain...  the maintainer took another job and is hard to reach for
information on how to maintain it.  But I am digressing, apologies for
that.

Sorry for the long post, it's an awful lot to take in, but there is a
lot to think about if I'm going to actually make things better at this
company.  Let me rephrase that, if I'm going to come up with a plan
that the owner can execute on to make things better at this company. 
Come up with a good plan and convince the owner to follow it, not easy
considering that this plan will cost money.

    --  Michael C. Robinson




More information about the PLUG mailing list