Christopher2222

5790 Reputation

24 Badges

15 years, 313 days

MaplePrimes Activity


These are replies submitted by Christopher2222

@sursumCorda I've found the file and am uploading it again here.

1_ListBuffer.zip

@Joe Riel 

You have to start into your last name, your name came up after 10 seconds

Ok I did find my name

The key is to pick the author name in the drop down list.  If your name doesn't come up, you actually have to wait until the list repopulates - and you will see that when the cursor in the box stops blinking.  I had to wait 5 seconds after typing in my name before the cursor stopped flashing, then I had to wait another 10 seconds before the list repopulated and I could see my name to select. 

Also either include a date range or put a space in the keyword for an alltime list.  It's not necessary for the space in the keyword if a date range is specified.  But the main key is to have to select an author from the drop down list.

Just noticed a replies and answers tab under the profile.  This is good. 

@Joe Riel Hmm, I'm discovering some tricks that sort of make it work better.

In the author select box, it seems one has to pick an author that pops up in the list.  Just typing a name doesn't work.  But also to make it work - in the keyword box it can't be left blank, so a space bar has to be entered then it works.  My name Christopher2222 does not come up as an option and so far I've been unable to fine it. 

I did find your name "Joe Riel" so picking it and entering a space in the keyword box seems to pick them up. 

Thanks to both. 

I noticed in LinearAlgebra, there is a Map function which can have an optional [filter] which on the help page selects the positions for it to be applied.  Can that be used in this case?

@C_R Yes that seems to be the case.

So the reason behind doing the advanced search by author is because there doesn't exist a "Search all replies by ..." button.
Search all posts by ... and Search all questions by ... exists.

That looks like redshift vs. the velocity.  It can be done in a creative way.  I have done it in the past but after an extensive search on mapleprimes, I can't find what I was looking for. 

Basically, if I recall correctly, you have to convert it to an array or matrix then you can scale an axis and go from there.  I'll keep hunting for my idea, but mapleprimes search is somewhat limited (broken) so it's difficult to search properly.

@lemelinm I had a post on that many years ago.  Maple can create graphs like that, but it hasn't included any such features yet (I'm slightly surprised they haven't).  Anyways my post here is probably what you are looking for regarding the gridding in 3d space.  https://www.mapleprimes.com/posts/126570-Gridlines-In-3d-Space

So I put together a quick collection of routines for
  - Linear Algebra
  - Numeric Integration
  - Statistics
  - Polynomials

all with their reference author in brackets and a mapleprimes reference from where I got it from.

The idea is to use the collection to generate times for your system to compare to others, benchmarking.  Mainly just to see how fast a particular code is run in comparison to different hardware setups. 

Anyways, the initial test worksheet is attached.  Test_Maple_performance_-_suite.mw

Ok, this question about benchmarking has come up a number of times on mapleprimes.  Just earlier this year it was mentioned that Maple should have a benchmark test suite, to me that sounds like a great idea (a package could be put together). 

So, everyone uses Excel, it wouldn't be hard to collect data and tabulate it in an excel sheet and post it to the beginning of the thread.  Acer had a quick list to build from, Linear Algebra, Statistics, integration, differential equations, polynomial manipulation and special function evaluation. 

Based on that, can we come up with 6 seperate algorithms/routines to time?

Yes, we are somewhat timing the efficiency of the builtin routines between different versions ( and with many samples a pattern will emerge and we can conclude that "x" processor with "n" amount of RAM is better than "y" processor with "x" amount of RAM), but also across multiple system setups on the same version of Maple, the differences will also be noticed.  Of course the larger number of samples the better. 

Here's an example for Linear Algebra routine we could use that I dug up from a while ago by Roman Pearce

restart;
with(LinearAlgebra);
N := 9;
A := RandomMatrix(2^N, 2^N + 1, generator = -10^5 .. 10^5);
TIMER := time(Modular:-IntegerLinearSolve(A, 2^N));
print(2^N = TIMER);

On my machine the output is

By all means a rough draft, we can refine them as we go.

@acer You've obviously pointed out my method is ridiculous and terrible.  Someone else can come up with a more concrete way of my intentions, I quit. 

I'll start it off.  I've made one run here but we should make 5 runs and take the average for the result
Here's a sample test problem run with Maple 2022.0 on a Win7 64bit, 8Gb RAM, Toshiba Tecra S10, Intel Core2 Duo, P8700 @2.53GHz

restart;
UseHardwareFloats := true;
n := 3500;
t0 := time();
M := LinearAlgebra:-RandomMatrix(n, n, generator = 0. .. 1, outputoptions = [datatype = float[8]]);
randTime := time() - t0;
LinearAlgebra:-LA_Main:-Rank(M);
rankTime := time() + (-t0 - randTime);
total := rankTime + randTime;

There are performace comparisons across the different versions of Maple and competing products, but no real comparisons across various platform setups (which would be befificial in your case - but would also be interesting to others)

I wonder if we could set up a section or table somewhere (in a post) with specific test calculations and time them with various system steups by users in the community.  As a side post I will start it. 

@Carl Love I can confirm that works in MapleFlow.

As well the select command will be more compact

 

2 3 4 5 6 7 8 Last Page 4 of 155