Hi
My office hours today, Nov 7, are cancelled.
There will be a talk on numerical analysis that
is designed to be accessible to undergraduates today
at 3:30 in Gore 102. The speaker is Prof. Rick Falk
of Rutgers who will be speaking on
Mathematical Modeling and Numerical Approximation.
He is an award winning professor in mathematics and is
the PhD advisor of our current department chair,
Dr. Monk.
If anyone goes, please find me and let me know
you were there. I'll give 10 extra credit homework
points to anyone who attends.
RJB
Wednesday, November 7, 2007
talk this afternoon
miscellaneous for week of Nov 5
Hi
The homework due next week is posted.
The third project, on image compression using the
SVD, is posted. I have it due on the Wed before
Thanksgiving, but we can talk about this.
There is no lecture on Wed 11/21, but I can
come in and open the classroom to help you
finish it off if you like.
I will talk some about rootfinding before finishing up
with iterative methods. In terms of sections of the book,
I will finish with 5.3 and 5.4 today, and will at the least start 6.1
on Friday; I'll come back to 5.5.
For blogging, doing 6.1 for Friday would be good.
If you haven't already, 5.4 by then will be ok too.
6.2 for Monday will be good also.
The second exam for the course will be on Monday,
November 19th. More about the exam later.
RJB
The homework due next week is posted.
The third project, on image compression using the
SVD, is posted. I have it due on the Wed before
Thanksgiving, but we can talk about this.
There is no lecture on Wed 11/21, but I can
come in and open the classroom to help you
finish it off if you like.
I will talk some about rootfinding before finishing up
with iterative methods. In terms of sections of the book,
I will finish with 5.3 and 5.4 today, and will at the least start 6.1
on Friday; I'll come back to 5.5.
For blogging, doing 6.1 for Friday would be good.
If you haven't already, 5.4 by then will be ok too.
6.2 for Monday will be good also.
The second exam for the course will be on Monday,
November 19th. More about the exam later.
RJB
Thursday, October 18, 2007
macs and motivation (project 1)
Hi
Thank you to an alert student who supplied this info about Macs:
Since I am on a Mac, I thought I might just relay this information to you. To find the system info, all you have to do is bring up 'About this Mac' under the Apple menu. There will be a dialog box which will display basic processor and memory info but there is an option to display 'More info...'. This brings up an application called 'System Profiler.' There you have every piece of relevant hardware info you could wish for in 'Hardware Overview.' This does include size and speed of cache, the bus speed, etc. For RAM speed, you can click on the memory tab in the drawer to the left.
Why would you want cpu info? Say you buy a new laptop. How much faster is it than your old machine? And your next computer after that? Or, say you're trying to decide which computer to buy, and of course, running Matlab rapidly is now your top priority. You need as much info as you can get about this computer and its cpu to accurately compare one vs another to evaluate their relative performance.
You can find comparisons of cpu performance at places like www.tomshardware.com, which has been accused of having "the mother of all cpu charts."
RJB
Thank you to an alert student who supplied this info about Macs:
Since I am on a Mac, I thought I might just relay this information to you. To find the system info, all you have to do is bring up 'About this Mac' under the Apple menu. There will be a dialog box which will display basic processor and memory info but there is an option to display 'More info...'. This brings up an application called 'System Profiler.' There you have every piece of relevant hardware info you could wish for in 'Hardware Overview.' This does include size and speed of cache, the bus speed, etc. For RAM speed, you can click on the memory tab in the drawer to the left.
Why would you want cpu info? Say you buy a new laptop. How much faster is it than your old machine? And your next computer after that? Or, say you're trying to decide which computer to buy, and of course, running Matlab rapidly is now your top priority. You need as much info as you can get about this computer and its cpu to accurately compare one vs another to evaluate their relative performance.
You can find comparisons of cpu performance at places like www.tomshardware.com, which has been accused of having "the mother of all cpu charts."
RJB
Wednesday, October 17, 2007
project 1 info
Hi
A few things about the project.
Only use the sparse version of A with the builtin lu
command.
I posted a new lu_time_ex.m that got rid of two extraneous
parts: asym* should be deleted, and the defined v wasn't needed.
I must have been half asleep when I was trimming things out of
a bigger code when I posted that; sorry about the mistakes.
You can find out somethings about the hardware on strauss
using fpversion or uname -pX or uname -aX.
But, you only get general information about the fact that
it's a 12 cpu sparc processor. I'm not sure yet how to get
more info. I believe the cpu is a 750MHz clock and 150MHz memory.
On a pc, you can use dxdiag from the run program dialog box
and you will get info like the following.
My desktop is a 2.66 GHz Intel Core2 duo (2 processors)
model 6700 with 3 GB or memory. (I'm not sure of the speed of
the memory in my desktop yet.) Also of interest may be cache
memory for the cpu.
I don't have any idea yet how to get similar info for a Mac.
RJB
A few things about the project.
Only use the sparse version of A with the builtin lu
command.
I posted a new lu_time_ex.m that got rid of two extraneous
parts: asym* should be deleted, and the defined v wasn't needed.
I must have been half asleep when I was trimming things out of
a bigger code when I posted that; sorry about the mistakes.
You can find out somethings about the hardware on strauss
using fpversion or uname -pX or uname -aX.
But, you only get general information about the fact that
it's a 12 cpu sparc processor. I'm not sure yet how to get
more info. I believe the cpu is a 750MHz clock and 150MHz memory.
On a pc, you can use dxdiag from the run program dialog box
and you will get info like the following.
My desktop is a 2.66 GHz Intel Core2 duo (2 processors)
model 6700 with 3 GB or memory. (I'm not sure of the speed of
the memory in my desktop yet.) Also of interest may be cache
memory for the cpu.
I don't have any idea yet how to get similar info for a Mac.
RJB
Wednesday, October 10, 2007
Info for Fri Oct 12 Exam
Hi
Just so we're all on the same page, there will be an exam on this Friday, Oct 12 during our 11:15 to 12:05 lecture. I'll give you around 55 minutes or a little more, by trying to begin the exam early and ending a little late, but the exam will be designed for 50 minutes. I don't have a practice exam.
Topics that are fair game for the exam include: Conditioning (both functions and matrices), stability, norms, operation counts, LU factorization, Gaussian elimination, special matrices, linear least squares, normal equations, orthogonality. As stated in the syllabus, these are topics up to and including section 3.2.
You will be allowed one 8.5x11 inch paper (both sides) that is a cheat sheet for the exam. Write down definitions, properties, whatever you want on that sheet.
You may also use Matlab, though the test will not be designed to significantly test you on Matlab syntax, for example. You may be asked: what a given bit of matlab is computing and whether it is a desired quantity; operation counts in Matlab (simpler than the qrfact.m problem from the last homework); or it may be advantageous to be able to do a command or two in matlab (nothing like writing a function or script).
You are not allowed: calculators, any other software besides matlab, cell phones, or any other electronic device (including ipods). I'll supply plain paper and the exam.
Advice for preparing: (1) Be able to solve the homework problems. If you didn't get them the first time, go back and be able to do them. (2) Make sure you remember your linear algebra or work extra problems in it if you don't feel confident about it. (3) Work the beginning Matlab exercises if you haven't already. (4) Don't wait until the last minute to start studying. Start tonight, and stop by and ask me questions if things seem murky. (5) Check out other numerical analysis books if reading something different may be helpful.
Best wishes,
RJB
Just so we're all on the same page, there will be an exam on this Friday, Oct 12 during our 11:15 to 12:05 lecture. I'll give you around 55 minutes or a little more, by trying to begin the exam early and ending a little late, but the exam will be designed for 50 minutes. I don't have a practice exam.
Topics that are fair game for the exam include: Conditioning (both functions and matrices), stability, norms, operation counts, LU factorization, Gaussian elimination, special matrices, linear least squares, normal equations, orthogonality. As stated in the syllabus, these are topics up to and including section 3.2.
You will be allowed one 8.5x11 inch paper (both sides) that is a cheat sheet for the exam. Write down definitions, properties, whatever you want on that sheet.
You may also use Matlab, though the test will not be designed to significantly test you on Matlab syntax, for example. You may be asked: what a given bit of matlab is computing and whether it is a desired quantity; operation counts in Matlab (simpler than the qrfact.m problem from the last homework); or it may be advantageous to be able to do a command or two in matlab (nothing like writing a function or script).
You are not allowed: calculators, any other software besides matlab, cell phones, or any other electronic device (including ipods). I'll supply plain paper and the exam.
Advice for preparing: (1) Be able to solve the homework problems. If you didn't get them the first time, go back and be able to do them. (2) Make sure you remember your linear algebra or work extra problems in it if you don't feel confident about it. (3) Work the beginning Matlab exercises if you haven't already. (4) Don't wait until the last minute to start studying. Start tonight, and stop by and ask me questions if things seem murky. (5) Check out other numerical analysis books if reading something different may be helpful.
Best wishes,
RJB
Wednesday, October 3, 2007
Hw 6 posted, due Mon 10/8 5:30pm
Hi
Assignment 6 is posted on the course web page. I tried to
make it a little shorter, and it is due Mon 10/8 at 5:30pm.
I'll grade it for Wednesday, so that you will have it back
for the midterm that will be held on Friday Oct 12.
RJB
Assignment 6 is posted on the course web page. I tried to
make it a little shorter, and it is due Mon 10/8 at 5:30pm.
I'll grade it for Wednesday, so that you will have it back
for the midterm that will be held on Friday Oct 12.
RJB
Blogging procedure and week of 10/1
Hi
The usual procedure is to go by the syllabus for the
blogging. We are basically on schedule now, and I've
only been sending messages about it when I thought
we were off schedule
Belatedly, for tomorrow you should
write an entry for 3.3, and for Friday, write
one for 3.4.
For those of you who have been delinquent,
it really does correlate with a better grade
from my experience with doing this last
semester.
RJB
The usual procedure is to go by the syllabus for the
blogging. We are basically on schedule now, and I've
only been sending messages about it when I thought
we were off schedule
Belatedly, for tomorrow you should
write an entry for 3.3, and for Friday, write
one for 3.4.
For those of you who have been delinquent,
it really does correlate with a better grade
from my experience with doing this last
semester.
RJB
Subscribe to:
Posts (Atom)