| Topic: |
Science > Physics |
| User: |
"Uncle Al" |
| Date: |
15 Oct 2003 07:45:15 PM |
| Object: |
Parity Eotvos experiment, update |
If anybody cares...
http://www.mazepath.com/uncleal/qz.pdf
Fig. 2, Table VI.
We recently crunched quartz to 3.3 quadrillion atoms or a 43 micron
diameter ball - 2.8 GHz Xeon and 1.8 GHz Opteron CPUs in Linux gave
identical data - up from 0.33 Quad and 20 microns. The graphic fit to
parity divergence, CHI, tweaked in the fourth decimal place, slope and
intercept. 15 Quads is cooking right now, or 71 microns diameter.
The native 32-bit source code is being ported to 64-bit and is being
parallelized for a dual-Opteron box. If this goes through, we go for
42 quadrillion atoms or a 100 micron diameter ball (1/10 millimeter)
in a CPU-month. That is a classical physics mass to even the most
skeptical pundit. No surprises vs. scale can hide.
If we can get into an 8-Opteron box, 4 CPU-weeks does 205 Quads or a
170 micron diameter ball. In a world dripping with *****, the
Gifted across the globe still shout defiance. Appreciation goes out
to four code poets in four different countries who offered to toss
some unused CPU-time Uncle Al's way - and boasted that the code could
run faster.
If anybody has a big cluster not otherwise engaged or needs a burn-in
for new equipment, I expect to have a program that will flex every
transistor in every CPU but not bother other components. All I ask in
return is the data generated. An Opteron running 32-bit code
massively outperforms a Xeon running at 156% of Opteron clock speed.
Windows is 40% slower than Linux in all iron and we abandoned any
thought of it.
No more "Intel Inside" for Uncle Al. Go AMD.
--
Uncle Al
http://www.mazepath.com/uncleal/eotvos.htm
(Do something naughty to physics)
.
|
|
| User: "Bilge" |
|
| Title: Re: Parity Eotvos experiment, update |
17 Oct 2003 06:47:12 PM |
|
|
Uncle Al:
If we can get into an 8-Opteron box, 4 CPU-weeks does 205 Quads or a
170 micron diameter ball. In a world dripping with *****, the
Gifted across the globe still shout defiance. Appreciation goes out
to four code poets in four different countries who offered to toss
some unused CPU-time Uncle Al's way - and boasted that the code could
run faster.
Aside from all of the neato, gee-whiz computations, have you
considered computing the actual cost of doing the experiment
including an analysis of any trade-off in precision vs. cost
for the calculations you've already done? Someone who would
consider doing the experiment might like an accurate figure
for the cost of setting a limit and knowing what that limit
is, broken down into each aspect of the experiment.
.
|
|
|
| User: "Uncle Al" |
|
| Title: Re: Parity Eotvos experiment, update |
17 Oct 2003 07:18:14 PM |
|
|
Bilge wrote:
Uncle Al:
If we can get into an 8-Opteron box, 4 CPU-weeks does 205 Quads or a
170 micron diameter ball. In a world dripping with *****, the
Gifted across the globe still shout defiance. Appreciation goes out
to four code poets in four different countries who offered to toss
some unused CPU-time Uncle Al's way - and boasted that the code could
run faster.
Aside from all of the neato, gee-whiz computations, have you
considered computing the actual cost of doing the experiment
including an analysis of any trade-off in precision vs. cost
for the calculations you've already done? Someone who would
consider doing the experiment might like an accurate figure
for the cost of setting a limit and knowing what that limit
is, broken down into each aspect of the experiment.
The cost of a parity Eotvos experiment in quartz is that of a
composition Eotvos experiment in anything. The major cost of a test
mass is in its precise fabrication, not its composition (unless you go
very exotic like tellurium). Everything else is identical, which is
the point of the exercise. Any output in the parity case can only
have one source - test mass opposite parity.
The parity Eotvos experiment has received some indirect support.
Jeffrey Weeks' dodecahedral model of the universe from WMAP data made
big headlines. What was not commonly reported (but was in the
"Science" article) is that such a connected dodecahedral universe is
chiral. Opposite faces of a dodecahedron are rotated 36 degrees.
That puts a chiral twist on the whole construct. The parity Eotvos
experiment might fit right in. I've e-mailed Weeks for an opinion.
Science 302(5643) 209 (2003)
http://arXiv.org/abs/astro-ph/0310253
Nature 425(6958) 593 (2003)
http://www.nature.com/nsu/031006/031006-8.html
--
Uncle Al
http://www.mazepath.com/uncleal/qz.pdf
http://www.mazepath.com/uncleal/eotvos.htm
(Do something naughty to physics)
.
|
|
|
| User: "Bilge" |
|
| Title: Re: Parity Eotvos experiment, update |
19 Oct 2003 02:38:10 PM |
|
|
Uncle Al:
Bilge wrote:
Aside from all of the neato, gee-whiz computations, have you
considered computing the actual cost of doing the experiment
including an analysis of any trade-off in precision vs. cost
for the calculations you've already done? Someone who would
consider doing the experiment might like an accurate figure
for the cost of setting a limit and knowing what that limit
is, broken down into each aspect of the experiment.
The cost of a parity Eotvos experiment in quartz is that of a
composition Eotvos experiment in anything. The major cost of a test
mass is in its precise fabrication, not its composition (unless you go
very exotic like tellurium). Everything else is identical, which is
the point of the exercise. Any output in the parity case can only
have one source - test mass opposite parity.
That doesn't really cut it in a proposal. Every proposal I've
ever seen has things like:
cost to produce test mass from [ insert material ]
test mass from [ insert material ]
etc.
cost of facility for X hours at $Y per hour
[additional cost for personnel where needed]
as well as things like who is likely to manufacture the samples
assuming you don't have the equipment to do so or at the level
of precision required. Or, if the samples don't require any
extraordinary skill to manufacture, some quantitative analysis
that demonstrates that fact.
Other items include who would be involved in the experiment, the
facilities that are ppotentially available, a detailed analysis of
how any trade offs in the number of facility hours required, costs
of materials, etc., are expected to result in the ability of the
experiment to realize a physically meaningful precision for the
measurements.
The parity Eotvos experiment has received some indirect support.
Jeffrey Weeks' dodecahedral model of the universe from WMAP data made
big headlines.
Yes, I saw that. It fits my theory that the age of the universe
gets revised every couple of years with a "better" one that is
closer to a prediction earlier than the last "better" one (among
other things).
What was not commonly reported (but was in the
"Science" article) is that such a connected dodecahedral universe is
chiral. Opposite faces of a dodecahedron are rotated 36 degrees.
That puts a chiral twist on the whole construct.
That gives you an opportunity (not an guarantee) for a plausibility
argument, depending upon how seriously anyone takes the observation of the
dodecahedral universe). I realize your basic argument is "it's cheap, it's
easy, so someone should look", so don't tell me that. If you could connect
that observation to your experiment with something mmore concrete than
speculation, in my opinion, your experiment would be a shoe-in. Since the
connection is not obvious to me, coming up with a plausible connection
might prove to be non-trivial.
The parity Eotvos experiment might fit right in.
I've e-mailed Weeks for an opinion.
It might, but unless someone can figure out how much it will
realistically cost in actual dollars and weigh its evaluate it on how much
it costs vs. the physical question(s) it addresses, no one can decide
whether or not to do it. That can work either way. If it were all laid out
in dollar figures, someone might decide it was better value for the money
than some other experiment planned for the same facility. Your experiment
_will_ cost someone some money, whether that is the DOE, NSF, the Welch
Foundation, the funds obtained by some research group or whatever. Someone
has to actually do the experiment and get paid for it. Unless they know
how much it costs, they aren't going to consider it. Even someone you
would hope to convince to include your experiment as part of their current
research would want to know how much of their time and money it would
require to produce a useful number.
.
|
|
|
| User: "tj Frazir" |
|
| Title: Re: Parity Eotvos experiment, update |
19 Oct 2003 03:37:53 PM |
|
|
Thats it Bilge ,,hit em where the money is.
Bilge must be a diesel duck too.
Tell Al he should be making tanker oil hard so it cant sink or running
the mad river down the coast underwater and soliving califorias wosse.
Ill have to tun a upside down changel out of concreet 70 feet wide with
15 foot sides from 285 feet abouve see level up the eel river and send
the river to the farms and la down the shore under water and solve all
their problems with out a pump.
20 miles a day done. 1000 miles to mexico.
700 miles to la . 75 million acre feet a year.
60 million acre feet oragnge county .
Apple valley 3 maf. ALL from the mad and eel rivers. 200 new
canneries ,,,I new port.
The best physics makes life on earth better.
Even if you must get a hat rack.
.
|
|
|
| User: "MorituriMax" |
|
| Title: Re: Parity Eotvos experiment, update |
20 Oct 2003 12:03:13 AM |
|
|
"tj Frazir" <GravityPhysics@webtv.net> wrote in message
news:9930-3F92F621-272@storefull-2155.public.lawson.webtv.net...
Thats.. .. ..
BS Filter Activated.
.
|
|
|
|
| User: "Randy Poe" |
|
| Title: Re: Parity Eotvos experiment, update |
21 Oct 2003 03:15:01 PM |
|
|
(tj Frazir) wrote in message news:<9930-3F92F621-272@storefull-2155.public.lawson.webtv.net>...
Thats it Bilge ,,hit em where the money is.
Bilge must be a diesel duck too.
Tell Al he should be making tanker oil hard so it cant sink or running
the mad river down the coast underwater and soliving califorias wosse.
Ill have to tun a upside down changel out of concreet 70 feet wide with
15 foot sides from 285 feet abouve see level up the eel river and send
the river to the farms and la down the shore under water and solve all
their problems with out a pump.
20 miles a day done. 1000 miles to mexico.
700 miles to la . 75 million acre feet a year.
60 million acre feet oragnge county .
Apple valley 3 maf. ALL from the mad and eel rivers. 200 new
canneries ,,,I new port.
The best physics makes life on earth better.
Even if you must get a hat rack.
Did you send that $4 Billion check to California yet? When
can I expect to read the press release?
.
|
|
|
| User: "Jim" |
|
| Title: Re: Parity Eotvos experiment, update |
21 Oct 2003 05:26:36 PM |
|
|
On 21 Oct 2003 13:15:01 -0700, (Randy Poe)
wrote:
GravityPhysics@webtv.net (tj Frazir) wrote in message news:<9930-3F92F621-272@storefull-2155.public.lawson.webtv.net>...
Thats it Bilge ,,hit em where the money is.
Bilge must be a diesel duck too.
Tell Al he should be making tanker oil hard so it cant sink or running
the mad river down the coast underwater and soliving califorias wosse.
Ill have to tun a upside down changel out of concreet 70 feet wide with
15 foot sides from 285 feet abouve see level up the eel river and send
the river to the farms and la down the shore under water and solve all
their problems with out a pump.
20 miles a day done. 1000 miles to mexico.
700 miles to la . 75 million acre feet a year.
60 million acre feet oragnge county .
Apple valley 3 maf. ALL from the mad and eel rivers. 200 new
canneries ,,,I new port.
The best physics makes life on earth better.
Even if you must get a hat rack.
Did you send that $4 Billion check to California yet? When
can I expect to read the press release?
Hey!! Only 3bil. He promised the other 1bil to me. :)
I have it in writing, in the Google archives.
Jim
.
|
|
|
| User: "tj Frazir" |
|
| Title: Re: Parity Eotvos experiment, update |
21 Oct 2003 06:12:58 PM |
|
|
Yes ,,the one billion is allready there JIM ..
its there ,,I showed you pics of it .
you wll nead to convert it jim . take a mecanical pulp grinder refiner
and grind it into a billion ..based on $ 550 per ton pulp price index an
the barkless soft aspin pines on the beach.
Are you lazy or just stupid Jim ? 1 billion bucks is on the beach free
as ***** and you cant get it threw your head !
I told you if you neaded any kind of help just ask, I posted evry
major buyer and noted that it is 3 billion if you add a machine and
make it tolet paper instead of pulp . You can sell the tolet paper on
contract as recycle save a tree for the very best tolet paper ( no bark
).
wallmart ,,dolar stores ,,would all love your contract. I can send you
a list so you can call one as soon as you fork over all the math .
Im still waiting . we not waiting ..we on our way south inthe new ice
breaker.
.
|
|
|
| User: "MorituriMax" |
|
| Title: Re: Parity Eotvos experiment, update |
22 Oct 2003 12:19:14 AM |
|
|
"tj Frazir" <GravityPhysics@webtv.net> wrote in message
news:27378-3F95BD7A-250@storefull-2154.public.lawson.webtv.net...
Yes ,,the one billion is allready there JIM ..
Yeah, right... so sad.
.
|
|
|
|
|
|
|
|
|
|
|
| User: "P.Lj" |
|
| Title: Re: Parity Eotvos experiment, update |
16 Oct 2003 04:09:56 AM |
|
|
Uncle Al <UncleAl0@hate.spam.net> wrote in news:3F8DEA1B.EA836159
@hate.spam.net:
If anybody cares...
http://www.mazepath.com/uncleal/qz.pdf
Fig. 2, Table VI.
We recently crunched quartz to 3.3 quadrillion atoms or a 43 micron
diameter ball - 2.8 GHz Xeon and 1.8 GHz Opteron CPUs in Linux gave
identical data - up from 0.33 Quad and 20 microns. The graphic fit to
parity divergence, CHI, tweaked in the fourth decimal place, slope and
intercept. 15 Quads is cooking right now, or 71 microns diameter.
The native 32-bit source code is being ported to 64-bit and is being
parallelized for a dual-Opteron box. If this goes through, we go for
42 quadrillion atoms or a 100 micron diameter ball (1/10 millimeter)
in a CPU-month. That is a classical physics mass to even the most
skeptical pundit. No surprises vs. scale can hide.
If we can get into an 8-Opteron box, 4 CPU-weeks does 205 Quads or a
170 micron diameter ball. In a world dripping with *****, the
Gifted across the globe still shout defiance. Appreciation goes out
to four code poets in four different countries who offered to toss
some unused CPU-time Uncle Al's way - and boasted that the code could
run faster.
If anybody has a big cluster not otherwise engaged or needs a burn-in
for new equipment, I expect to have a program that will flex every
transistor in every CPU but not bother other components. All I ask in
return is the data generated. An Opteron running 32-bit code
massively outperforms a Xeon running at 156% of Opteron clock speed.
Windows is 40% slower than Linux in all iron and we abandoned any
thought of it.
No more "Intel Inside" for Uncle Al. Go AMD.
--
Uncle Al
http://www.mazepath.com/uncleal/eotvos.htm
(Do something naughty to physics)
Is that code available for personal testing on other architecures ?
^P.Lj
.
|
|
|
| User: "Uncle Al" |
|
| Title: Re: Parity Eotvos experiment, update |
16 Oct 2003 12:05:43 PM |
|
|
"P.Lj" wrote:
Uncle Al <UncleAl0@hate.spam.net> wrote in news:3F8DEA1B.EA836159
@hate.spam.net:
If anybody cares...
http://www.mazepath.com/uncleal/qz.pdf
Fig. 2, Table VI.
We recently crunched quartz to 3.3 quadrillion atoms or a 43 micron
diameter ball - 2.8 GHz Xeon and 1.8 GHz Opteron CPUs in Linux gave
identical data - up from 0.33 Quad and 20 microns. The graphic fit to
parity divergence, CHI, tweaked in the fourth decimal place, slope and
intercept. 15 Quads is cooking right now, or 71 microns diameter.
The native 32-bit source code is being ported to 64-bit and is being
parallelized for a dual-Opteron box. If this goes through, we go for
42 quadrillion atoms or a 100 micron diameter ball (1/10 millimeter)
in a CPU-month. That is a classical physics mass to even the most
skeptical pundit. No surprises vs. scale can hide.
If we can get into an 8-Opteron box, 4 CPU-weeks does 205 Quads or a
170 micron diameter ball. In a world dripping with *****, the
Gifted across the globe still shout defiance. Appreciation goes out
to four code poets in four different countries who offered to toss
some unused CPU-time Uncle Al's way - and boasted that the code could
run faster.
If anybody has a big cluster not otherwise engaged or needs a burn-in
for new equipment, I expect to have a program that will flex every
transistor in every CPU but not bother other components. All I ask in
return is the data generated. An Opteron running 32-bit code
massively outperforms a Xeon running at 156% of Opteron clock speed.
Windows is 40% slower than Linux in all iron and we abandoned any
thought of it.
No more "Intel Inside" for Uncle Al. Go AMD.
--
Uncle Al
http://www.mazepath.com/uncleal/eotvos.htm
(Do something naughty to physics)
Is that code available for personal testing on other architecures ?
E-mail address in the heading of
http://www.mazepath.com/uncleal/qz.pdf
Source code is C++, 32-bit bandwidth, 80-bit long_double_precision
calculation. 64-bit bandwith, parallelized version is being coded and
benchmarked. We are only interested in hardware that can exceed the
performance of four CPU-weeks in a 1.8 GHz Opteron, or seven CPU-weeks
in a 2.8 GHz Xeon.
--
Uncle Al
http://www.mazepath.com/uncleal/
(Toxic URL! Unsafe for children and most mammals)
"Quis custodiet ipsos custodes?" The Net!
.
|
|
|
|
|
| User: "Graham Lee" |
|
| Title: Re: Parity Eotvos experiment, update |
15 Oct 2003 08:36:12 PM |
|
|
Uncle Al wrote:
If anybody cares...
http://www.mazepath.com/uncleal/qz.pdf
Fig. 2, Table VI.
We recently crunched quartz to 3.3 quadrillion atoms or a 43 micron
diameter ball - 2.8 GHz Xeon and 1.8 GHz Opteron CPUs in Linux gave
identical data
One would hope that digital computers running the same code would give
the same results as each other (to within limits of the precision of the
data types employed).
BTW, I find "experiment" a bit of a strong term to describe a numerical
simulation. I like to call them Numerical Simulations, Brute-Force
modelling, or lettuce mathematics (the Monte Zuma method).
As for switching to 64-bit code; overrated. What's commonly forgotten
is that the general design of the processor improves with time, not only
the register sizes which for floating point number-crunching aren't so
important; time taken to load the data into the registers is a
high-order perturbation in comparison with the time taken to manipulate
it. For instance, I'd rather use a 32-bit PowerPC unit than a 64-bit
UltraSPARC 1[*], despite the latter having bigger bittage.
Graham.
[*]Frankly I'd rather use an M68k than an US1: the 68k parts tend to die
on their ar53s less often.
.
|
|
|
| User: "Richard Herring" |
|
| Title: Re: Parity Eotvos experiment, update |
16 Oct 2003 06:02:48 AM |
|
|
Graham Lee <graham.lee@wadham.ox.ac.invalid.uk> wrote in message news:<bmkskb$r3e$1@news.ox.ac.uk>...
Uncle Al wrote:
If anybody cares...
http://www.mazepath.com/uncleal/qz.pdf
Fig. 2, Table VI.
We recently crunched quartz to 3.3 quadrillion atoms or a 43 micron
diameter ball - 2.8 GHz Xeon and 1.8 GHz Opteron CPUs in Linux gave
identical data
One would hope that digital computers running the same code would give
the same results as each other (to within limits of the precision of the
data types employed).
If one doesn't take sufficient care with the numerical analysis,
sometimes the only thing the code actually calculates _is_ the
precision of the data types employed :-(
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&threadm=3EF8661B.103DC815%40hate.spam.net&rnum=1&prev=/groups%3Fas_q%3Djacobi%2520transformation%26safe%3Dimages%26ie%3DUTF-8%26oe%3DUTF-8%26as_ugroup%3Dsci.physics%26as_uauthors%3Dal%26lr%3D%26hl%3Den
.
|
|
|
| User: "Uncle Al" |
|
| Title: Re: Parity Eotvos experiment, update |
16 Oct 2003 11:45:49 AM |
|
|
Richard Herring wrote:
Graham Lee <graham.lee@wadham.ox.ac.invalid.uk> wrote in message news:<bmkskb$r3e$1@news.ox.ac.uk>...
Uncle Al wrote:
If anybody cares...
http://www.mazepath.com/uncleal/qz.pdf
Fig. 2, Table VI.
We recently crunched quartz to 3.3 quadrillion atoms or a 43 micron
diameter ball - 2.8 GHz Xeon and 1.8 GHz Opteron CPUs in Linux gave
identical data
One would hope that digital computers running the same code would give
the same results as each other (to within limits of the precision of the
data types employed).
If one doesn't take sufficient care with the numerical analysis,
sometimes the only thing the code actually calculates _is_ the
precision of the data types employed :-(
80-bit double_long_precision only in our calculation. Two separate
analyses show we are well beyond the precision needed. All results
exactly overlap to 18 decimal places - Windows or Linux OS; Intel and
AMD consumer PC CPUs; Xeon and Opteron CPUs.
Double_precision shows noise in the ~11th decimal place.
--
Uncle Al
http://www.mazepath.com/uncleal/
(Toxic URL! Unsafe for children and most mammals)
"Quis custodiet ipsos custodes?" The Net!
.
|
|
|
| User: "Richard Herring" |
|
| Title: Re: Parity Eotvos experiment, update |
17 Oct 2003 03:54:04 AM |
|
|
Uncle Al <UncleAl0@hate.spam.net> wrote in message news:<3F8ECB3D.5E42BEE7@hate.spam.net>...
Richard Herring wrote:
Graham Lee <graham.lee@wadham.ox.ac.invalid.uk> wrote in message news:<bmkskb$r3e$1@news.ox.ac.uk>...
Uncle Al wrote:
If anybody cares...
http://www.mazepath.com/uncleal/qz.pdf
Fig. 2, Table VI.
We recently crunched quartz to 3.3 quadrillion atoms or a 43 micron
diameter ball - 2.8 GHz Xeon and 1.8 GHz Opteron CPUs in Linux gave
identical data
One would hope that digital computers running the same code would give
the same results as each other (to within limits of the precision of the
data types employed).
If one doesn't take sufficient care with the numerical analysis,
sometimes the only thing the code actually calculates _is_ the
precision of the data types employed :-(
80-bit double_long_precision only in our calculation. Two separate
analyses show we are well beyond the precision needed.
Good. I assume you'll be publishing them?
All results
exactly overlap to 18 decimal places - Windows or Linux OS; Intel and
AMD consumer PC CPUs; Xeon and Opteron CPUs.
But all _that_ proves is that the hardware is performing as specified.
If the results from two entirely different _algorithms_ overlapped to
18 places I'd be more impressed.
Double_precision shows noise in the ~11th decimal place.
"Noise" is too vague. Where does rounding error start to eat into
truncation error?
.
|
|
|
| User: "Uncle Al" |
|
| Title: Re: Parity Eotvos experiment, update |
17 Oct 2003 09:59:47 AM |
|
|
Richard Herring wrote:
Uncle Al <UncleAl0@hate.spam.net> wrote in message news:<3F8ECB3D.5E42BEE7@hate.spam.net>...
Richard Herring wrote:
Graham Lee <graham.lee@wadham.ox.ac.invalid.uk> wrote in message news:<bmkskb$r3e$1@news.ox.ac.uk>...
Uncle Al wrote:
If anybody cares...
http://www.mazepath.com/uncleal/qz.pdf
Fig. 2, Table VI.
We recently crunched quartz to 3.3 quadrillion atoms or a 43 micron
diameter ball - 2.8 GHz Xeon and 1.8 GHz Opteron CPUs in Linux gave
identical data
One would hope that digital computers running the same code would give
the same results as each other (to within limits of the precision of the
data types employed).
If one doesn't take sufficient care with the numerical analysis,
sometimes the only thing the code actually calculates _is_ the
precision of the data types employed :-(
80-bit double_long_precision only in our calculation. Two separate
analyses show we are well beyond the precision needed.
Good. I assume you'll be publishing them?
http://www.mazepath.com/uncleal/gz.pdf
(Quartz to .33 quad, not the full 3.3 qaud)
The calculation is subsidiary to the physics - and mostly a show of
bravado for the Referees. Periodic crystals are self-similar by
definition. Even a small radial sample beyond a few unit cells sets
the trend to arbitrarily large dimensions. However, because an atomic
lattice is not homogeneous (atom or no atom) and not istropic
(certainly not for the thre qualified pairs of enantiomorphic space
groups) we get a full standard deviation of noise calculated points
vs. best fit curve. As with a real world test mass, things average
out with accumulating size and only the trend remains...
....but you never know. A 12,000 atom lattice isn't much of an
argument (a cube 23 atoms on an edge). A 3.3 quadrillion atom lattice
is a better example (a cube 150,000 atoms on an edge). The surface
doesn't intrude so much. 42 quadrillion atoms (a cube 350,000 atoms
on an edge) is better still.
All results
exactly overlap to 18 decimal places - Windows or Linux OS; Intel and
AMD consumer PC CPUs; Xeon and Opteron CPUs.
But all _that_ proves is that the hardware is performing as specified.
If the results from two entirely different _algorithms_ overlapped to
18 places I'd be more impressed.
Double_precision shows noise in the ~11th decimal place.
"Noise" is too vague. Where does rounding error start to eat into
truncation error?
80-bit long_double-precision code in an 80-bit precision CPU is more
than sufficient to comfortably retain 18 decimal places through
arithmetic operations for the number of points considered. The lowest
precision operation is a square root. No trig functions.
1) Two programmers have rigorously looked at the precision needed
vs. dataset size and are satisfied. That's theory.
2) A Xeon and an Opteron gave absolutely identical results for the
3.3 quad run. We see no hardware dependence. Random error is not
ocurring.
3) The fitted slope and intercept shifted slightly in their fourth
decimal place when teh 3.3 qaud run was added in, increasing the atoms
considered tenfold. If there were accumulating error for whatever
reason, the 3.3 quad run would not be consistent with numbers from
1000 atoms. We see a large dataset breakdown using double_precision.
--
Uncle Al
http://www.mazepath.com/uncleal/
(Toxic URL! Unsafe for children and most mammals)
"Quis custodiet ipsos custodes?" The Net!
.
|
|
|
|
|
|
|
| User: "Randy Poe" |
|
| Title: Re: Parity Eotvos experiment, update |
21 Oct 2003 03:13:02 PM |
|
|
Graham Lee <graham.lee@wadham.ox.ac.invalid.uk> wrote in message news:<bmkskb$r3e$1@news.ox.ac.uk>...
Uncle Al wrote:
If anybody cares...
http://www.mazepath.com/uncleal/qz.pdf
Fig. 2, Table VI.
We recently crunched quartz to 3.3 quadrillion atoms or a 43 micron
diameter ball - 2.8 GHz Xeon and 1.8 GHz Opteron CPUs in Linux gave
identical data
One would hope that digital computers running the same code would give
the same results as each other (to within limits of the precision of the
data types employed).
One would hope so, but with large-scale simulations it's
rarely the case. Just because you can get 64-bit precision
on one multiply doesn't mean you have 64-bit accuracy
after a couple of billion successive multiplies and
adds. Far from it.
- Randy
.
|
|
|
|
| User: "Uncle Al" |
|
| Title: Re: Parity Eotvos experiment, update |
15 Oct 2003 09:27:13 PM |
|
|
Graham Lee wrote:
Uncle Al wrote:
If anybody cares...
http://www.mazepath.com/uncleal/qz.pdf
Fig. 2, Table VI.
We recently crunched quartz to 3.3 quadrillion atoms or a 43 micron
diameter ball - 2.8 GHz Xeon and 1.8 GHz Opteron CPUs in Linux gave
identical data
One would hope that digital computers running the same code would give
the same results as each other (to within limits of the precision of the
data types employed).
BTW, I find "experiment" a bit of a strong term to describe a numerical
simulation. I like to call them Numerical Simulations, Brute-Force
modelling, or lettuce mathematics (the Monte Zuma method).
[snip]
The calculation qualifies test masses with which to perform the
experiment. If the parity Eotvos experiment gives net output, then
the Equivalence Principle is counterdemonstrated and all metric
theories of gravitation are WRONG at their founding postulate. Affine
theories take over.
General Relativity is a metric theory of gravitation.
--
Uncle Al
http://www.mazepath.com/uncleal/
(Toxic URL! Unsafe for children and most mammals)
"Quis custodiet ipsos custodes?" The Net!
.
|
|
|
| User: "tj Frazir" |
|
| Title: Re: Parity Eotvos krackpot update |
15 Oct 2003 11:50:34 PM |
|
|
Now unce Al just went over the line and became a crackpot.
spupidity plus stupidity is just stupid.
equivalance is mass and energy are interchangable and identical to
conservation and Uncle dumbass thinks he will prove conservation WRONG !
dumbass ,,you should be working on oil
.
|
|
|
| User: "MorituriMax" |
|
| Title: Re: Parity Eotvos krackpot update |
16 Oct 2003 01:30:43 AM |
|
|
"tj Frazir" <GravityPhysics@webtv.net> wrote in message
news:24370-3F8E239A-23@storefull-2156.public.lawson.webtv.net...
Now.. .. ..
BS Filter Activated.
.
|
|
|
|
| User: "Jim" |
|
| Title: Re: Parity Eotvos krackpot update |
16 Oct 2003 02:58:59 PM |
|
|
On Thu, 16 Oct 2003 00:50:34 -0400 (EDT), (tj
Frazir) wrote:
Now unce Al just went over the line and became a crackpot.
spupidity plus stupidity is just stupid.
equivalance is mass and energy are interchangable and identical to
conservation and Uncle dumbass thinks he will prove conservation WRONG !
dumbass ,,you should be working on oil
Why the interest in employing the services of someone you
think is such a dumbass/idiot?
If you're so smart, do it yourself.
BTW, where's that 4bil you owe me?
Jim
.
|
|
|
| User: "The Ghost In The Machine" |
|
| Title: Re: Parity Eotvos krackpot update |
16 Oct 2003 11:00:10 PM |
|
|
In sci.physics, Jim
<lose30lb@workfromhome.com>
wrote
on Thu, 16 Oct 2003 19:58:59 GMT
<qvttovo4enf9j7gq44s0tf8dg4dv0pqa9f@4ax.com>:
On Thu, 16 Oct 2003 00:50:34 -0400 (EDT), (tj
Frazir) wrote:
Now unce Al just went over the line and became a crackpot.
spupidity plus stupidity is just stupid.
equivalance is mass and energy are interchangable and identical to
conservation and Uncle dumbass thinks he will prove conservation WRONG !
dumbass ,,you should be working on oil
Why the interest in employing the services of someone you
think is such a dumbass/idiot?
If you're so smart, do it yourself.
BTW, where's that 4bil you owe me?
Jim
I think he spent it on his boat.... :-)
--
#191,
It's still legal to go .sigless.
.
|
|
|
| User: "MorituriMax" |
|
| Title: Re: Parity Eotvos krackpot update |
17 Oct 2003 12:59:35 AM |
|
|
"The Ghost In The Machine" <ewill@sirius.athghost7038suus.net> wrote
in message news:1o2461-3if.ln1@lexi2.athghost7038suus.net...
I think he spent it on his boat.... :-)
Nah, she spent it on cheetos..
.
|
|
|
|
| User: "tj Frazir" |
|
| Title: Re: Parity Eotvos krackpot update |
17 Oct 2003 12:19:10 AM |
|
|
Im on chemest ,,al is no physicsist .
But if al wants 10 million dumped in his lap all he has to do is find a
way to make a tanker of oil a solid .
make that 50 million in cold hard cash.
That is for anyone that can !
His priorities are wrong. 50 million is a chunk of change bud ! the
tanker wount haul 1/2 hardner but get it down to 200 tons vrs 100000
tons and uncle al will be the 50 million grand prise winner.
BTW yer all dumbfucks ,,no one knows it all.
I learn evry day and Im so rich you would not comprehend it.
He has a hummer he has not seen untill they know here he wants it wih
an open title.
it was 160 grand untill I sent it back and said wheel chair driver one
hand ,,,that was 40 more. probly worth more than his house.
.
|
|
|
| User: "MorituriMax" |
|
| Title: Re: Parity Eotvos krackpot update |
17 Oct 2003 01:00:00 AM |
|
|
"tj Frazir" <GravityPhysics@webtv.net> wrote in message
news:5732-3F8F7BCE-377@storefull-2157.public.lawson.webtv.net...
Im.. .. ..
BS Filter Activated.
.
|
|
|
|
| User: "Jeff Relf" |
|
| Title: . Freeze it Under Pressure . |
17 Oct 2003 06:01:33 AM |
|
|
Hi TJ ,
You say :
" find a way to make a tanker of oil a solid . "
That's an easy one : Freeze it under pressure .
P.S. regarding the hummer ,
Are you in a wheel chair , and a one handed driver ?
.
|
|
|
| User: "tj Frazir" |
|
| Title: Re: . Freeze it Under Pressure . |
17 Oct 2003 08:52:46 AM |
|
|
That would force te oil out the leaking tanker.
Keep in mind the oil co dont know what the question is yet. It would
be the answer befor they could ask it. If you made it a solid or so
jelled it couldnt do any damage you would save billions and make your
self very rich just for filing a working pattent.
Theres a pattent that would sell !
It would become law !
If you had a hardner for oil the oil slick would never be seen again .
We would name the stuff after the inventor. schwarzts oil ?
Biuld a house with waist oil !
Tanker breaks at sea or runs aground ,,a schwarzts oil tug pulls up and
injects the oil .
Then its a solid ,,the ship wount sink the oil wount slick and the
planet is safe and the inventor saves the day !
Insurance on a tanker would be cut in half.
Uncle al would go down in history .
He would colect off all the other things oil would be used for. The
reward would be imedete. The 10 million from me would jjust be the down
playment from planet earth,
uncle al would stick the cash in the refiner that makes oil hard and be
the worlds poroducer of the stuff.
A rich man knows what he is good at and dose one thing very well ,,a
poor man will do evrything average .
A rich man will have poeple do things very well
and the combination makes the world a better place.
Isnt that evryones end goal ? Better world ?
If anyone had the stuff today ,,anyone would get rich today.
Ill send a brinks truck !
.
|
|
|
| User: "MorituriMax" |
|
| Title: Re: . Freeze it Under Pressure . |
17 Oct 2003 11:09:11 AM |
|
|
"tj Frazir" <GravityPhysics@webtv.net> wrote in message
news:27808-3F8FF42E-33@storefull-2153.public.lawson.webtv.net...
That.. .. ..
poor tj.. Jeff has figured out how to do it and tj can't cope with
it... freeze it under pressure.. so simple...
.
|
|
|
|
| User: "Jeff Relf" |
|
| Title: . Oil Spills . |
17 Oct 2003 11:25:24 PM |
|
|
Hi TJ ,
You say :
" Tanker breaks at sea or runs aground ...
a schwarzts oil tug pulls up and injects the oil .
Then its a solid ...
the ship won't sink the oil wount slick
and the planet is safe
and the inventor saves the day ! "
Double hulls have solved the running aground problem .
Sometimes tankers break apart in storms ...
In places too remote for a tug .
Stronger hulls might be the best solution for that .
.
|
|
|
| User: "tj Frazir" |
|
| Title: Re: . Oil Spills . |
18 Oct 2003 08:43:19 AM |
|
|
Dobble hull improved it but no where near the cure. A 6 year old tanker
broke off spain a few years ago and cosy 13 billion to clean off the
beach. valdes cost 80 billion . A container ship with 600 tons of
fuel is bad enouph and one is down evry day.
50 ships ran aground this year so far !
It dont matter if a dobble hull never breaks ,,the inshurance would blow
your mind !
And that dont matter ,,its all irrelivant.
The fist fact is I will put 10 million in a brinks truck in cash for it
! I dont care if it cost 1 million to load the tug ,,Ill load 200 tugs
!
The tug will just sit in port and look good.
Hope we never had to move the tug.
NO ONE else on the planet can inevnt it unless you read these post.
No one esle ask te question and nothing is ever invented till the
question is ask. Not even the oil co understands the question they
never thought to ask. Its part physics part chem and part money.
If uncle invents it then its going down in history as uncle al oil.
swcharets oil or what ever.
.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Related Articles |
|
|
