Computer modeling and simulation

[BigAl]

Why all the worry about GPU?
The big problem is the actual calculations. The non-linearity of the system takes all the CPU you can spare. Most of the current programs (IDEAS, NASTRAN, and I would assume LS_DYNA can make use of distributed processing--defined as "whatever CPU is not beuing used".
It is the actual CALCULATIONS that take the time.
Rendering and graphics are just bells and whistles--part of the "dog and pony show" -making pretty pictures to present the results to managers and other idiots who cannot understand the actual results​

These days, the GPU on the video card can be programmed to do general arithmetic and it does it at a screaming rate compared to the CPU and on a good day both can be kept busy in parallel. Standard run-time libraries make the programming of the GPU available to any programmer. (The architecture of the problem may or may not be a good fit the vector-processor architecture of the GPU.)

Photoshop is the one application I can name that uses the GPU to apply algorithms to images and this is unrelated to displaying images on the screen.

Given that NIST used a 16(?) CPU cluster based on 2004(?) hardware to model the WTC1/2 collapse, by 2009 standards, this really isn't very much computer power, then and especially now.

The Sony Playstation 3 makes a screaming node in a cluster, several times faster than the fastest Intel chip. If the software NIST used runs on it, it's the way to go.

 

[TriskettheKid]

Why all the worry about GPU?
The big problem is the actual calculations. The non-linearity of the system takes all the CPU you can spare. Most of the current programs (IDEAS, NASTRAN, and I would assume LS_DYNA can make use of distributed processing--defined as "whatever CPU is not beuing used".
It is the actual CALCULATIONS that take the time.
Rendering and graphics are just bells and whistles--part of the "dog and pony show" -making pretty pictures to present the results to managers and other idiots who cannot understand the actual results​

Because if you can code properly, you can offload computation to the parallel processors on the GPU.

Games have been doing something kind of similar for a while now, with PhysX. For a while, game physics were computed using the CPU, which proved fairly taxing on the older architectures (P4, etc). The work around was to create PhysX, which was a second (or third, if using SLI) card in a PCIe slot that was dedicated solely to physics calcs. As not many games used it, and processing power at both the CPU and GPU increased monumentally, PhysX was "dropped" and melded into CUDA (sorry, I'm probably glossing over a bit of nuance here).

What this means is, with the proper "conversion" or coding, you can redirect a program to utilize the heavy computing power of yoru GPU.

I linked to a paper at CalTech that discussed just this kind of idea, though it was from an older generation of CPUs and GPUs (it references the horrid nVidia FX series).

Here's the paper:

http://multires.caltech.edu/pubs/GPUSim.pdf

 

[rwguinn]

Because if you can code properly, you can offload computation to the parallel processors on the GPU.

Games have been doing something kind of similar for a while now, with PhysX. For a while, game physics were computed using the CPU, which proved fairly taxing on the older architectures (P4, etc). The work around was to create PhysX, which was a second (or third, if using SLI) card in a PCIe slot that was dedicated solely to physics calcs. As not many games used it, and processing power at both the CPU and GPU increased monumentally, PhysX was "dropped" and melded into CUDA (sorry, I'm probably glossing over a bit of nuance here).

What this means is, with the proper "conversion" or coding, you can redirect a program to utilize the heavy computing power of yoru GPU.

I linked to a paper at CalTech that discussed just this kind of idea, though it was from an older generation of CPUs and GPUs (it references the horrid nVidia FX series).

Here's the paper:

http://multires.caltech.edu/pubs/GPUSim.pdf

You are talking a commercial product here. It is set up to run on several platforms (actually, you buy for your platform)
Usually these are UNIX, Windows, Possibly VMS and NOS (or whatever the equivalent Mainframes are today). Don't know about the others.
So you have to design your OS to emulate these, and do whatever parallel stuff you can do. Not sure how all that works--I'm a user, not a pusher

So I will bow to your expertise in this matter

 

[Toke]

Ok, so getting a grafics card to help with calculations is possible if you know what you are doing.
That excludes me, too bad, games rendering/compilation will be an overnight activity.

(I hoped there were some simple command to engage grafics card along with 2. cpu.)
Thanks for the hints.

 

[Klimax]

Ok, so getting a grafics card to help with calculations is possible if you know what you are doing.
That excludes me, too bad, games rendering/compilation will be an overnight activity.

(I hoped there were some simple command to engage grafics card along with 2. cpu.)
Thanks for the hints.

However one has to be very carefull about precision...
ETA: (And One CPU has to dedicate some percentage of time to keep GPU fed. This sort of computing is now in BOINC world and it does not appear to be that faster at this time)

 

[TriskettheKid]

You are talking a commercial product here. It is set up to run on several platforms (actually, you buy for your platform)
Usually these are UNIX, Windows, Possibly VMS and NOS (or whatever the equivalent Mainframes are today). Don't know about the others.
So you have to design your OS to emulate these, and do whatever parallel stuff you can do. Not sure how all that works--I'm a user, not a pusher

So I will bow to your expertise in this matter

Oh, no I understand fully.

What I'm saying is, if you had the know-how, and the source code, you could try and "convert" the source code a GPU-heavy setup could use, and that would allow you to put something together for a cheaper amount.

It may not run as fast, or as well, as the nodes NIST used, but GPUs are coming into their own in terms of power and ability.

Of course, you're still going to have to test EVERY processor and stick of RAM used for errors (Prime95, memtest, etc), so taht you know which parts will run error-free for the time needed.

As a base comparison, I ran a 24-hour stress test on my Core i7 when I OC'ed it from stock to 3.6gHz. It was just one part, but it showed no errors. The work we're talking about running would require all necessary parts to run, under stress, for weeks on end, without having an error.

 

[Mr.D]

All of this of course begs the question of what the Truthers would do even if they went ahead and did this.

I mean, just suppose a Truther out there had the financial wherewithall to purchase, assemble, test, validate and operate such a cluster with the same software and model that NIST used.

Then what?

Remember that NIST's simulation was of collapse initiation. So what happens when the new simulation comes within a small margin of error of reproducing NIST's result which would still lead to global collapse?

Will they produce an alternate model to defend any of their demolition/thermite/noplane theories?

Alternately, is the TM ready to defend a result that is either too perfect or too unphysical?

 

[dropzone]

Alternately, is the TM ready to defend a result that is either too perfect or too unphysical?

They've done it for eight years. Why stop now?

 

[cludgie]

This is still true in Formula One, although I don't think they scale the cars down much if at all; the few picures I see look to be a full or nearly full sized chassis. Anyway, some teams have two, not just one, wind tunnels for development; there's tons of money spent in F1, and the top teams go all out. However, in an attempt to cut costs and speed aerodynamic development, some teams are switching to CFD (rwguinn and all the engineers here will know what a "Computational Fluid Dynamics" system is) in place of wind tunnels; here's a story about the Renault team increasing their use of such "virtual" wind tunnels. And here's a different one from a few years back advertising an association with a now defunct team. Real tunnels are expensive, and do need down time for routine maintenance, although I don't read anything about any F1 team completely ditching the real world wind buildings. I don't think any of them are willing to go that far.

Oddly enough, back when Indycar's direct rival - Champ Car World Series - still existed, they actually had a rule against wind tunnel testing. What they did in place of that, I don't know. Indycar had no such restriction.

To bring this back on topic: Yes, it's true, virtual modeling for race cars at least is very advanced. If it gives a competetive advantage, you can bet F1 teams will somehow dig up cash for it, and in a racing series where fractions of a second per lap matter (yes, that's not an exaggeration; watch an F1 race or qualifying session some time), those teams will demand the utmost out of every step of their design process. CFD and FEA are important tools in the multi-hundred million dollar series (link is not to a definition, but an article about one team's integration of FEA into their design process; even though not about the Twin Towers, it's actually a good on-topic read).

The other thing with F1 is you can no longer do on-track testing between races during the course of a season. So a wind tunnel is the only place you can validate performance of a real component outside of trying it in valuable practice/qualifying session time during a race weekend.

 

[tfk]

Prof. Steven Jones during a lecture at UC Davis, May 2009:

"A theory, which cannot be tested by other scientists, independently, is not true science."

try to listen at 15:20

.
Hi bio,

Do you not understand yet? Steven Jones is a rank amateur at Engineering, at Mechanics, at Structures, at Demolitions. At ALL of the subjects into which he has thrown himself.

He is a first rate particle physicist. His work in muon mediated cold fusion is considered (last time I heard) first rate.

He knows no more about mechanical engineering than I do about cold fusion.

This is precisely why he was run out of BYU, with the vast majority of the engineering faculty taking the unheard-of step of publicly writing a letter that stayed an eyelash away from academically censurable action. There are, hard to believe, stringent limits placed on how & when one academic can criticize another academic in public.

Jones will NEVER appear in front of a board of experienced engineers. He'd have his butt handed to him. And, if they were working engineers & not academics, then it'd be a bloodbath. One that I'd frankly pay to see. (Maybe a Pay Per View opportunity here. Texas cage match.)

Like science work, engineering work does need to be independently validated. And it is done constantly. But it cannot be done by amateurs. And it can absolutely not be done by people with political agendas.

Any honest, open researcher in engineering would be ecstatic to share his results, his raw materials, his methods, etc. with anyone that wanted to join in. Any honest open researcher would also make damn sure that his team was made up of competent professionals in the field in which he is investigation. The most accomplished professional that he/she could find.

Look at the group that Dr. Jones has put together. Competent & accomplished are two of the last adjectives that leap to mind.

Now look at the NIST reports. Look at the experts called upon. Look at their degrees. Look at their fields of expertise. Look at their numbers.

It's night & day.

Tom

 

[bio]

.
Hi bio,

Do you not understand yet? Steven Jones is a rank amateur at Engineering, at Mechanics, at Structures, at Demolitions. At ALL of the subjects into which he has thrown himself.

He is a first rate particle physicist. His work in muon mediated cold fusion is considered (last time I heard) first rate.

He knows no more about mechanical engineering than I do about cold fusion.

This is precisely why he was run out of BYU, with the vast majority of the engineering faculty taking the unheard-of step of publicly writing a letter that stayed an eyelash away from academically censurable action. There are, hard to believe, stringent limits placed on how & when one academic can criticize another academic in public.

Jones will NEVER appear in front of a board of experienced engineers. He'd have his butt handed to him. And, if they were working engineers & not academics, then it'd be a bloodbath. One that I'd frankly pay to see. (Maybe a Pay Per View opportunity here. Texas cage match.)

Like science work, engineering work does need to be independently validated. And it is done constantly. But it cannot be done by amateurs. And it can absolutely not be done by people with political agendas.

Any honest, open researcher in engineering would be ecstatic to share his results, his raw materials, his methods, etc. with anyone that wanted to join in. Any honest open researcher would also make damn sure that his team was made up of competent professionals in the field in which he is investigation. The most accomplished professional that he/she could find.

Look at the group that Dr. Jones has put together. Competent & accomplished are two of the last adjectives that leap to mind.

Now look at the NIST reports. Look at the experts called upon. Look at their degrees. Look at their fields of expertise. Look at their numbers.

It's night & day.

Tom

1. There are more and more architects and engineers (726), who give their reputation and name, and say essentially the same like Prof. Steven E. Jones. What are you guys here saying? "These architects and engineers do not work on "high-rise", steel-framed buildings, so they cannot say anything." Good - you cannot say anymore "you have no architects and engineers" .
2. Prof. Steven E. Jones published his expertise regarding 9/11 in peer-reviewed scientific magazines. What are you guys here saying: "Their peer-review were not scientific enough."
3. NIST does not publish all the data from their computer modelling. What are you saying? "Yes, but you have not the hardware to check it anyhow."
4. NIST is a goverment-run organisation! Do you really expect, that they will suspect their chief of mass-murder? So it could never become a "honest, open" research and the premise was indeed, that the towers did collapse due the the plane-impacts. NIST should somehow prove that and had a very hard time. Why is NIST so secretive otherwise?

What is "debunking" too often? Personal attacks combined with pseudo-argumentation under the aim of preventing the truth.

 

[Baylor]

1. There are more and more architects and engineers (726), who give their reputation and name, and say essentially the same like Prof. Steven E. Jones. What are you guys here saying? "These architects and engineers do not work on "high-rise", steel-framed buildings, so they cannot say anything." Good - you cannot say anymore "you have no architects and engineers" .
2. Prof. Steven E. Jones published his expertise regarding 9/11 in peer-reviewed scientific magazines. What are you guys here saying: "Their peer-review were not scientific enough."
3. NIST does not publish all the data from their computer modelling. What are you saying? "Yes, but you have not the hardware to check it anyhow."
4. NIST is a goverment-run organisation! Do you really expect, that they will suspect their chief of mass-murder? So it could never become a "honest, open" research and the premise was indeed, that the towers did collapse due the the plane-impacts. NIST should somehow prove that and had a very hard time. Why is NIST so secretive otherwise?

What is "debunking" too often? Personal attacks combined with pseudo-argumentation under the aim of preventing the truth.

Jones is a fraud. Why do you support the work of a fraud? He manipulated the photos yet you still hold on to his every word. It's impossible to use thermite to fell a skyscraper. Stoichiometry and common sense prove it. What's with the tenacity? Proof Jones is a fraud:
Jones's manipulated photo

Untouched photo

After you're done with that, chew on this for a while. I'm sure Germans would embrace mormons if they immigrated to Motherland.
http://space.crono911.net/EBook/805_Jones_Jesus.pdf


Bio: would it be acceptable if NIST had manipulated photos in order to deceive its readers? would you take Dr. Sunder seriously if he had written a paper about Jesus reappearing in ancient America?

 

[tfk]

bio,

There are two fields of expertise that would allow you to KNOW who & what to believe in all of this. You can get into either one of them. The first will take you about 35 years (4 years college, 3-5 years post grad & about 25+ years working experience) and will provide you expertise in one field. The second can be started today, and in about 2 months of effort, you'll be well on your way to becoming an expert in about 1000 fields. I'd recommend strongly that you choose option B.

The first field is structural engineering. The second is epistemology.

Learn epistemology. Learn it inside & out. Backwards & forwards.

Epistemology will tell you, in objective fields of study:
1. Believe only REAL experts.
2. Expertise is incredibly specific. (physicists & electrical engineers, chemical engineers, etc. are NOT experts in structural engineering)

Now, with that as a starting point...

1. There are more and more architects and engineers (726), who give their reputation and name, and say essentially the same like Prof. Steven E. Jones. What are you guys here saying? "These architects and engineers do not work on "high-rise", steel-framed buildings, so they cannot say anything." Good - you cannot say anymore "you have no architects and engineers" .

.
The real experts say that Steven Jones is a rank amateur (he is) who doesn't know what he is talking about. (He doesn't)
The real experts say that Gates is a rank amateur (he is) who doesn't know what he's talking about. (He doesn't)
The real experts say that 90+% of ae911t are rank amateurs (they are) who don't know what they're talking about. (They don't)
Epistemology will tell you that, in any field of study, there are a small number of cranks & incompetents. You need to learn how to identify them & to ignore them.
.

2. Prof. Steven E. Jones published his expertise regarding 9/11 in peer-reviewed scientific magazines. What are you guys here saying: "Their peer-review were not scientific enough."

.
Who cares? Jones is an amateur.
.

3. NIST does not publish all the data from their computer modelling. What are you saying? "Yes, but you have not the hardware to check it anyhow."

.
Nonsense.

NIST has transparently published their methods, their assumptions, their results & their conclusions.

Let me draw an analogy.
You are telling me that David Copperfield dematerialized the Statue of Liberty. I'm telling you "no, he didn't. It was a simple trick."

I'm telling you that he put the audience on a turn-table, aligned the stature between two towers with wide curtains extending to either side, closed a third curtain covering the statue, pivoted the table while the curtain was closed, and when the curtain opened, the two towers were pointed in a different direction. The side curtains prevented people from seeing the skyline move.

I've given you sworn statements from other experts (aka magicians), the audience members, from the camera crews & from the company that designed & built the stage. And you're telling me that, unless I provide the drawings & specifications for the stage, motor & control system, then my explanation is unbelievable.

And you're probably saying this, NOT because you actually believe it. But because some other charlatan WANTS you to believe that people can magically disappear & reappear statues. Because he's got another purpose for you & that particular belief.

Open your eyes & recognize your role in this farce, bio. You're being used.

I could tell you how this trick & the next trick & the next trick are done. But pretty soon, I'm gonna be tired of your pestering annoyance. Recognize the big picture. They are all tricks. Nobody does "real magic".

I could explain how this twoofer "fact" is wrong & the next twoofer fact and the next. But pretty soon, I'm gonna be tired of your ... (you get the picture.) Recognize the big picture, bio.
.

4. NIST is a goverment-run organisation! Do you really expect, that they will suspect their chief of mass-murder? So it could never become a "honest, open" research and the premise was indeed, that the towers did collapse due the the plane-impacts. NIST should somehow prove that and had a very hard time. Why is NIST so secretive otherwise?

.
This comment is infuriatingly "Young & Stupid", bio. Get your ego back in check. You & your generation are NOT the first honest, honorable generation on the planet. Have a bit of respect for the people that preceded you.

NIST is a bunch of Americans. Americans don't care if the guy that committed mass murder was the president. There would have been a lynching. And the engineers at NIST (& the 50 other companies & universities that helped with the analyses) would have been first in line with the pitchforks & torches.
.

What is "debunking" too often? Personal attacks combined with pseudo-argumentation under the aim of preventing the truth.

.
Look into that ego-bypass procedure, would ya, kid.

Tom

 

[Justin39640]

What is "debunking" too often? Personal attacks combined with pseudo-argumentation under the aim of preventing the truth.

Look into that ego-bypass procedure, would ya, kid.

Tom

the other day a truther called me a "zionist jew lover" and that 911 was a "jew false flag"

so i think bio has the right idea he just applied it to the wrong group

 

[Bluesky]

The last thing America wants to do is give the raw data to the ae911tm. We have all seen how they sculpt the truth, manipulate data, cherry pick comments, and focus on the irrelevant.

They could take the data and make a few discrete changes and it would take months for a team to track down and explain what they did wrong. Who pays for the team to explain, modify, and train them as they bumble there way through the analysis and contrive a scenario where failure does not occur?

And of course, we have seen what the ae911tm do with anything that fits their agenda: publish first and then ascertain truth and relevance later.

 

[bio]

Jones is a fraud. Why do you support the work of a fraud? He manipulated the photos yet you still hold on to his every word. It's impossible to use thermite to fell a skyscraper. Stoichiometry and common sense prove it. What's with the tenacity? Proof Jones is a fraud:
Jones's manipulated photo
(...)

what is your evidence, that Prof. Jones "manipulated" the photos?

 

[TruthersLie]

what is your evidence, that Prof. Jones "manipulated" the photos?

He just showed you the two images side by side.

Jones has used that photo in his speaches to prove "molten metal" when in fact it comes from a video that shows BEYOND A SHADOW OF A DOUBT that it was a FLASHLIGHT.

Yet Jones used those images as his "proof."

If he didn't manipulate them, then he didn't bother to do ANY research into a manipulated image, and then claimed they were something they were not.

don't you call that dishonest? unethical?

 

[bio]

He just showed you the two images side by side.

Jones has used that photo in his speaches to prove "molten metal" when in fact it comes from a video that shows BEYOND A SHADOW OF A DOUBT that it was a FLASHLIGHT.

Yet Jones used those images as his "proof."

If he didn't manipulate them, then he didn't bother to do ANY research into a manipulated image, and then claimed they were something they were not.

don't you call that dishonest? unethical?

it is dishonest to pretend, that Prof. Jones should have "manipulated" something although you have obviously no proof.

 

[Dave Rogers]

it is dishonest to pretend, that Prof. Jones should have "manipulated" something although you have obviously no proof.

Technically, bio has a point here. We don't know for certain that Jones is reprehensibly dishonest; he may, alternatively, simply be laughably incompetent. However, there isn't a third option.

Dave

 

[TruthersLie]

it is dishonest to pretend, that Prof. Jones should have "manipulated" something although you have obviously no proof.

ROFLMAO.

That is your defense? Really?

He took the image and used it as PROOF OF MOLTEN METAL for YEARS.

The image he used was ALTERED from the original.

That leaves 3 options.
1. He altered it.
2. He took an altered image and used it w/out doing any research (which would show where it came from and that it was altered)
3. He took an altered image and used it KNOWING it had been altered.

There is no, "oopsie I used the wrong image."

he continued to use it EVEN AFTER it was shown to be manipulated.

So either he is
1. dishonest
or
2. Incompetent

which one do you choose?

ETA: And it has been shown conclusively to be from a firefighters flashlight, yet he (s Jones) and crew have claimed it was proof of molten metal.

So again, who is being dishonest? (not that it is anything new for twoofs)