2 ай бұрын
3 ай бұрын
3 ай бұрын
4 ай бұрын
4 ай бұрын
4 ай бұрын
5 ай бұрын
5 ай бұрын
6 ай бұрын
6 ай бұрын
6 ай бұрын
6 ай бұрын
6 ай бұрын
6 ай бұрын
7 ай бұрын
Пікірлер
@BunrithViwatthanak 4 күн бұрын
I don't understand, because I'm a Assembly programmer who uses NASM x86_64 bit
@relaxmindset5957 6 күн бұрын
I wish you do quantom computing exact same you did for design a cpu, can't wait to see that's coming
@RossMcgowanMaths 6 күн бұрын
So I now have the full floating point unit plumbed into the cpu. I think it's quite a cool fpu working all functions on a single clock pulse and using a funky fast invesrse square root algorithm with a Newton Raphson iteration. Sexy !
@kirillpitolin1046 14 күн бұрын
I am testing this algorithm in matlab, I tried the example from the video, it is calculated correctly, for the sake of interest I decided to set "a" equal not to 4, but to 121, and at this point we are not even talking about accuracy, instead of the expected 11, I get 35... Maybe the algorithm has a limit on the input data, like when getting a sine/cosine?
@RossMcgowanMaths 11 күн бұрын
Yes the CORDIC functions are input limited. This is something that is outwith the scope of the course.
@arielseidman5548 Ай бұрын
In ~18 min, you taught me more than my professor has in 10 weeks. thanks
@RossMcgowanMaths Ай бұрын
Thank you.
@ThePaulTM Ай бұрын
Nice Work! There is also the rotation about the long diagonals this is how the Dodecahedron can be formed by five Cubes Golden Ratio to each other.
@RossMcgowanMaths Ай бұрын
Sounds interesting , I will have a look.
@Rnjeazy Ай бұрын
Beautiful!
@RossMcgowanMaths Ай бұрын
Thank you! Cheers!
@MrThegame000 Ай бұрын
Great video ! How can we get your animation ?
@RossMcgowanMaths Ай бұрын
www.desmos.com/calculator/1nngiw4uzv
@kierawhite6807 2 ай бұрын
17.125 isn't that 17+1/8 so 1001.001 before
@premkumarsr4021 2 ай бұрын
Is cordic better than taylor series in terms of runtime?
@RossMcgowanMaths 2 ай бұрын
CORDIC is typically more time-efficient in hardware with fixed-point arithmetic, while Taylor series may be faster in software when the CPU handles floating-point operations efficiently. The choice between the two depends on the specific context of the application (hardware vs. software, required precision, function being computed).
@roryawesometv6356 2 ай бұрын
yes
@CameronMcgowan-w4c 2 ай бұрын
cool
@RossMcgowanMaths 2 ай бұрын
A channel update. I'm still here and progressing with design a cpu 4 and also deign a cpu 5.
@w3math_channel 2 ай бұрын
Hello What software do you use to write with an electronic pen?
@RossMcgowanMaths 2 ай бұрын
Smoothdraw. It's free and very simple to use. Use a Gaomon S620 pen and tablet and a software app called Nebo which I use to do all my mathematical equations. Record video on OBS Studio.
@КалоянАтанасов-з6п 2 ай бұрын
Great course, as always! It would be awesome to see future courses dive into building a superscalar CPU that can handle multitasking and more advanced features. No matter what you decide to cover next however, I’ll definitely be looking forward to it when I have more time to dig in. Keep up the great work!
@RossMcgowanMaths 2 ай бұрын
Thank you. Very glad you like my work.
@nirselickter 2 ай бұрын
Thanks for the great courses you've created, I'm trying to do them at my own pace. I also like to see in some future features like multitasking , compile high level language to assembly (c language?), and maybe loading the cpu to some fpga evaluation board. Many thanks !🙂
@proxodik6573 3 ай бұрын
Thanks for video
@RossMcgowanMaths 3 ай бұрын
Most welcome
@studyfreak1655 3 ай бұрын
If want to calculate the thresold voltage of nmos inverter with saturated load can we consider Vgs as Vth ? Can you please make a detailed video on saturated load condition. Thank you so much
@bitdecay7773 3 ай бұрын
Would this course help in implementing my own sin/cos functions in C?
@RossMcgowanMaths 3 ай бұрын
Yes , as it gives you the mathematical algorithm to do just that.
@ninejaguar 3 ай бұрын
If this the 2nd Part of a 3 Part Introductory Mathematics course, where are Parts 1 and 3?
@RossMcgowanMaths 3 ай бұрын
I have still to create the first and third parts.
@Kilogya 3 ай бұрын
Literally the best explanation of how a CPU works. I'm learning C at the moment and thought that maybe making my own CPU would be a good idea if I want to learn how assembly and CPUs work. This series is a life saver since I'm 100% self taught on everything.
@RossMcgowanMaths 3 ай бұрын
Glad it helped!
@RossMcgowanMaths 4 ай бұрын
A quick video to show my progress and let you all know that I am busy with my digital designs.
@akhilesh.kallath 4 ай бұрын
This is very good. Makes very good understanding of the floating point binary numbers. Thank you very much.
@RossMcgowanMaths 4 ай бұрын
Glad it was helpful!
@RossMcgowanMaths 4 ай бұрын
I haven't really been visible over the last set of videos so I added this talking head video to remind everyone that there is a real person behind these videos and I do actually have a plan and all this mathematics is leading somewhere , exactly where at the moment is not completely clear to me , but this is part of the fun in doing something 'new'.
@timmorgan3673 4 ай бұрын
Hi Ross - It'll be really interesting to see all this implemented in "hardware" - Looking forward to it :)
@timmorgan3673 4 ай бұрын
A nice revision of Newton-Raphson - Cheers :) I was kinda just plugging in numbers blindly for this part ...
@RossMcgowanMaths 4 ай бұрын
Looks like the value 250 works well , I will try it out with say 1000 values in python and see how well it works. Tao be honest I'm not sure how simple / difficult it will be to implement in Logisim , I'm just kind ' winging it' at the moment.
@RossMcgowanMaths 4 ай бұрын
Actually 99.29 % when rounded
@RossMcgowanMaths 4 ай бұрын
SO IT DOES WORK !!!
@RossMcgowanMaths 4 ай бұрын
A correction to the last video.
@RossMcgowanMaths 4 ай бұрын
Ok an error in the equation Iy which was noted by Tim and still took my half an hour of brain scratching to work out for myself. the tern ax should be (a-1)x also note that the x is in fact the matissa m , I just called it x as I had to use the value x for it to work in desmos the graphing calculator. I will cover this error and the derivation if Iy in another video.
@timmorgan3673 4 ай бұрын
Fascinating Stuff :)
@RossMcgowanMaths 4 ай бұрын
Taking a bit of a 'tangent' here , pardon the pun. Extending the idea of a magic number to a magic line. Not sure if I am going anywhere with it , but the mathematics is interesting so let's travel down this road for a while and see where it takes us.
@RossMcgowanMaths 4 ай бұрын
At 5.38 I made a mistake interpreting the mantissa values , it is a value grater than or equal to zero and a number just less than 1 by a factor of 2^23
@RossMcgowanMaths 4 ай бұрын
I will be redoing this video.
@RossMcgowanMaths 4 ай бұрын
I have added another 3 videos which are scheduled to go live on the 2nd August 6th August and 8th August at 5pm covering the fast inverse square root method. Or you can get them now on my patron account www.patreon.com/RossMcgowanMaths if you are a paid member
@RossMcgowanMaths 4 ай бұрын
You don't have my image as my wife is using the camera today. Let me know if the sound is ok ?
@andreasproteus1465 4 ай бұрын
Sound is good. Videos are better without the talking head. Keep up the good work.
@timmorgan3673 4 ай бұрын
Hi Ross - My vote definitely goes to "Capitals" - Methinks that this is easier to read & also looks nicer IMHO ...
@RossMcgowanMaths 4 ай бұрын
Yes I think so too.
@RossMcgowanMaths 4 ай бұрын
Also I'm not quite in the middle of the screen , not sure how that happened , the camera is maybe a bit scewed.
@RossMcgowanMaths 4 ай бұрын
I have to fix my volume , I think it's a little low. I will drop down stairs now and put the rug back into the office room where I do my recordings as it gets rid of the tiny echoing sounds.
@CameronMcgowan-w4c 5 ай бұрын
yes
@timmorgan3673 5 ай бұрын
Lovin' it :)
@TalsBadKidney 5 ай бұрын
Big crowd pleaser right here
@RossMcgowanMaths 5 ай бұрын
The algorithm is called the Fast Inverse Square Root , in fact it doesn't work out the inverse but the reciprocal. Sometime in computer science we use the term inverse as it means the multiplicative inverse as opposed to the functional inverse.
@RossMcgowanMaths 5 ай бұрын
Yes , I deleted the audio music.
@fatlimey 5 ай бұрын
The 2016 paper from Moroz et al "Fast calculation of inverse square root with the use of magic constant" goes deeper into the Fast Inverse SQRT and finds smaller, and more accurate versions of the calculation. Available at arXiv:1603.04483
@roryawesometv6356 5 ай бұрын
oh hey the audio bug is gone!
@xarhspapapadatos 5 ай бұрын
I don't get why the rolling ball has only 2 degrees of freedom. The degrees of freedom of a system is defined as the number of independent variables that can vary due to a small displacement. When the ball is displaced, the euler angles have to change as well. Also now that I'm thinking about it, I don't think displacement is very well defined for a system with more than 1 point paricles.
@RossMcgowanMaths 5 ай бұрын
Been a while since I looked at this , but off the top of my head the degrees of freedom are the 2 independent axis , x and y.
@kirillpitolin1046 5 ай бұрын
do you think this is a good algorithm for implementing division in FPGAs or are there others?
@RossMcgowanMaths 5 ай бұрын
I don't think it is a particularly common algorithm , more common ones are digital recurrence , Newton-Raphson which I have covered in my assembly language course (Code the Machine) and Polynomial based division algorithms.
@patpol4885 5 ай бұрын
nice music
@RossMcgowanMaths 5 ай бұрын
As you move up the temperature axis the maximum Energy (density) occurs further along the frequency axis. This shows us that as we heat things up they will glow first at lower frequencies then higher frequencies in the E/M spectrum , that is red , orange , yellow , green , blue , violet.