[mujahid_e3]

(Original post by compy)
Lots of questions here: http://en.wikibooks.org/wiki/A-level...g/Real_Numbers

P.S. the attached doeesn't seem to be attached

thanks for your reply, will have a look at this..

have now attached the question... please help if u can.

thanks

[Yash13]

(Original post by mujahid_e3)
can someone explain the floating point numbers to me or give me a good link/tutorial that i can understand from.. my teacher didnt no how to explain it and just skipped it... i really need the help..

also can someone tell me how i answer the question (in attachment) its about floating point numbers...

all help appreciated

thanks

The book has these special numbers to remember. The lowest positive normalised value would have to be 0.1000000 for the mantissa and 1000 for the exponent.

Part b) I did these calculations :
Mantissa(m) x 2^exponent(e)
m= 1.011(dont worry about zeros)
e= 2
1.011 x 2^2=
101.1 = -4 + 1 + 0.5
= -2.5

[mujahid_e3]

(Original post by Yash13)
The book has these special numbers to remember. The lowest positive normalised value would have to be 0.1000000 for the mantissa and 1000 for the exponent.

Part b) I did these calculations :
Mantissa(m) x 2^exponent(e)
m= 1.011(dont worry about zeros)
e= 2
1.011 x 2^2=
101.1 = -4 + 1 + 0.5
= -2.5

thanks, but i never knew you actually have to remember that table in the book..

thanks for you working but how comes from 1.011 it became 101.1? and then later its -4 and not +4?

[compy]

(Original post by mujahid_e3)
thanks, but i never knew you actually have to remember that table in the book..

thanks for you working but how comes from 1.011 it became 101.1? and then later its -4 and not +4?

Yash is correct. Using the Noorgat Dance:

________1.011000000 0010
sign_____-ve
slide__________________+2
bounce__101.1
flip______010.1
swim ____2__.1/2 = -2.5 (remembering the sign from earlier)

the flip rule is where you convert the twos completement number if it's negative. In this case it is so Yash has applied twos complement to it, the method I've used above is slightly different

[mujahid_e3]

(Original post by compy)
Yash is correct. Using the Noorgat Dance:

________1.011000000 0010
sign_____-ve
slide__________________+2
bounce__101.1
flip______010.1
swim ____2__.1/2 = -2.5 (remembering the sign from earlier)

the flip rule is where you convert the twos completement number if it's negative. In this case it is so Yash has applied twos complement to it, the method I've used above is slightly different

oh ok i see.... need to learn this method

thanks

[mujahid_e3]

can anyone explain how i convert 13.625 in a floating point representation?
thanks

[exam2k10]

(Original post by mujahid_e3)
can anyone explain how i convert 13.625 in a floating point representation?
thanks

Got an answer of
Mantissa
0.1101101
Exponent
0100

Not sure if this is right though.
Where did u get the question from, so that I can check
Thanks

[compy]

(Original post by mujahid_e3)
can anyone explain how i convert 13.625 in a floating point representation?
thanks

From wikibooks:
1. work out the binary equivalent
2. work out how far to move the binary point (y)
3. set the exponent to be reverse of the number of places you moved the binary point (-y)
4. pad the number with extra bits

1. 13.625 = 01101.101
2. 0.1101101 = move 4 places to the left = -4
3. = -(-4) = 4 = 0100
4. 0.1101101 0100

[Kya]

(Original post by compy)
From wikibooks:
1. work out the binary equivalent
2. work out how far to move the binary point (y)
3. set the exponent to be reverse of the number of places you moved the binary point (-y)
4. pad the number with extra bits

1. 13.625 = 01101.101
2. 0.1101101 = move 4 places to the left = -4
3. = -(-4) = 4 = 0100
4. 0.1101101 0100

Nice way of explaining the method tbh. Do you have one for negative numbers? Since its always them that confuse me!


This was posted from The Student Room's iPhone/iPad App

[compy]

(Original post by Kya)
Nice way of explaining the method tbh. Do you have one for negative numbers? Since its always them that confuse me!


This was posted from The Student Room's iPhone/iPad App

There's an example at the bottom of this page: http://en.wikibooks.org/wiki/A-level..._point_numbers

[FredrickTrott]

(Original post by exam2k10)
Got a high B last year . So I need roughly 70/100 for an A. Thing is, this is such a long module, and I am gonna be doing Computer Science at university. Luckily, it wont be just pure Computer Science, I can branch out after the first year.

I plan on just going over notes today, tomorrow and tuesday, and then just start doing all the CPT4, CPT5 papers

Aah cool, i'm hoping to do Computing at Uni as well . What did you do for your project if you don't mind me asking?

[SecondHand]

I'm going to do Computer Science too..

Ok, P10 from the textbook.

Explain why using the pigeonhole principle why a lossless compression algorithm that makes one input file smaller will make some other input file bigger.

Anyone?

[exam2k10]

(Original post by FredrickTrott)
Aah cool, i'm hoping to do Computing at Uni as well . What did you do for your project if you don't mind me asking?

No its kl. Got 70/75 . What about you?

BTW what university (firm) are you going to go to if u don't mind me asking?

I really hope COMP3 is not gonna be bad. I'm ok with networks. Just need to do some practice on real numbers, and databases. OS is just pure theory. Programming paradigms is just recap.

[exam2k10]

(Original post by SecondHand)
I'm going to do Computer Science too..

Ok, P10 from the textbook.


Anyone?

Hi Secondhand,

I'm not sure if this is right but I will explain anyway.

Basically the pigeonhole principle works like this. For n objects, there will be n-1 outcomes. Good example will be a football match. 6 teams enter for a tournament. Only 5 matches can be played with these 6 teams.

Now to your question about the loseless compression. Because there are more objects than outcomes, I assume that it will be large.

Like I said, Im not sure if I am right. Would be interesed to see, as they could potentially ask about this in our exam

[SecondHand]

On another note, Fig 1.4.1 on P35.

[SecondHand]

(Original post by exam2k10)
Hi Secondhand,

I'm not sure if this is right but I will explain anyway.

Basically the pigeonhole principle works like this. For n objects, there will be n-1 outcomes. Good example will be a football match. 6 teams enter for a tournament. Only 5 matches can be played with these 6 teams.

Now to your question about the loseless compression. Because there are more objects than outcomes, I assume that it will be large.

Like I said, Im not sure if I am right. Would be interesed to see, as they could potentially ask about this in our exam

I think the pigeonhole principle is better defined as - if you have n outcomes and n+1 objects then at least two must have the same outcome.

Assume: A lossless compression algorithm always reduces the size of the size.
This leads the the reasoning, if an input is of length n bits and a lossless algorithm is applied it will be reduced to it then an input between 1 and 2^n-1 bits will be created but how many inputs are there of n bits? There are precisely 2^n. How many inputs are there of 1->n-1 bits? 2^n-1. Therefore if such an algorithm existed then at least two inputs of length n would become the same file when compressed. This contradicts our original hypothesis that a lossless algorithm always reduces the size of the input as if an input got compressed to the same bit pattern then it would be the same even though the original would be clearly different, information would have been lost. Therefore for some inputs the file size will increase.

Any clarifications don't hesitate to ask.

[FredrickTrott]

(Original post by exam2k10)
No its kl. Got 70/75 . What about you?

BTW what university (firm) are you going to go to if u don't mind me asking?

I really hope COMP3 is not gonna be bad. I'm ok with networks. Just need to do some practice on real numbers, and databases. OS is just pure theory. Programming paradigms is just recap.

I made a Maze Solver, and got 73/75 (Lost 2 on my Testing ). I've firmed Imperial College London, and Insured UCL, what about you?

I need 102 UMS in Comp 3 to get 90% overall at A2, and 102 translates to a High A, so i am hoping this paper doesn't surprise me with too many questions i don't know.

[FredrickTrott]

Is anyone in particular going to be making Post Exam Solutions for this, or could we somehow figure out a system where we all do a question, and post it on here to get the full paper?

[Yash13]

(Original post by FredrickTrott)
I made a Maze Solver, and got 73/75 (Lost 2 on my Testing ). I've firmed Imperial College London, and Insured UCL, what about you?

I need 102 UMS in Comp 3 to get 90% overall at A2, and 102 translates to a High A, so i am hoping this paper doesn't surprise me with too many questions i don't know.

Hey I got the same grade for cwk. I was quite happy before I realised I still have to do excellent in the exam for an A*. 102 is mid A tbf but still

[SecondHand]

How do you guys know your UMS for your coursework? Did you submit it in January?

Do you have one of them crazy offers from Imperial?