DirectCompute vs. CUDA vs. OpenCL

[1.1.2.3.5...]

I've repeatedly invited and ASKED you to provide your opinion. You've repeatedly ignored the request. You haven't questioned my position, you've nitpicked a statement to hell and back.

I do not care about your opinion. Understand that? I do not really want to discuss my opinion, my opinion is completey beside the point and of no consequence.

But if you honestly care and it will move this discussion forward then.... it is pretty much in line with most of your opinion (or what I think your opinion is) on this point. Except my definition of failure is not in line with your by any measure. This is not a sublet point...it is critical.

If you frame a discussion under false premises..the discussion is all invalid.

Get this straight...I would prefer if there were neutral open standards in common use for gpgpu computing. I think we agree on most or all of our opinions...I cannot be sure though because you are so frickin impossible to understand. In fact...I have had a more difficult time understanding you than any book, any person or anything ever in all my life. And my very high gpa in undergrad and grad school tells me I am good at understanding things. You take the cake for having the most obfuscated implied meanings in seemingly simple sentences.

I still have no clue what your statement is supposed to mean at this point. If it is just an opinion.....then say so. If you are stating it as an objective fact...beyond dispute...as solid a piece of knowledge as anything we humans can know...then I expect proof.

My opinion will have NO impact on the truth value of your statement. If I agree with your statement, it will remain unproven...you do understand that, correct?

I would even prove or disprove it for you. Just give me what I need...make your statement precise and define all terms.

If you think I am attacking your position or you, then you failed to understand EVERYTHING I have said.

Your statement can be false, and you can still have a valid position. Your statement can be true, and you can still have an invalid position. Your position does not hinge on this statement.

But you want to defend your statement to the death, but you do not want to actually prove it. You asked for examples of fallacies you committed....Burden of proof....

That is an informal logical fallacy. Look it up. It means the burden of proof is on you. You cannot give me an analogy as proof. That does not prove anything, but can give evidence to support your claim. Also, you cannot say that since I have not disproved your statement, then it is true. You must to prove it all on your own or you have committed that logical fallacy.

If I am wrong on this point...I should be fired and be striped of my degrees because this is a very minor and elementary piece of logic.

 

[kllrnohj]

I do not care about your opinion. Understand that? I do not really want to discuss my opinion, my opinion is completey beside the point and of no consequence.

Then why are you even posting in an opinion thread?

But if you honestly care and it will move this discussion forward then.... it is pretty much in line with most of your opinion (or what I think your opinion is) on this point. Except my definition of failure is not in line with your by any measure. This is not a sublet point...it is critical.

What is your definition then?

Get this straight...I would prefer if there were neutral open standards in common use for gpgpu computing. I think we agree on most or all of our opinions...I cannot be sure though because you are so frickin impossible to understand. In fact...I have had a more difficult time understanding you than any book, any person or anything ever in all my life. And my very high gpa in undergrad and grad school tells me I am good at understanding things. You take the cake for having the most obfuscated implied meanings in seemingly simple sentences.

And I find you hard to understand. Out of curiosity, is English your native language? If so, where are you from?

And I'd like to point out that others here got what I was saying without much difficulty.

I still have no clue what your statement is supposed to mean at this point. If it is just an opinion.....then say so. If you are stating it as an objective fact...beyond dispute...as solid a piece of knowledge as anything we humans can know...then I expect proof.

Of course its an opinion. I said it, and I didn't call it a fact.

My opinion will have NO impact on the truth value of your statement. If I agree with your statement, it will remain unproven...you do understand that, correct?

Yes, but the entire point of forums is to share opinions and information.

But you want to defend your statement to the death, but you do not want to actually prove it. You asked for examples of fallacies you committed....Burden of proof....

Burden of proof I can agree with, but you posted a huge list of things I didn't do.

That said, many things being discussed simply can't be proven one way or the other. For example, did Nvidia pay the JC2 devs to include CUDA features? You can't prove that one way or the other since none of us truly know the answer.

Also, you cannot say that since I have not disproved your statement, then it is true.

Agreed.

 

[1.1.2.3.5...]

Then why are you even posting in an opinion thread?

To discuss the facts....

And I find you hard to understand. Out of curiosity, is English your native language? If so, where are you from?

Born Minneapolis and raised in Orange County, going to school in the mid west. I have a fair sampling of English.

And I'd like to point out that others here got what I was saying without much difficulty.

I do not think they do...

Of course its an opinion. I said it, and I didn't call it a fact.

Not until now....unless I missed it...this is all a complete waste of time.

Yes, but the entire point of forums is to share opinions and information.

Do you mean that all that is discussed in forums are opinion? I spend most of my time...

www.physicsforums.com

Burden of proof I can agree with, but you posted a huge list of things I didn't do.

It does not really matter...you did them to some degree or another, but I do not honestly care.

That said, many things being discussed simply can't be proven one way or the other. For example, did Nvidia pay the JC2 devs to include CUDA features? You can't prove that one way or the other since none of us truly know the answer.

They can be proven true or false at least in theory. Opinions are not facts and they cannot be prove one way or another. If your statement is your opinion on the matter...fine. I do not care about that.

But for the record...it has not been proven that: if nv had to pay devs to use cuda, then cuda is a failure.

I am soo over this waste of time discussion....

 

[phide]

But for the record...it has not been proven that: if nv had to pay devs to use cuda, then cuda is a failure.

That is his opinion. "Failure" is relative and variable in terms of its definition and subject to personal interpretation.

Personally, I find it hard to label CUDA a failure in any respect simply because it hasn't seen any natural adoption in one sector, particularly when it's flourishing in others. CUDA is still out there, being well-supported by NVIDIA and waiting to be utilized for games. It will not be ready to be labelled a failure until the time has come in which CUDA is no longer an available means of implementing GPGPU functionality.

Labeling it a failure is a touch premature. It may not be a success, but I fail to see how it can be viewed as a failure.

 

[1.1.2.3.5...]

That is his opinion. "Failure" is relative and variable in terms of its definition and subject to personal interpretation.

I asked for clarification and for more specific terms.

Personally, I find it hard to label CUDA a failure in any respect simply because it hasn't seen any natural adoption in one sector, particularly when it's flourishing in others. CUDA is still out there, being well-supported by NVIDIA and waiting to be utilized for games. It will not be ready to be labelled a failure until the time has come in which CUDA is no longer an available means of implementing GPGPU functionality.

Labeling it a failure is a touch premature. It may not be a success, but I fail to see how it can be viewed as a failure.

Yep...I have said this. Mostly your last line...

 

[LeninGHOLA]

Start again, from the top!

 

[MrWizard6600]

stop arguing with mathematics sequences kllrnohj, particularly those with lim x->∞ F(x) = ∞ (if you argue forever he will too).
edit:

Do you mean that all that is discussed in forums are opinion? I spend most of my time...

www.physicsforums.com

lol, physics nerds and compsci nerds doing battle; those wondering his name is Fibonacci series.

Developers use CUDA over CL or DirectCompute because CUDA is a good product. Nvidia's pushing it, and its proprietary. The fact that between 50% and 70% of dev's target demographic isn't CUDA capable means CUDA based effects are never going to penetrate game development in any really serious way.

We have yet to see this whole concept of GPGPU's and parallel computing really hit the market in any meaningful way, and its going to take time for any code house to adopt: this isn't just a change in feature sets its a whole paradigm shift.

Cool stuff is coming, but for now, wait, or get invovled.

 

[gescom]

If Nvidia needs to pay developers to use CUDA, then CUDA is a failure, pure and simple.

It's simply just a part of Nvidia (PR) strategy. Nothing wrong with it I guess - our way or the money way.

 

[gescom]

I know for a fact CUDA is selling some cards. I know of several places with ten of thousands of dollars worth of NV telsa cards. All to run CUDA apps.

Out of curiosity - stock exchange market? post-production?

 

[trinibwoy]

Take it a few steps forward. If a game was released that is solely for nVidia users, would that be considered fair? Using the same reasoning, ATi users are not getting cheated out of an experience, because it wouldn't have been there without nVidia's help anyways. Even if a small amount of content is missing, you're applying the same principle.

If nVidia were to continue this exaggerated trend, it would force ATi out of the market simply because nVidia has more power and sway in the game developing world. Correct me if I'm wrong, but that's not legal.

A software developer isn't legally bound to write code that can run on multiple platforms. Hence console exclusives. Even Microsoft isn't obligated to write software that runs on everything.

I think the general point, however, is that if you have software that is better than the competition's, you shouldn't have to pay people to use it.

Subsidizing adoption of a new product is very common. Ever try free samples in the supermarket? Similiar concept. It's not just whether it's better than everything else. There is an initial cost to a developer to learn/adopt a new technology so obviously incentives will help get past that cost of entry. I'm evaluating a software solution right now and the vendor is bending over backwards to give me everything I need during the evaluation period including free support, licenses etc. That's a cost to them that they are willing to incur in order to win future business.