Timing normal 32K RAM vs 32K 16bit RAM

[RXB]

1 minute ago, apersson850 said:

You believe this just because, as you have showed before, you don't understand math. The fact that the time was 45 for both shows that the difference is no more than one second. Which is about 2%. Which is why we say the difference is so small that it's neglectible, since that's what we consider 2% to be. Nobody ever said there wouldn't be any difference. Just that it's too small to be important under these circurmstances, i.e. when using BASIC.

I think you are a troll and have proven you do not care about accuracy at all, just your bloody opinion!

[apersson850]

14 minutes ago, RXB said:

NO! I indicated that as you stated normal RAM is not as fast as 16bit RAM there is a difference.

Never once did I say what you just said, not once.

Oh yes you did. You said we needed a larger sample. Larger samples are only meaningful to even out variations between the samples. So you claimed there are variations, or the larger sample wouldn't be needed.

But you said this since you don't understand math, or at least not statistics.

 

14 minutes ago, RXB said:

What I said over and over and over is your test that got 45 vs 45 sucks and is worthless as a test as it does not even show the difference you yourself says exists.

There for to expand that difference in each RAM versions (i.e. 16bit RAM vs Normal RAM) you need a loop to expand that speed difference by using a loop.

As we've explained numerous times, it's enough to show that the difference is no more than 2%. Which is all it takes, at least for those who understand math.

[apersson850]

23 minutes ago, RXB said:

"On the other hand, INT(3.14159) is indeed the same as 3, but that has nothing to do with what we're discussing."

 

Ok that is 100% proof you are just flat out trolling!

That is a rounding error and would flunk you in math, no one asked for a rounding, what is needed is factual numbers not rounding to make you happy!

Tell you what try to use 3 for the rest of your life in programing for PI!

Oh, the master of trolling here is yourself, Rich. It was you who stated that there was a difference between int(3.14159) and 3, which there isn't. But you probably would have written (had you understood math) that there's a difference between pi and 3, which there of course is. It's just that it's irrelevant in this case, as each loop takes the same time (give or take a minute difference the interrupt services may imply).

 

Try this: Take PI and calculate the average of it. You'll get PI. Now take one hundred PI and calculate the average of them. Since PI*100/100 is still PI, you get the same result.

Edited June 10, 2022 by apersson850

[apersson850]

7 minutes ago, RXB said:

it seems at this point that I am being bullied for entertainment purposes.

None of the arguments thus far are for accurate math or but instead want to go for rounding estimations, like judging a leap of faith.

Sorry that is never accurate, I like facts and accuracy.

Perfect example is 3 is not PI and never has been.

 

A am asking for ACCURACY not GUESSTIMATES!

If you are bullied then it's just because you persist in claiming things it should be obvious that you do understand are wrong.

 

We are saying that if you time something with one second resolution, a runtime of 50 seconds is enough to find a variation as small as 2%. As soon as you understand that, you'll grasp the whole thing. We are not interested in any better accuracy than that, since we consider that 2% is too little a difference to be important. Nobody has ever claimed there's no difference, just that it's too small to be important.

[apersson850]

16 minutes ago, RXB said:

I think you are a troll and have proven you do not care about accuracy at all, just your bloody opinion!

You see that statement in the mirror, right?

 

My opionion, bloody or not, is that an inaccuracy of 2% is good enough to prove the point here.

Now, in an attempt to return to some seriousness, what I'm actually trying to do is to get into your head that there's nothing wrong with the calculations we've shown you. I fail to believe that somebody talented enough to create RXB wouldn't comprehend that, if he just start thinking about it in the right way. It seems you've wound yourself up in the wrong direction here, so you just don't see the obvious.

Edited June 10, 2022 by apersson850

[RXB]

11 minutes ago, apersson850 said:

Oh yes you did. You said we needed a larger sample. Larger samples are only meaningful to even out variations between the samples. So you claimed there are variations, or the larger sample wouldn't be needed.

But you said this since you don't understand math, or at least not statistics.

 

As we've explained numerous times, it's enough to show that the difference is no more than 2%. Which is all it takes, at least for those who understand math.

And again a LOOP would show the difference. 

You know how I know you are trolling?

My girlfriend has been reading these and understand the loop would show the difference and she can not even program.

Just using simple math, but you refuse to even acknowledge this and want to say PI=3 rounding something no one with a brain would do.

[RXB]

10 minutes ago, apersson850 said:

You see that statement in the mirror, right?

 

My opionion, bloody or not, is that an inaccuracy of 2% is good enough to prove the point here.

Since when has your opinion is supposed to be taken as FACT? 

45 vs 45 is not proof, matter of fact it is the opposite of what you claim they are the same.

Same goes for PI=3 they are not the same as PI=3.14159, they are not the same and one is more accurate. PERIOD!

[apersson850]

1 minute ago, RXB said:

And again a LOOP would show the difference. 

You know how I know you are trolling?

My girlfriend has been reading these and understand the loop would show the difference and she can not even program.

Just using simple math, but you refuse to even acknowledge this and want to say PI=3 rounding something no one with a brain would do.

Yes, a loop would show the difference. Even the most simple math knowledge would give you that. But take one step further, and you realize that 45 seconds, with a resolution of one second, shows that the difference is max one second. If it was more, it would be 45 and 46 seconds, at least.

And that's all it takes! We need not know better than that the difference is no more than 2%, since it gives us a max value it can be. We all know that there is some difference, since the faster memory will create that. There is no need to show that. Just how little is is.

If we wanted to know if the difference is smaller than one percent, then yes, then we would need one more loop or two.

[apersson850]

2 minutes ago, RXB said:

45 vs 45 is not proof, matter of fact it the opposite of what you claim they are the same.

Because the resolution is known to be one second, it's proof that the difference is no more than one. Which is all it takes. PERIOD!

[RXB]

1 minute ago, apersson850 said:

Yes, a loop would show the difference. Even the most simple math knowledge would give you that. But take one step further, and you realize that 45 seconds, with a resolution of one second, shows that the difference is max one second. If it was more, it would be 45 and 46 seconds, at least.

And that's all it takes! We need not know better than that the difference is no more than 2%, since it gives us a max value it can be. We all know that there is some difference, since the faster memory will create that. There is no need to show that. Just how little is is.

If we wanted to know if the difference is smaller than one percent, then yes, then we would need one more loop or two.

Just admit your opinion matters more to you then being accurate and factual.

[apersson850]

2 minutes ago, RXB said:

Just admit your opinion matters more to you then being accurate and factual.

That the difference in no more than one second is not an opinion. It's a fact. Admit that, to begin with. One is at least 45 and the other is not as much as 46. Admit that, to begin with.

Once you do that, I'll admit something else.

Edited June 10, 2022 by apersson850

[RXB]

1 minute ago, apersson850 said:

Because the resolution is known to be one second, it's proof that the difference is no more than one. Which is all it takes. PERIOD!

45 vs 45 says the same. So Proof that does not work is proof. PERIOD?????

[youxia]

Alright, like I said before this has been fun for a while, but you two are in an infinte loop yourselves, and it's getting ugly. Time to hit IGNORE THREAD....

 

(btw, the only way to "solve" your argument , after several pages of going in circles, is for one of you  - or both - to stop posting)

[RXB]

Just now, apersson850 said:

That the difference in no more than one second is not an opinion. It's a fact. Admit that, to begin with. One is at least 45 and the other is not as much as 46. Admit that, to begin with.

But you do not know the difference, as you are using ROUNDING to discount that difference. 

You have no idea factual what it is as you are GUESSING which I call a GUESSTIMATE!

[RXB]

1 minute ago, youxia said:

Alright, like I said before this has been fun for a while, but you two are in an infinte loop yourselves, and it's getting ugly. Time to hit IGNORE THREAD....

 

(btw, the only way to "solve" your argument , after several pages of going in circles, is for one of you  - or both - to stop posting)

Ok fine, just a simple request for factual accuracy is a problem I guess.

[apersson850]

1 minute ago, RXB said:

45 vs 45 says the same. So Proof that does not work is proof. PERIOD?????

No, it doesnt' say they are the same, since the resolution is one second. We don't know if one is 45.12345 and the other is 45.6789. We only know that both are at lest 45.00000 and not more than 45.99999. Admit that.

[apersson850]

4 minutes ago, RXB said:

But you do not know the difference, as you are using ROUNDING to discount that difference. 

You have no idea factual what it is as you are GUESSING which I call a GUESSTIMATE!

No, we are not rounding, since we indeed do have a very good idea about the actual value. It's 45 <= x < 46. We know that for a fact, and that accuracy is good enough. At least for the rest of us.

Thus we know for a fact that the difference can't be bigger than 46/45, which is roughly 2% (that's rounding), and that's good enough to know.

Edited June 10, 2022 by apersson850

[RXB]

2 minutes ago, apersson850 said:

No, it doesnt' say they are the same, since the resolution is one second. We don't know if one is 45.12345 and the other is 45.6789. We only know that both are at lest 45.00000 and not more than 45.99999. Admit that.

Sure correct you do not know, you only know your test does not work and is at least off by under a second, so you are guessing the value.

But a loop would tell you how much to the millisecond which I guess you do not want to know.

[RXB]

2 minutes ago, apersson850 said:

No, we are not rounding, since we indeed do have a very good idea about the actual value. It's 45 <= x < 46. We know that for a fact, and that accuracy is good enough. At least for the rest of us.

What rest of us, you speak for entire world now?

[senior_falcon]

Over and over and over, I have invited you to do your own tests. That way you can provide the answer to the question that the entire world is waiting for with bated breath: Is a TI99 running on expansion memory 1.2% or 1.4% faster than an unexpanded console. You have all the tools you need, yet you refuse to do it. Why?

Instead, you object constantly to everyone else's points. I am beginning to think you may be the Irving Kanarek of retrocomputing.

[HOME AUTOMATION]

2 hours ago, RXB said:

Who asked for rounding down to ignore difference? Who?

In the name of decency, I'm going out on a limb here...

22 hours ago, RXB said:

Sorry but this is like saying INT(3.14159) is the same as 3, is this your argument?

...Perhaps this is not exactly the same as asking for rounding down to ignore difference.

I too assumed you meant to say "Sorry but this is like saying PI(3.14159) is the same as 3, is this your argument?"

 

Maybe there is an alternate context that I'm missing.

[RXB]

1 hour ago, senior_falcon said:

Over and over and over, I have invited you to do your own tests. That way you can provide the answer to the question that the entire world is waiting for with bated breath: Is a TI99 running on expansion memory 1.2% or 1.4% faster than an unexpanded console. You have all the tools you need, yet you refuse to do it. Why?

Instead, you object constantly to everyone else's points. I am beginning to think you may be the Irving Kanarek of retrocomputing.

ENTIRE WORLD? Is everyone on Earth on Atari Age and all are watching this?

Accuracy does not matter much to you, does it?

All this is about to me is an argument over being accurate and factual, not guesswork passed off as fact.

Same as PI is not int(3) it is useless as it is not factual or accurate at all.

You refuse to admit that accuracy matters, instead your personal incredulity as only your opinion matters.

Facts be dammed. As for Irving Kanarek I am not a lawyer or would I defend Charlie Manson so insult taken.

Again facts matter not opinion!

[RXB]

47 minutes ago, HOME AUTOMATION said:

In the name of decency, I'm going out on a limb here...

...Perhaps this is not exactly the same as asking for rounding down to ignore difference.

I too assumed you meant to say "Sorry but this is like saying PI(3.14159) is the same as 3, is this your argument?"

 

Maybe there is an alternate context that I'm missing.

Make it simple they are not the same and one is impossible to use as PI. 

Try using 3 as PI in any calculation and watch how bad the result is!

[apersson850]

1 hour ago, RXB said:

Sure correct you do not know, you only know your test does not work and is at least off by under a second, so you are guessing the value.

But a loop would tell you how much to the millisecond which I guess you do not want to know.

On the contrary. The test works, since it does show that the difference is under one second. Within that second, we don't know. But there's no guessing, as we say the difference is no more than about 2%, and that we know for a fact. We don't guess if it's 0.87% or 1.452%. That would indeed be guesswork.

The only opinion I have here (not fact) is that 2% is too little a difference to be important. Thus there's no need to increase the accuracy. If you instead would think it's important to know if it's 1.336% or 1.342%, then for sure, more loops are needed. But it's pointless. It doesn't matter if you wait 45.3 seconds or 45.6 seconds.

 

And no, I'm not talking about the entire world, but the rest of us at this forum. That was implied. Sorry if it wasn't as obvious as I presumed.

[Reciprocating Bill]

1 hour ago, RXB said:

Sure correct you do not know, you only know your test does not work and is at least off by under a second, so you are guessing the value.

We can easily make it more accurate. For example, the maze executes 10x in 451 seconds with memory expansion and in 461 seconds without on a stock console. So now we know, to 1/10 second accuracy, that a single execution requires 45.1 seconds with, 46.1 seconds without memory expansion. That's a 2.2% difference. 

Edited June 10, 2022 by Reciprocating Bill