Software revolution?...

[Jan Panteltje]

On a sunny day (Fri, 11 Dec 2020 11:16:00 +1100) it happened Clifford Heath
<no.spam@please.net> wrote in <6TyAH.45217$vr5.30064@fx48.iad>:

On 11/12/20 4:16 am, Jan Panteltje wrote:
Water is not the big problem, look what Saudia Arabia did:
https://en.wikipedia.org/wiki/Water_supply_and_sanitation_in_Saudi_Arabia

Saudi groundwater extraction is a disaster. Natural surface water oases
have vanished; 80% of all the estimated ancient groundwater has been
extracted and the natural water table lowered by up to 150 metres. They
haven\'t invested heavily in solar (that could drive de-sal), and when
the oil dries up their investments in gold Rolls Royces won\'t save them.

https://en.wikipedia.org/wiki/Water_supply_and_sanitation_in_Saudi_Arabia#Fossil_groundwater

CH

Of course, but I was referring to desalination, one paragraph lower:
https://en.wikipedia.org/wiki/Water_supply_and_sanitation_in_Saudi_Arabia#Desalination
it clearly is the solution when groundwater levels are too low.

 

[Jan Panteltje]

On a sunny day (Thu, 10 Dec 2020 13:01:11 -0700) it happened Don Y
<blockedofcourse@foo.invalid> wrote in <rqtuqo$f3$1@dont-email.me>:

On 12/10/2020 10:47 AM, Jan Panteltje wrote:
On a sunny day (Thu, 10 Dec 2020 10:26:41 -0700) it happened Don Y
blockedofcourse@foo.invalid> wrote in <rqtloo$ptv$1@dont-email.me>:

On 12/10/2020 10:16 AM, Jan Panteltje wrote:
On a sunny day (Thu, 10 Dec 2020 09:29:53 -0700) it happened Don Y
blockedofcourse@foo.invalid> wrote in <rqtieg$s2r$1@dont-email.me>:

On 12/10/2020 8:21 AM, Jan Panteltje wrote:

Not only are we dependant on water for our own biological needs, but,
imagine what happens when sanitation fails in a \"heavily\" populated
area.

Yes
I live close to the sea.
been experimenting with desalination, heating salty water..
You can do that with solar, but if you still have electrickety then with that.
But it takes a lot of power.
On boats it is often done with high pressure pumps and filters.
About 1000$ for a 12V set IIRC.
You could use solar panels to charge the batteries.
https://www.cruisingworld.com/desalination-decisions-watermakers/
For emergency (say in a liferaft) there exist manual pumps.

Catching rain is cheaper...

When you only get 11 inches of precipitation, annually, catching rain is
a dubious proposition! Likewise, when the RH hovers around 15%, capturing
dew is just as improbable.

And, water isn\'t convenient to store in the quantities people need
to survive (large volumes).

Really?
https://en.wikipedia.org/wiki/List_of_tallest_dams_in_the_United_States
Anyways when all cars go \'lectric the tankers can be re-purposed...
Water is not the big problem, look what Saudia Arabia did:
https://en.wikipedia.org/wiki/Water_supply_and_sanitation_in_Saudi_Arabia
water can also easily be transported over long distances by pipes.
As long as you have electrickety that is
Most of planet earth is covered by H2O, plus food in the form of fishs etc.
Snow, glaciers, poles..

Those solutions require considerable time/effort to implement.
If you have been \"attacked\", you have folks looking for water (to drink)
within a few DAYS.

Yes, under attack, and all teh internet connected smart stuff will
not help pumps pushing water through teh pipes eather
So face it: Its all over!

A city water employee can visit all of the wells around the town in a
day or two; given that they have multiple employees (and getting the
wells back on-line would be an obvious priority), they could have folks at
every well before end of business, the next day.

But, that assumes the wells have been \"attacked\" by attempting to
confuse their controls. If they have been attacked by tainting
the ground water (in or near the intake), then the pumps would
have to be shut down while the extent -- and nature -- of the
contamination is determined.

[Note that we rely on Central Arizona Project water to recharge
our ground-water supply. This is a 300 mile, overground,
unprotected, man-made canal that transports water from the
northern part of the state. No idea what precautions are
taken against deliberate attacks on that source. And, of course,
even the ability to DETECT an attack doesn\'t mean the water can
be accessed in spite of the attack!]

https://en.wikipedia.org/wiki/Central_Arizona_Project

Yes, canals like that are as old as the Roman empire..
It says:
\"The canal loses approximately 16,000 acre-feet (5.2 billion gallons) of water each year to evaporation,\"
Why not use pipes?
OTOH it probably irrigates the land it crosses....

They also mention climate change (warming) will increase the evaporation.
Climate change (if it really is wat is told to the masses as ever increasing \'heating\')
should then lead to mass migration...
But I think climate has always changed, and is cyclic:
http://old.world-mysteries.com/alignments/mpl_al3b.htm
Note the Milankovich Cycles
All the way down in that link you see the stages of glacation...
The world has to sell, fear helps, now everybody a new electric car
better aircos, later (.... ) better heaters I suppose.
This reminds me of a fun movie I have seen where the \'star wreck\' travellers return to earth after a time warp and end up in a frozen world.
\"star wreck in the pirkinnin\"
is free for download I think.

 

[Don Y]

On 12/11/2020 12:20 AM, Jan Panteltje wrote:

On a sunny day (Thu, 10 Dec 2020 13:01:11 -0700) it happened Don Y
blockedofcourse@foo.invalid> wrote in <rqtuqo$f3$1@dont-email.me>:

On 12/10/2020 10:47 AM, Jan Panteltje wrote:
On a sunny day (Thu, 10 Dec 2020 10:26:41 -0700) it happened Don Y
blockedofcourse@foo.invalid> wrote in <rqtloo$ptv$1@dont-email.me>:

On 12/10/2020 10:16 AM, Jan Panteltje wrote:
On a sunny day (Thu, 10 Dec 2020 09:29:53 -0700) it happened Don Y
blockedofcourse@foo.invalid> wrote in <rqtieg$s2r$1@dont-email.me>:

On 12/10/2020 8:21 AM, Jan Panteltje wrote:

Not only are we dependant on water for our own biological needs, but,
imagine what happens when sanitation fails in a \"heavily\" populated
area.

Yes
I live close to the sea.
been experimenting with desalination, heating salty water..
You can do that with solar, but if you still have electrickety then with that.
But it takes a lot of power.
On boats it is often done with high pressure pumps and filters.
About 1000$ for a 12V set IIRC.
You could use solar panels to charge the batteries.
https://www.cruisingworld.com/desalination-decisions-watermakers/
For emergency (say in a liferaft) there exist manual pumps.

Catching rain is cheaper...

When you only get 11 inches of precipitation, annually, catching rain is
a dubious proposition! Likewise, when the RH hovers around 15%, capturing
dew is just as improbable.

And, water isn\'t convenient to store in the quantities people need
to survive (large volumes).

Really?
https://en.wikipedia.org/wiki/List_of_tallest_dams_in_the_United_States
Anyways when all cars go \'lectric the tankers can be re-purposed...
Water is not the big problem, look what Saudia Arabia did:
https://en.wikipedia.org/wiki/Water_supply_and_sanitation_in_Saudi_Arabia
water can also easily be transported over long distances by pipes.
As long as you have electrickety that is
Most of planet earth is covered by H2O, plus food in the form of fishs etc.
Snow, glaciers, poles..

Those solutions require considerable time/effort to implement.
If you have been \"attacked\", you have folks looking for water (to drink)
within a few DAYS.

Yes, under attack, and all teh internet connected smart stuff will
not help pumps pushing water through teh pipes eather
So face it: Its all over!

A city water employee can visit all of the wells around the town in a
day or two; given that they have multiple employees (and getting the
wells back on-line would be an obvious priority), they could have folks at
every well before end of business, the next day.

But, that assumes the wells have been \"attacked\" by attempting to
confuse their controls. If they have been attacked by tainting
the ground water (in or near the intake), then the pumps would
have to be shut down while the extent -- and nature -- of the
contamination is determined.

[Note that we rely on Central Arizona Project water to recharge
our ground-water supply. This is a 300 mile, overground,
unprotected, man-made canal that transports water from the
northern part of the state. No idea what precautions are
taken against deliberate attacks on that source. And, of course,
even the ability to DETECT an attack doesn\'t mean the water can
be accessed in spite of the attack!]

https://en.wikipedia.org/wiki/Central_Arizona_Project

Yes, canals like that are as old as the Roman empire..

The pertinent question is: how are they protected from (deliberate)
contaminants? Presumably, they could gate the water from entering
the muni water system. But, it\'s got to *go* SOMEWHERE!

It says:
\"The canal loses approximately 16,000 acre-feet (5.2 billion gallons) of water each year to evaporation,\"
Why not use pipes?

Likely cost. Some reservoirs are covered to minimize evaporation
(at a lower cost than a more hermetic solution). And, we \"store\"
lots of water in the ground.

But, folks want to ignore how dire our water situation is. Photo of
Lake Mead, source of the CAP water (note boat for reference scale
when judging drop in water level):

<https://s1.ibtimes.com/sites/www.ibtimes.com/files/2015/04/22/lake-mead-main.jpg>

If concern over water \"waste\" was high on the priority list, the 25-60%
home pool ownership would be targeted. A pool\'s water level typically
falls about an inch a week, on average. Even a small pool (400-500 sq ft)
loses more than 10,000G to evaporation, annually.

Add to that, transpiration losses in things like \"lawns\" (discouraged)
and wasteful irrigation (where the water is \"thrown\" over long distances
through the air -- 40% losses!) and there are many opportunities to
conserve that are missed.

Of course, letting people like *us* grow citrus is almost as bad as
letting the cotton growers operate! :<

> OTOH it probably irrigates the land it crosses....

 

[Jan Panteltje]

On a sunny day (Fri, 11 Dec 2020 01:18:47 -0700) it happened Don Y
<blockedofcourse@foo.invalid> wrote in <rqva1n$86a$1@dont-email.me>:

But, folks want to ignore how dire our water situation is. Photo of
Lake Mead, source of the CAP water (note boat for reference scale
when judging drop in water level):

https://s1.ibtimes.com/sites/www.ibtimes.com/files/2015/04/22/lake-mead-main.jpg

Yea that looks really bad...

If concern over water \"waste\" was high on the priority list, the 25-60%
home pool ownership would be targeted. A pool\'s water level typically
falls about an inch a week, on average. Even a small pool (400-500 sq ft)
loses more than 10,000G to evaporation, annually.

Add to that, transpiration losses in things like \"lawns\" (discouraged)
and wasteful irrigation (where the water is \"thrown\" over long distances
through the air -- 40% losses!) and there are many opportunities to
conserve that are missed.

Of course, letting people like *us* grow citrus is almost as bad as
letting the cotton growers operate! :

OTOH it probably irrigates the land it crosses....

We had, last summer here, a request not to spray your garden as water levels were low.
But basically The Netherlands where I am is largely below sea level, it is just that
the inflow of sweet water from the rivers (from Germany) was very low.
They are tinkering with that water level all the time, opening and closing locks.
I did some design work for that long ago, water measurement systems are
all over the place and connected to a central computer where they determine
what all the pumping stations should be doing etc.
It is an essential system, if it fails Amsterdam will flood for example.

 

[Clifford Heath]

On 11/12/20 6:02 pm, Jan Panteltje wrote:

On a sunny day (Fri, 11 Dec 2020 11:16:00 +1100) it happened Clifford Heath
no.spam@please.net> wrote in <6TyAH.45217$vr5.30064@fx48.iad>:

On 11/12/20 4:16 am, Jan Panteltje wrote:
Water is not the big problem, look what Saudia Arabia did:
https://en.wikipedia.org/wiki/Water_supply_and_sanitation_in_Saudi_Arabia

Saudi groundwater extraction is a disaster. Natural surface water oases
have vanished; 80% of all the estimated ancient groundwater has been
extracted and the natural water table lowered by up to 150 metres. They
haven\'t invested heavily in solar (that could drive de-sal), and when
the oil dries up their investments in gold Rolls Royces won\'t save them.

https://en.wikipedia.org/wiki/Water_supply_and_sanitation_in_Saudi_Arabia#Fossil_groundwater

CH

Of course, but I was referring to desalination

So was I. You need to have energy to drive it. They should have
unlimited solar, but while there\'s still oil, why bother? And when there
isn\'t, how are they gonna pay for it?

CH

 

[Jan Panteltje]

On a sunny day (Fri, 11 Dec 2020 22:38:45 +1100) it happened Clifford Heath
<no.spam@please.net> wrote in <9TIAH.27362$li2.20937@fx22.iad>:

On 11/12/20 6:02 pm, Jan Panteltje wrote:
On a sunny day (Fri, 11 Dec 2020 11:16:00 +1100) it happened Clifford Heath
no.spam@please.net> wrote in <6TyAH.45217$vr5.30064@fx48.iad>:

On 11/12/20 4:16 am, Jan Panteltje wrote:
Water is not the big problem, look what Saudia Arabia did:
https://en.wikipedia.org/wiki/Water_supply_and_sanitation_in_Saudi_Arabia

Saudi groundwater extraction is a disaster. Natural surface water oases
have vanished; 80% of all the estimated ancient groundwater has been
extracted and the natural water table lowered by up to 150 metres. They
haven\'t invested heavily in solar (that could drive de-sal), and when
the oil dries up their investments in gold Rolls Royces won\'t save them.

https://en.wikipedia.org/wiki/Water_supply_and_sanitation_in_Saudi_Arabia#Fossil_groundwater

CH

Of course, but I was referring to desalination

So was I. You need to have energy to drive it. They should have
unlimited solar, but while there\'s still oil, why bother? And when there
isn\'t, how are they gonna pay for it?

They are building 2 big nuclear plants and maybe some smaller ones for desalination:
https://www.world-nuclear.org/information-library/country-profiles/countries-o-s/saudi-arabia.aspx

IMO nuclear is the correct option when oil runs out for them, and it is basically the correct option everywhere,
Indeed hey will have to find an other way to make money. selling / exporting electrickety could be one.
I have read about ideas to get \'trickety from solar plants in Africa, have not really looked into it.
Anywhere where there is plenty sun should work.

There is fear for nuclear, but if you look at the wildlife where Tjernobyl happened, it is thriving.
Too bad they removed the youtube video showing the wildlife there (wolves I think it was).

And of course fear of the bombs that could perhaps be made using plutonium from those nuclear plants.

But thousands die each year in coal mines! So far very few did in nuclear mishaps.
Not even counting the victims from air pollution from burning coal and oil.

We should use nuclear propulsion to colonize space too.
Politics.
I hope China will have no fear to use it for that.
NASA\'s nuclear propulsion experiments were scrapped long ago.
Fear

 

[Don Y]

On 12/11/2020 3:31 AM, Jan Panteltje wrote:

On a sunny day (Fri, 11 Dec 2020 01:18:47 -0700) it happened Don Y
blockedofcourse@foo.invalid> wrote in <rqva1n$86a$1@dont-email.me>:

But, folks want to ignore how dire our water situation is. Photo of
Lake Mead, source of the CAP water (note boat for reference scale
when judging drop in water level):

https://s1.ibtimes.com/sites/www.ibtimes.com/files/2015/04/22/lake-mead-main.jpg

Yea that looks really bad...

If concern over water \"waste\" was high on the priority list, the 25-60%
home pool ownership would be targeted. A pool\'s water level typically
falls about an inch a week, on average. Even a small pool (400-500 sq ft)
loses more than 10,000G to evaporation, annually.

Add to that, transpiration losses in things like \"lawns\" (discouraged)
and wasteful irrigation (where the water is \"thrown\" over long distances
through the air -- 40% losses!) and there are many opportunities to
conserve that are missed.

Of course, letting people like *us* grow citrus is almost as bad as
letting the cotton growers operate! :

OTOH it probably irrigates the land it crosses....

We had, last summer here, a request not to spray your garden as water levels were low.
But basically The Netherlands where I am is largely below sea level, it is just that
the inflow of sweet water from the rivers (from Germany) was very low.

It is unlawful to \"waste\" water, here -- though any evidence of this
is usually met with education instead of financial penalties ($250 - $2500 per
incident).

You can\'t let water leave your property and travel onto public property
(alleys or streets) or onto another person’s property. When a lot is
planned, you must show how you will keep all water (incl rainfall)
ON that property.

You can\'t let water to pond to a depth greater than 1/4 inch or over a
cumulative surface area greater than 150 square feet on any street or
parking lot.

You can\'t \"wash\" driveways, sidewalks, parking areas, or other surface
areas with an open hose or spray nozzle attached to an open hose. In
other places where I\'ve lived, it was commonplace to \"hose down the
driveway\", etc.

You can\'t operate a misting system (spraying a fine water mist into
the air for the purpose of cooling that outdoor space -- common at
cafes and outdoor eateries, as well as in private yards) in *unoccupied*
non-residential areas.

You can\'t operate an irrigation system with a broken head or emitter, or
with a head(s) that is spraying more than 10% of the spray on a street,
parking lot, or sidewalk (where it obviously isn\'t serving any \"irrigation\"
purpose)

You must repair \"controllable leaks\", including broken sprinkler heads,
leaking valves, or a leaking faucet. Our (electronic) water meters detect
likely leaks and can report this condition to the (electronic) meter reader.

You must meet the 50% rainwater harvesting requirement for landscape
irrigation. Note that this need not be \"active\" measures.

In practice, no one is cited for these. But, you\'ll often hear of
any abuses from your neighbors.

Note that it is *legal* to \"drain\" your swimming pool onto the roadway.
In fact, as we have no storm sewers, this is the only way to dispose
of that water. The point being that such use is regular maintenance
and not \"waste\" (which assumes you consider pools not to be inherently
wasteful!)

 

[Sjouke Burry]

On 11.12.20 8:20, Jan Panteltje wrote:

On a sunny day (Thu, 10 Dec 2020 13:01:11 -0700) it happened Don Y
blockedofcourse@foo.invalid> wrote in <rqtuqo$f3$1@dont-email.me>:

On 12/10/2020 10:47 AM, Jan Panteltje wrote:
On a sunny day (Thu, 10 Dec 2020 10:26:41 -0700) it happened Don Y
blockedofcourse@foo.invalid> wrote in <rqtloo$ptv$1@dont-email.me>:

On 12/10/2020 10:16 AM, Jan Panteltje wrote:
On a sunny day (Thu, 10 Dec 2020 09:29:53 -0700) it happened Don Y
blockedofcourse@foo.invalid> wrote in <rqtieg$s2r$1@dont-email.me>:

On 12/10/2020 8:21 AM, Jan Panteltje wrote:

Not only are we dependant on water for our own biological needs, but,
imagine what happens when sanitation fails in a \"heavily\" populated
area.

Yes
I live close to the sea.
been experimenting with desalination, heating salty water..
You can do that with solar, but if you still have electrickety then with that.
But it takes a lot of power.
On boats it is often done with high pressure pumps and filters.
About 1000$ for a 12V set IIRC.
You could use solar panels to charge the batteries.
https://www.cruisingworld.com/desalination-decisions-watermakers/
For emergency (say in a liferaft) there exist manual pumps.

Catching rain is cheaper...

When you only get 11 inches of precipitation, annually, catching rain is
a dubious proposition! Likewise, when the RH hovers around 15%, capturing
dew is just as improbable.

And, water isn\'t convenient to store in the quantities people need
to survive (large volumes).

Really?
https://en.wikipedia.org/wiki/List_of_tallest_dams_in_the_United_States
Anyways when all cars go \'lectric the tankers can be re-purposed...
Water is not the big problem, look what Saudia Arabia did:
https://en.wikipedia.org/wiki/Water_supply_and_sanitation_in_Saudi_Arabia
water can also easily be transported over long distances by pipes.
As long as you have electrickety that is
Most of planet earth is covered by H2O, plus food in the form of fishs etc.
Snow, glaciers, poles..

Those solutions require considerable time/effort to implement.
If you have been \"attacked\", you have folks looking for water (to drink)
within a few DAYS.

Yes, under attack, and all teh internet connected smart stuff will
not help pumps pushing water through teh pipes eather
So face it: Its all over!

A city water employee can visit all of the wells around the town in a
day or two; given that they have multiple employees (and getting the
wells back on-line would be an obvious priority), they could have folks at
every well before end of business, the next day.

But, that assumes the wells have been \"attacked\" by attempting to
confuse their controls. If they have been attacked by tainting
the ground water (in or near the intake), then the pumps would
have to be shut down while the extent -- and nature -- of the
contamination is determined.

[Note that we rely on Central Arizona Project water to recharge
our ground-water supply. This is a 300 mile, overground,
unprotected, man-made canal that transports water from the
northern part of the state. No idea what precautions are
taken against deliberate attacks on that source. And, of course,
even the ability to DETECT an attack doesn\'t mean the water can
be accessed in spite of the attack!]

https://en.wikipedia.org/wiki/Central_Arizona_Project

Yes, canals like that are as old as the Roman empire..
It says:
\"The canal loses approximately 16,000 acre-feet (5.2 billion gallons) of water each year to evaporation,\"
Why not use pipes?
OTOH it probably irrigates the land it crosses....

They also mention climate change (warming) will increase the evaporation.
Climate change (if it really is wat is told to the masses as ever increasing \'heating\')
should then lead to mass migration...
But I think climate has always changed, and is cyclic:
http://old.world-mysteries.com/alignments/mpl_al3b.htm
Note the Milankovich Cycles
All the way down in that link you see the stages of glacation...
The world has to sell, fear helps, now everybody a new electric car
better aircos, later (.... ) better heaters I suppose.
This reminds me of a fun movie I have seen where the \'star wreck\' travellers return to earth after a time warp and end up in a frozen world.
\"star wreck in the pirkinnin\"
is free for download I think.

Yep. Just done it.

 

[Sjouke Burry]

On 11.12.20 11:31, Jan Panteltje wrote:

On a sunny day (Fri, 11 Dec 2020 01:18:47 -0700) it happened Don Y
blockedofcourse@foo.invalid> wrote in <rqva1n$86a$1@dont-email.me>:

But, folks want to ignore how dire our water situation is. Photo of
Lake Mead, source of the CAP water (note boat for reference scale
when judging drop in water level):

https://s1.ibtimes.com/sites/www.ibtimes.com/files/2015/04/22/lake-mead-main.jpg

Yea that looks really bad...

If concern over water \"waste\" was high on the priority list, the 25-60%
home pool ownership would be targeted. A pool\'s water level typically
falls about an inch a week, on average. Even a small pool (400-500 sq ft)
loses more than 10,000G to evaporation, annually.

Add to that, transpiration losses in things like \"lawns\" (discouraged)
and wasteful irrigation (where the water is \"thrown\" over long distances
through the air -- 40% losses!) and there are many opportunities to
conserve that are missed.

Of course, letting people like *us* grow citrus is almost as bad as
letting the cotton growers operate! :

OTOH it probably irrigates the land it crosses....

We had, last summer here, a request not to spray your garden as water levels were low.
But basically The Netherlands where I am is largely below sea level, it is just that
the inflow of sweet water from the rivers (from Germany) was very low.
They are tinkering with that water level all the time, opening and closing locks.
I did some design work for that long ago, water measurement systems are
all over the place and connected to a central computer where they determine
what all the pumping stations should be doing etc.
It is an essential system, if it fails Amsterdam will flood for example.

Hmmm... That would be terrible... Or maybe not?

 

[Jeroen Belleman]

On 2020-12-11 19:06, Sjouke Burry wrote:

On 11.12.20 11:31, Jan Panteltje wrote:
[...] if it fails Amsterdam will flood for example.


Hmmm... That would be terrible... Or maybe not?

Why not?

Jeroen Belleman (Who has two siblings living there)

 

[three_jeeps]

On Thursday, December 10, 2020 at 8:44:35 PM UTC-5, Don Y wrote:

On 12/10/2020 4:07 PM, three_jeeps wrote:
There are architectures that address parallelism...VLIW, MPP, Vector
machines, all exploit parallelism at different levels of granularity. It is
not only about execution concurrency it is about the data dependencies
associated with concurrency.
It is *that* which most \"coders\" are incapable of ferreting out.
Their thinking is of the \"first do THIS, then do THAT\" variety.

There are also opportunities for rethinking chosen algorithms
to avoid artificial dependencies.

But, unless you have that as a *goal*, you are unlikely to
expend the effort -- nor have the experience -- in doing so.
As a \"coder\", you\'ve probably not had much formal theory in
developing algorithms and just see applications as exercises
in generating equivalent \"computer prose\" from \"English prose\".

I agree, coders don\'t consider that aspect in their work, partly because compiler optimizations try to understand the data dependencies within the context of the program itself but also in consideration of the capabilities of the target hw arch. Loop unrolling is a common optimization, at the hw arch, out of order ex units, register coloring, etc. are techniques to extract as much parallelism as possible at the fine level of granularity of instruction streams. Tomasulu\'s algorithm is a classic example and was employed on the IBM 360. There are a host of others in the years that followed.
Programming a vector machine necessitated 1) restructuring your code base a certain way then 2) running it through a \'vectorizer\' which looked for data dependencies and tries to restructure the execution stream.
Of course, there are other aspects of machine architecture that work against parallelization. One approach to increase speedup is super pipelining or deep pipelining. Works great on \'well conditioned\' code but given a data dependency in the pipeline, flushing the execution stream and starting over works against the anticipated performance gains. This gets into the discussion of synthetic benchmarks and profiling.....
j

 

[Don Y]

On 12/11/2020 12:36 PM, three_jeeps wrote:

On Thursday, December 10, 2020 at 8:44:35 PM UTC-5, Don Y wrote:
On 12/10/2020 4:07 PM, three_jeeps wrote:
There are architectures that address parallelism...VLIW, MPP, Vector
machines, all exploit parallelism at different levels of granularity. It
is not only about execution concurrency it is about the data
dependencies associated with concurrency.
It is *that* which most \"coders\" are incapable of ferreting out. Their
thinking is of the \"first do THIS, then do THAT\" variety.

There are also opportunities for rethinking chosen algorithms to avoid
artificial dependencies.

But, unless you have that as a *goal*, you are unlikely to expend the
effort -- nor have the experience -- in doing so. As a \"coder\", you\'ve
probably not had much formal theory in developing algorithms and just see
applications as exercises in generating equivalent \"computer prose\" from
\"English prose\".

I agree, coders don\'t consider that aspect in their work, partly because
compiler optimizations try to understand the data dependencies within the
context of the program itself but also in consideration of the capabilities
of the target hw arch. Loop unrolling is a common optimization, at the hw
arch, out of order ex units, register coloring, etc. are techniques to
extract as much parallelism as possible at the fine level of granularity of
instruction streams. Tomasulu\'s algorithm is a classic example and was
employed on the IBM 360. There are a host of others in the years that
followed.

There are two conflicting guidelines operating, here.

On the one hand, you *don\'t* want to have folks thinking (in detail)
about the hosting hardware when building an algorithm. You want
the algorithm to be as portable as possible AND don\'t want to
bind the implementation to a \"current hosting choice\" (which may
prove to be silly -- or contrary! -- in future hardware developments)

On the other, you don\'t want folks abrogating their responsibility
to develop effective solutions. You don\'t expect a hardware designer to
let the tools decide the ratings of every component in the design
and NOT consider the consequences of particular design decisions.

Programming a vector machine necessitated 1) restructuring your code base a
certain way then 2) running it through a \'vectorizer\' which looked for data
dependencies and tries to restructure the execution stream.

I think the parallelism that eludes coders is more fundamental than
optimizing execution orderings. They seem to implement algorithms
in serial pieces instead of thinking about the demands of the
application, itself.

I designed a device to \"sanitize\" disk drives. The stated goal of
the device is to ensure no \"previous contents\" of media remain on
those drives. And, to verify the storage integrity of each drive as
well as its dynamic performance.

As processing the entire surface of a modern disk is a time consuming
operation, you want to be able to process MANY at a time.

A candidate drive needs to be mounted to a carrier (a mechanical
requirement of the hardware that I was using). Then, installed
into the system. Once installed, it needs to be recognized
(make, model, serial number, capacity, etc.) and a permanent
record made of its existence (and the processing which follows).

If the drive is below a certain size/performance threshold (based on
observations of past drives of this make/model) AND if there are
alternate means of destroying the drive (e.g., shredder), then
the drive should be rejected and marked for that \"other\" processing.

The entire surface of the drive should be overwritten to obliterate
past contents (I have different algorithms for doing this -- though
that is largely a political choice to satisfy the drives\' original
owner(s)). The medium should then be verified as having no
vestige of the previous contents.

While doing this, a measure of the drive\'s performance should be
ascertained (read/write faults, remapped sectors, bandwidth, etc.)
to determine if the drive is suitable for reuse AFTER sanitizing.

Reading through these requirements, there are very few serial
steps dictated.

E.g., there is no reason you can\'t start processing drive #1 as
soon as it is introduced to the machine -- why wait around
for drives #2 - #60? (It takes about 2 minutes to physically
prepare a drive so drive #1 will have been sitting idle for
almost 2 hours before drive #60 was installed!)

There\'s no reason you have to wait for the database to record
the drive\'s identity/characteristics before starting processing.

There\'s no reason you have to make your \"disposal/processing\"
decision before starting processing (given that there may be
59 other drives being processed/assessed while this one is
being processed, that decision could come WHILE you\'re
processing this drive instance)

There\'s no reason you have to wait for the drive to finish processing
before coming to a decision regarding its integrity or performance.

There\'s no reason you can\'t \"eject\" the drive before recording
the results of that processing in the database.

If you naively approach the problem as a bunch of SERIAL steps,
you find that you\'ll end up with a device that considerably
underperforms an implementation that exploits opportunities for
parallelism.

But, that means restructuring the design with that in mind;
there\'s no \"optimizer\" that is going to make such sweeping
assessments of the process!

This \"smarter\" approach lets me keep the device AND THE OPERATOR
busy, continuously (he\'s inserting unprocessed drives while
removing finished drives, one after another).

Of course, there are other aspects of machine architecture that work against
parallelization. One approach to increase speedup is super pipelining or
deep pipelining. Works great on \'well conditioned\' code but given a data
dependency in the pipeline, flushing the execution stream and starting over
works against the anticipated performance gains. This gets into the
discussion of synthetic benchmarks and profiling..... j

But these are still constrained to the actual chosen implementation.
E.g., You can\'t automagically redefine my choice of data structures
in the memory layout to better exploit cache line locality.

[This returns to the competing goals stated initially, above;
do you want the developer to be THAT aware of hardware details
to be able to exploit those mechanisms that can\'t magically be
exploited FOR him?]

Regardless, if you\'re looking to write performant code, you can\'t
just sit down and write \"serially\".

 

[Jan Panteltje]

On a sunny day (Fri, 11 Dec 2020 19:06:00 +0100) it happened Sjouke Burry
<burrynulnulfour@ppllaanneett.nnll> wrote in
<5fd3b508$0$18018$e4fe514c@textnews.kpn.nl>:
------------------------------------^^^^^
....

It is an essential system, if it fails Amsterdam will flood for example.


Hmmm... That would be terrible... Or maybe not?

Well I grew up there, it is a nice place.
https://en.wikipedia.org/wiki/Flood_control_in_the_Netherlands

Amsterdam is about 2 meters below sea level.
For the farmers around the city the ground water level has to be controlled to a few cm.
Too high an the crops start rotting, too low and the crops die.
To accomplish that there are pumping stations and water level sensors.

Same goes for large parts of the Netherlands, dikes, pumps...
much land was reclaimed from the sea.


I see you post from KPN, you would not happen to live in Amsterdam / Netherlands ;-)?
What did they do to you??

Sjouke is Fries?