Differences

This shows you the differences between two versions of the page.

--- haas:fall2017:discrete:projects:pnc2 [2017/07/13 20:36] – wedge
+++ haas:fall2017:discrete:projects:pnc2 [2017/09/18 15:58] (current) – [Evaluation Criteria] wedge
@@ Line 6: / Line 6: @@
 ~~TOC~~
-======Project: ALGORITHMS - PRIME NUMBER CALCULATION (pnc2)======
+======Project: ALGORITHMS AND OPTIMIZATIONS - PRIME NUMBER COMPUTATION (pnc2)======
 =====Errata=====
@@ Line 14: / Line 14: @@
 ====Revision List====
-  * revision 0: initial release (20170713)
   * revision #: <description> (DATESTRING)
@@ Line 29: / Line 28: @@
   * if optimizing for (or trying to take up less) space, the program runtime can grow.
-Let it be clear: using modern computers bound by fourth dimensional constraints, we **cannot** write solutions without some foothold in both space and time (any program has statements in it that accomplish the task at hand (space occupied), and that same program takes time to run (even if it happens rather quickly)). What we can do is influence where on the space-time spectrum our implementation falls, and if done strategically for the particular constraint at hand, can aid us in achieving desired improvements in program efficiency.
+Let it be clear: using modern computers bound by these so-called fourth dimensional constraints, we **cannot** write solutions without some foothold in both space and time (any program has statements in it that accomplish the task at hand (space occupied), and that same program takes time to run (even if it happens rather quickly)). What we can do is influence where on the space-time spectrum our implementation falls, and if done strategically for the particular constraint at hand, can aid us in achieving desired improvements in program efficiency.
 In the previous projects, we focused on algorithms that were more constrained by time- taking progressively more time the larger the data set to be processed became. These time-constrained algorithms also happen to be the type of algorithm you're most familiar with (or at least, were indoctrinated into thinking with...). Your third grade self, on the other hand,  potentially would have found more familiarity with the space-constrained methods this project will explore.
@@ Line 47: / Line 46: @@
   * unless otherwise specified, you are doing all values (odds AND evens).
-  * if you exceed some threshold of values get "--error!---" from "**make check**" that's okay: I have memory limits in place (just like I have time limits in place).
+  * if you exceed some threshold of values and get "--error!---" from "**make check**" that's okay: I have memory limits in place (just like I have time limits in place).
-  * as the space-based approach to solving our prime number computation problem is fundamentally different from those we took in previous projects, you will want to look to start from scratch. Trying to reuse old code and shoehorn it into what effectively amounts to an entirely different paradigm will create a lot of problems.
+  * as the storage-based approach to solving our prime number computation problem is fundamentally different from those we took in previous projects, you will want to look to start from scratch. Trying to reuse old code and shoehorn it into what effectively amounts to an entirely different paradigm will create a lot of problems.
-But, the conceptual level, at least for the baseline implementation, we'll be pursuing will be the space-based equivalent of **primereg**, that is, no fancy optimizations for odd numbers, no map traversals, nor the **sqrt()** tricks, just straight up processing of the values that need to be processed.
+But, the conceptual level, at least for the baseline implementation, what we'll be pursuing will be the space-based equivalent of **primereg**, that is, no fancy optimizations for odd numbers, no map traversals, nor the **sqrt()** tricks, just straight up processing of the values that need to be processed.
 You may find some of the previous optimizations aren't even applicable based on how this algorithm works (break on composite, for instance).
 =====Sieve of Eratosthenes (primesoe)=====
 Your next program, and first sieve, will be the Sieve of Eratosthenes. Perhaps among the best  and likely longest-known sieves, its origins date from antiquity.
@@ Line 86: / Line 86: @@
 This is a space-constrained algorithm, therefore we will need a chunk of space to store these values. Think about what a lot of this looks like with respect to how you know how to organize data.
-=====Optimizations on sieve of Eratosthenes=====
+=====prime algorithm optimizations=====
-Not only do we have a new algorithm to explore, but we can also explore new ways of implementing some familiar prime number optimizations:
+To give us a decent appreciation of the subtleties of algorithm development in a theme of programs, I have identified the following optimizations that we will be implementing.
-====odds-only processing (primesoeo)====
-Taking the **primesoe** codebase, enhance it to do odds-only processing, ideally considering overall storage used.
-====sqrt() trick (primesoes)====
+For simplicity, I have encoded this information in the file name (and therefore resulting executable/binary) that will correspond to the indicated algorithm+optimizations.
-Taking the **primesoe** codebase, enhance it to utilize the **sqrt()** trick.
-NOTE: this variant is still to process all values (odd and even).
+To break it down, all prime programs will be of the form:
+  * primeALG[O...]
+    * where each and every program starts with "prime"
+    * is immediately followed by a 3-letter (lowercase) abbreviation of the algorithm to be implemented (**reg**, or **soe**, for instance)
+    * and then is followed by 0 or more layered attributes describing the particular optimization that is applied (again, if any: **zero** or more).
-====approximated square root trick (primesoea)====
+The optimizations we will be implementing in this project (and their naming values) include:
-Taking the **primesoe** codebase, enhance it to utilize the approximated square root trick.
+  * **odds-only checking (o)** - aside from **2**, **all** other prime numbers are odd. Therefore, there is zero need to perform a composite check on an even number, allowing us to focus exclusively on odd values (luckily, they seem to occur in a predictable pattern).
+  * **sqrt() trick (s)** - mathematically it has been shown that if a number has any evenly divisible factors, at least one half of that factor pair will occur by the square root point of the number being tested.
+  * **sqrt()-less square root approximation (a)** - **sqrt()**, a function in the math library, does an industrial strength square root calculation. We don't need that, merely a whole integer value corresponding to the approximate square root. Here we will implement our own logic to approximate square root, hopefully with a considerable performance impact.
-NOTE: this variant is still to process all values (odd and even).
+Unless specified in the encoded name, your algorithm should only implement the algorithm and optimization(s) specified.
-====odd + sqrt() trick (primesoeos)====
+That is, if your program to implement is **primesoeo**, that means you are ONLY to implement the sieve of Eratosthenes algorithm and odds-only checking. **NO** sqrt() trick, etc. We are establishing separate data points for analytical comparison.
-Modify **primesoe** to implement both the odd traversal and sqrt() trick optimizations.
-====odd + approximated square root trick (primesoeoa)====
+On the other hand, if your program to implement is **primesoeos**, that means your are implementing the "odds traversal", and "sqrt() trick" optimizations.
-Modify **primesoe** to implement both the odd traversal and approximated square root trick optimizations.
+Some of these optimizations can co-exist easily (odd + sqrt(), odd + approx square root),  while others are mutually exclusive (sqrt() and approximated square root conflict). So there are definitely a few combinations that are possible using this scheme.
-=====Program=====
+For this project, you'll be implementing various combinations of optimizations, giving us a full set of programs to analyze performance.
+=====A note on comments=====
+Something I noticed (and have historically noticed) in relation to comments that I'd like to point out:
+Comments should be describing what is going on in your code.
+With projects like this, often relying on a common base, comments become even more important, as they allow me to see specifically what is changed or unique about one variant over the other.
+As such, when evaluating the project, I will be looking for pertinent comments specifically covering the how or why of the particular change unique to the variant in question.
+And notice I said the "how" and/or "why". NOT the "what". I see all the time vague comments like "<nowiki>// doing the sqrt() optimization</nowiki>"... but:
+  * WHY is that important to the process?
+  * HOW does it impact the efficiency of the algorithm?
+These are things I'd like to see addressed in your comments, as there were some cases where the WHAT was claimed, yet what actually followed had little resemblance (if any) on the requirements for that variant.
+Just like if you can't do it by hand you have no business trying to code it- if you cannot adequately explain the WHY and HOW, you similarly will have trouble.
+=====Programs=====
 It is your task to write the following Sieve of Eratosthenes-oriented prime number calculating programs:
@@ Line 117: / Line 139: @@
   * **primesoeos.c**: odd + sqrt() trick
   * **primesoeoa.c**: odd + approximated square root trick
 ====Program Specifications====
@@ Line 153: / Line 174: @@
     * as primes are being displayed, they are space-separated (first prime hugs the left margin), and when all said and done, a newline is issued.
     * the timing information will be displayed in accordance to code I will provide below (see the **timing** section).
+=====Implementation Restrictions=====
+As our goal is not only to explore the more subtle concepts of computing but to promote different methods of thinking (and arriving at solutions seemingly in different ways), one of the themes I have been harping on is the stricter adherence to the structured programming philosophy/paradigm. It isn't just good enough to be able to crank out a solution if you remain blind to the many nuances of the tools we are using, so we will at times be going out of our way to emphasize focus on certain areas that may see less exposure (or avoidance due to it being less familiar-- and from what I've witnessed so far this semester, avoidance is alive and well).
+As such, the following implementation restrictions are also in place:
+  * focus on **if()**, **ternary**, and **conditional chaining** over switch/case statements.
+  * keep your use of **continue** and **break** statements (especially **break** statements) to a necessary minimum).
+  * absolutely **NO** other non-structured program flow alteration (jumps, gotos, etc.)
+  * absolutely **NO** infinite loops (**while(1)**, which are more-or-less unviable anyway if you cannot **break** out of them).
+  * no forced redirection of the flow of the process (no seeking to the end of the file to grab a max size only to zip back somewhere else: deal with the data in as you are naturally encountering it).
+  * All "arrays" must be declared and referenced using ONLY pointer notation, NO square brackets.
+  * **NO** logic shunts (ie having an if statement nested inside a loop to bypass an undesirable iteration)- this should be handled by the loop condition!
+  * At most, **ONE** return statement per function (in the case of **void**, 0 return statements).
+  * No redundant duplication of code to address different top-level conditions or operational constraints (think quantity vs. range- these can successfully co-exist in the same block of code).
+  * Never leave an initialized or allocated resource unverified- always do proper error checking (was the file successfully opened? Was the memory successfully allocated?
+A common resistance or complaint I get with imposing these is that it may make your solutions more cumbersome or less optimal; that actually may not be an incorrect assertion, but remember: we are interested in the longer-term pursuit of structured thinking and effective problem solving. To foster your ability to think flexibly and differently. We tend to be naturally more averse to going against the grain, but to be an effective programmer/problem solver, this is absolutely necessary. It may be "annoying", and you may choose to make it more aggravating on both yourself and me by agonizing over it, but history and experience teaching has shown me, time and time again, that this is an investment and it pays off in the long run (assuming one actually plays along).
+I am seeking to make you all better thinkers and programmers, and I cannot do that if I cave to your innate desires not to change your ways of doing things. Yes, it may be unfamiliar; yes, it may be perceived as challenging. But you know what? Work through it and eventually it becomes the new normal- what was challenging is now no longer an issue.
 =====Grabit Integration=====
@@ Line 158: / Line 200: @@
 To "grab" it:
-NOTE: You will **NEED** to specify the semester as indicated as the semester in question has not yet started.
 <cli>
-lab46:~/src/discrete$ SEMESTER=fall2017 grabit discrete pnc2
+lab46:~/src/discrete$ grabit discrete pnc2
-make: Entering directory '/var/public/fall2017/discrete/pnc2'
+make: Entering directory '/var/public/SEMESTER/discrete/pnc2'
-Commencing copy process for fall2017 discrete project pnc2:
+Commencing copy process for SEMESTER discrete project pnc2:
  -> Creating project pnc2 directory tree           ... OK
  -> Copying pnc2 project files                     ... OK
@@ Line 172: / Line 212: @@
 *** Copy COMPLETE! You may now go to the '/home/USER/src/discrete/pnc2' directory ***
-make: Leaving directory '/var/public/fall2017/discrete/pnc2'
+make: Leaving directory '/var/public/SEMESTER/discrete/pnc2'
-lab46:~/src/discrete/$
 lab46:~/src/discrete$ cd pnc2
 lab46:~/src/discrete/pnc2$
@@ Line 190: / Line 229: @@
 <cli>
-lab46:~/src/discrete/pnc1$ make help
+lab46:~/src/discrete/pnc2$ make help
-******************[ Discrete Structures pnc1 Project ]******************
+******************[ Discrete Structures pnc2 Project ]******************
 ** make                     - build everything                        **
 ** make showerrors          - display compiler warnings/errors        **
-**                                                                    **
 ** make debug               - build everything with debug symbols     **
-** make check               - check implementation for validity       **
+** make checkqty            - runtime evaluation for qty              **
+** make checkrange          - runtime evaluation for range            **
+**                                                                    **
+** make verifyqty           - check implementation for qty validity   **
+** make verifyrange         - check implementation for range validity **
+** make verifyall           - verify project specifications           **
 **                                                                    **
 ** make link                - link in previous prime programs         **
@@ Line 211: / Line 254: @@
 ** make help                - this information                        **
 ************************************************************************
-lab46:~/src/discrete/pnc1$
+lab46:~/src/discrete/pnc2$
 </cli>
@@ Line 217: / Line 260: @@
   * **make**: compile everything
-    * any **warnings** or **errors** generated by the compiler will go into a file in the base directory of the project in a file called **errors**; you can **cat** it to view the information.
+    * any **warnings** or **errors** generated by the compiler will go into a file in the base directory of pnc0 in a file called **errors**; you can **cat** it to view the information.
   * **make debug**: compile everything with debug support
     * any **warnings** or **errors** generated by the compiler will be displayed to the screen as the programs compile.
@@ Line 223: / Line 266: @@
   * **make save**: make a backup of your current work
   * **make submit**: archive and submit your project
+The various "check" options do a runtime performance grid, allowing you to compare timings between your implementations.
+The various "verify" options do more aggressive checks, helping to ensure your project falls within stated project specifications.
 Just another "nice thing" we deserve.
 =====Command-Line Arguments=====
-To automate our comparisons, we will be making use of command-line arguments in our programs. As we have yet to really get into arrays, I will provide you same code that you can use that will allow you to utilize them for the purposes of this project.
+To automate our comparisons, we will be making use of command-line arguments in our programs.
 ====header files====
@@ Line 245: / Line 292: @@
 int main(int argc, char **argv)
 </code>
+There are two very important variables involved here (the types are actually what are important, the names given to the variables are actually quite, variable; you may see other references refer to them as things like "ac" and "av"):
+  * int argc: the count (an integer) of tokens given on the command line (program name + arguments)
+  * <nowiki>char **argv</nowiki>: an array of strings (technically an array of an array of char) that contains "strings" of the various tokens provided on the command-line.
 The arguments are accessible via the argv array, in the order they were specified:
-  * **argv[0]**: program invocation (path + program name)
+  * argv[0]: program invocation (path + program name)
-  * **argv[1]**: our maximum / upper bound
+  * argv[1]: our maximum / upper bound
+  * argv[2]: reserved value, should still be provided and be a 1 for this project
+  * argv[3]: conditionally optional; represents lower bound
+  * argv[4]: conditionally optional; represents upper bound
+Additionally, let's not forget the **argc** variable, an integer, which contains a count of arguments (argc == argument count). If we provided argv[0] through argv[4], argc would contain a 5.
+===example===
+For example, if we were to execute the **primeregbms** program:
+<cli>
+lab46:~/src/discrete/pnc1$ ./primeregbms 128 1 2 2048
+</cli>
+We'd have:
+  * <nowiki>argv[0]</nowiki>: "./primeregbms"
+  * <nowiki>argv[1]</nowiki>: "128" (note, NOT the scalar integer 128, but a string)
+  * <nowiki>argv[2]</nowiki>: "1"
+  * <nowiki>argv[3]</nowiki>: "2"
+  * <nowiki>argv[4]</nowiki>: "2048"
+and let's not forget:
+  * argc: 5   (there are 5 things, argv indexes 0, 1, 2, 3, and 4)
+With the conditionally optional arguments as part of the program spec, for a valid execution of the program, argc could be a value anywhere from 3 to 5.
 ====Simple argument checks====
-Although I'm not going to require extensive argument parsing or checking for this project, we should check to see if the minimal number of arguments has been provided:
+While there are a number of checks we should perform, one of the first should be a check to see if the minimal number of arguments has been provided:
 <code c>
-    if (argc < 2)  // if less than 2 arguments have been provided
+    if (argc < 3)  // if less than 3 arguments (program_name + quantity + argv[2] == 3) have been provided
     {
-        fprintf(stderr, "Not enough arguments!\n");
+        fprintf(stderr, "%s: insufficient number of arguments!\n", argv[0]);
         exit(1);
     }
 </code>
+Since argv[3] (lower bound) and argv[4] (upper bound) are conditionally optional, it wouldn't make sense to check for them in the overall count. But we can and do still want to stategically utilize **argc** to determine if an argv[3] or argv[4] is present.
 ====Grab and convert max====
-Finally, we need to put the argument representing the maximum value into a variable.
+Finally, we need to put the argument representing the maximum quantity into a variable.
 I'd recommend declaring a variable of type **int**.
@@ Line 270: / Line 350: @@
 <code c>
-    max  = atoi(argv[1]);
+    max  = atoi (argv[1]);
 </code>
@@ Line 528: / Line 608: @@
 And of course the same for the 3 variants, and the same error message reporting if invalid values are given.
 ====Performance changes====
 You may notice a change with the sieves as compared to the other algorithms you've implemented with respect to performance- there will likely be a lower bound of performance, ie you have to exceed a certain threshold before the algorithm truly enters its power band, and then it may truly be a step-up in terms of performance.
 =====Check Results=====
-To verify your program's correctness and view your implementation's performance, "**make check**" will recognize the indicated program names listed above:
+If you'd like to compare your implementations, I rigged up a Makefile checking rule called "**make checkqty**" and "**make checkrange**" which you can run to get a nice side-by-side runtime comparisons of your implementations.
+In order to work, you **MUST** be in the directory where your pnc2 binaries reside, and must be named as such (which occurs if you ran **make** to compile them).
+====check qty====
 <cli>
+lab46:~/src/discrete/pnc2$ make checkqty
 =========================================================
       qty   soe     soeo    soes    soea    soeos   soeoa
@@ Line 553: / Line 638: @@
  verify:     OK      OK      OK      OK      OK      OK
 =========================================================
+lab46:~/src/discrete/pnc2$
 </cli>
-If/when the runtime of a particular prime variant exceeds an upper threshold (1 second), it will be omitted from further tests, and a series of dashes will instead appear in the output.
+When the runtime of a particular prime variant exceeds an upper threshold (1 second), it will be omitted from further tests, and a series of dashes will instead appear in the output.
 If you don't feel like waiting, simply hit **CTRL-c** and the script will terminate.
-If the check is successful, you will see "OK" displayed beneath in the appropriate column; if unsuccessful, you will see "MISMATCH".
+====check range====
+<cli>
+lab46:~/src/discrete/pnc2$ make checkrange
+coming soon
+lab46:~/src/discrete/pnc2$
+</cli>
+====Verification====
+I also include a validation check- to ensure your prime programs are actually producing the correct list of prime numbers. If the check is successful, you will see "OK" displayed beneath in the appropriate column; if unsuccessful, you will see "MISMATCH".
+====Full Verification Compliance====
+There's also a more rigorous verification step you can take, which runs your programs through a series to tests to see if they conform to project specifications:
+<cli>
+lab46:~/src/discrete/pnc2$ make verifyall
+coming soon
+lab46:~/src/discrete/pnc2$
+</cli>
+===verifyall tests===
+The "**verifyall**" is an industrial grade verification; there are 13 specific tests performed, they are:
+  * **qtynorm**: a normal quantity run (2-max)
+    * **./primealg 2048 1 2 0**
+  * **qtypart**: an offset quantity run (24-max)
+    * **./primealg 2048 1 24 0**
+  * **rngnorm**: a normal range run (2-max)
+    * **./primealg 0 1 2 2048**
+  * **rngpart**: an offset range run (24-max)
+    * **./primealg 0 1 24 2048**
+  * **coop**: both qty and upper bounds set (q: 2048, ub: 8192)
+    * **./primealg 2048 1 2 8192**
+  * **coop2**: both qty and upper bounds set (q: 512, ub: 8192)
+    * **./primealg 512 1 2 8192**
+  * **coop3**: both qty and upper bounds set, offset start (24-max, q: 2048, ub: 8192)
+    * **./primealg 2048 1 24 8192**
+  * **noargs**:  no arguments provided on command line (invokes error message)
+    * **./primealg**
+  * **invargs**: insufficient number of arguments provided (invokes error)
+    * **./primealg 128**
+  * **invqty**: invalid value for quantity argument given (invokes error)
+    * **./primealg -2 1**
+  * **invnary**: invalid value given for n-ary (invokes error)
+    * **./primealg 128 2**
+  * **invlow**: invalid value given for lower bound (invokes error)
+    * **./primealg 128 1 1**
+  * **invhigh**: invalid value given for upper bound (invokes error)
+    * **./primealg 128 1 32 24**
+If you'd actually to see the output your program's output is being tested against, that can be found in the **/usr/local/etc** directory in the file **primeTEST**, where "TEST" is the name of the verify test mentioned above.
+For example, if you wanted to see the intended output of the **invnary** test, that would be found in:
+  * **/usr/local/etc/primeinvnary**
+You could easily run your program with the stated arguments for the test, then use **cat** to display the test results and do a visual comparison.
+====In general====
+Analyze the times you see... do they make sense, especially when comparing the algorithm used and the quantity being processed? These are related to some very important core Computer Science considerations we need to be increasingly mindful of as we design our programs and implement our solutions. Algorithmic complexity and algorithmic efficiency will be common themes in all we do.
 =====Submission=====
@@ Line 568: / Line 711: @@
   * Code must be nicely and consistently indented (you may use the **indent** tool)
   * Code must utilize the algorithm(s) presented above:
-    * **primeregmo.c**: map + odd traversal optimizations
+  * soe     soeo    soes    soea    soeos   soeoa
-    * **primeregms.c**: map traversal + sqrt() trick
+    * **primesoe.c**: baseline sieve of eratosthenes
-    * **primeregma.c**: map treversal + approximated square root trick
+    * **primesoeo.c**: odd traversal optimization
-    * **primeregos.c**: odd traversal + sqrt() trick
+    * **primesoes.c**: sqrt() trick
-    * **primeregoa.c**: odd traversal + approximated square root trick
+    * **primesoea.c**: approximated square root
-    * **primeregbmo.c**: break + map + odd traversal
+    * **primesoeos.c**: odd traversal + sqrt() trick
-    * **primeregbms.c**: break + map + sqrt() trick
+    * **primesoeoa.c**: odd traversal + approximated square root
-    * **primeregbma.c**: break + map + approximated square root trick
+  * Code must be commented, and comments must focus on the how and why of the process.
-    * **primeregbos.c**: break + odd + sqrt() trick
-    * **primeregboa.c**: break + odd + approximated square root trick
-    * **primeregmos.c**: map + odd traversal + sqrt() trick
-    * **primeregmoa.c**: map + odd traversal + approximated square root trick
-    * **primeregbmos.c**: break + map + odd + sqrt() trick
-    * **primeregbmoa.c**: break + map + odd + approximated square root trick
-  * Code must be commented
-    * have a properly filled-out comment banner at the top
-      * be sure to include any compiling instructions
-    * have at least 20% of your program consist of **<nowiki>//</nowiki>**-style descriptive comments
   * Output Formatting (including spacing) of program must conform to the provided output (see above).
   * Track/version the source code in a repository
@@ Line 593: / Line 726: @@
 <cli>
-lab46:~/src/discrete/pnc1$ make submit
+lab46:~/src/discrete/pnc2$ make submit
 Delinking ...
-removed ‘primerega.c’
-removed ‘primeregba.c’
-removed ‘primeregb.c’
-removed ‘primeregbm.c’
-removed ‘primeregbo.c’
-removed ‘primeregbs.c’
-removed ‘primereg.c’
-removed ‘primeregm.c’
-removed ‘primerego.c’
-removed ‘primeregs.c’
-removed ‘primeregbma’
-removed ‘primeregbmoa’
-removed ‘primeregbmo’
-removed ‘primeregbmos’
-removed ‘primeregbms’
-removed ‘primeregboa’
-removed ‘primeregbos’
-removed ‘primeregma’
-removed ‘primeregmoa’
-removed ‘primeregmo’
-removed ‘primeregmos’
-removed ‘primeregms’
-removed ‘primeregoa’
-removed ‘primeregos’
 removed ‘errors’
 Project backup process commencing
-Taking snapshot of current project (pnc1)      ... OK
+Taking snapshot of current project (pnc2)      ... OK
-Compressing snapshot of pnc1 project archive   ... OK
+Compressing snapshot of pnc2 project archive   ... OK
-Setting secure permissions on pnc1 archive     ... OK
+Setting secure permissions on pnc2 archive     ... OK
 Project backup process complete
-Submitting discrete project "pnc1":
+Submitting discrete project "pnc2":
-    -> ../pnc1-20170713-11.tar.gz(OK)
+    -> ../pnc2-DATESTRING-HR.tar.gz(OK)
 SUCCESSFULLY SUBMITTED
@@ Line 637: / Line 746: @@
 You should get that final "SUCCESSFULLY SUBMITTED" with no error messages occurring. If not, check for typos and or locational mismatches.
-What I will be looking for:
+====Evaluation Criteria====
+Grand total points:
 <code>
-:pnc1:final tally of results (364/364)
+:pnc2:final tally of results (390/390)
-*:pnc1:primeregbma.c performs proper argument checking [2/2]
-*:pnc1:primeregbma.c no negative compiler messages [2/2]
-*:pnc1:primeregbma.c implements only specified algorithm [6/6]
-*:pnc1:primeregbma.c consistent indentation and comments [4/4]
-*:pnc1:primeregbma.c data and output conform to specifications [6/6]
-*:pnc1:primeregbma.c make check runtime tests succeed [6/6]
-*:pnc1:primeregbmoa.c performs proper argument checking [2/2]
-*:pnc1:primeregbmoa.c no negative compiler messages [2/2]
-*:pnc1:primeregbmoa.c implements only specified algorithm [6/6]
-*:pnc1:primeregbmoa.c consistent indentation and comments [4/4]
-*:pnc1:primeregbmoa.c data and output conform to specifications [6/6]
-*:pnc1:primeregbmoa.c make check runtime tests succeed [6/6]
-*:pnc1:primeregbmo.c performs proper argument checking [2/2]
-*:pnc1:primeregbmo.c no negative compiler messages [2/2]
-*:pnc1:primeregbmo.c implements only specified algorithm [6/6]
-*:pnc1:primeregbmo.c consistent indentation and comments [4/4]
-*:pnc1:primeregbmo.c data and output conform to specifications [6/6]
-*:pnc1:primeregbmo.c make check runtime tests succeed [6/6]
-*:pnc1:primeregbmos.c performs proper argument checking [2/2]
-*:pnc1:primeregbmos.c no negative compiler messages [2/2]
-*:pnc1:primeregbmos.c implements only specified algorithm [6/6]
-*:pnc1:primeregbmos.c consistent indentation and comments [4/4]
-*:pnc1:primeregbmos.c data and output conform to specifications [6/6]
-*:pnc1:primeregbmos.c make check runtime tests succeed [6/6]
-*:pnc1:primeregbms.c performs proper argument checking [2/2]
-*:pnc1:primeregbms.c no negative compiler messages [2/2]
-*:pnc1:primeregbms.c implements only specified algorithm [6/6]
-*:pnc1:primeregbms.c consistent indentation and comments [4/4]
-*:pnc1:primeregbms.c data and output conform to specifications [6/6]
-*:pnc1:primeregbms.c make check runtime tests succeed [6/6]
-*:pnc1:primeregboa.c performs proper argument checking [2/2]
-*:pnc1:primeregboa.c no negative compiler messages [2/2]
-*:pnc1:primeregboa.c implements only specified algorithm [6/6]
-*:pnc1:primeregboa.c consistent indentation and comments [4/4]
-*:pnc1:primeregboa.c data and output conform to specifications [6/6]
-*:pnc1:primeregboa.c make check runtime tests succeed [6/6]
-*:pnc1:primeregbos.c performs proper argument checking [2/2]
-*:pnc1:primeregbos.c no negative compiler messages [2/2]
-*:pnc1:primeregbos.c implements only specified algorithm [6/6]
-*:pnc1:primeregbos.c consistent indentation and comments [4/4]
-*:pnc1:primeregbos.c data and output conform to specifications [6/6]
-*:pnc1:primeregbos.c make check runtime tests succeed [6/6]
-*:pnc1:primeregma.c performs proper argument checking [2/2]
-*:pnc1:primeregma.c no negative compiler messages [2/2]
-*:pnc1:primeregma.c implements only specified algorithm [6/6]
-*:pnc1:primeregma.c consistent indentation and comments [4/4]
-*:pnc1:primeregma.c data and output conform to specifications [6/6]
-*:pnc1:primeregma.c make check runtime tests succeed [6/6]
-*:pnc1:primeregmoa.c performs proper argument checking [2/2]
-*:pnc1:primeregmoa.c no negative compiler messages [2/2]
-*:pnc1:primeregmoa.c implements only specified algorithm [6/6]
-*:pnc1:primeregmoa.c consistent indentation and comments [4/4]
-*:pnc1:primeregmoa.c data and output conform to specifications [6/6]
-*:pnc1:primeregmoa.c make check runtime tests succeed [6/6]
-*:pnc1:primeregmo.c performs proper argument checking [2/2]
-*:pnc1:primeregmo.c no negative compiler messages [2/2]
-*:pnc1:primeregmo.c implements only specified algorithm [6/6]
-*:pnc1:primeregmo.c consistent indentation and comments [4/4]
-*:pnc1:primeregmo.c data and output conform to specifications [6/6]
-*:pnc1:primeregmo.c make check runtime tests succeed [6/6]
-*:pnc1:primeregmos.c performs proper argument checking [2/2]
-*:pnc1:primeregmos.c no negative compiler messages [2/2]
-*:pnc1:primeregmos.c implements only specified algorithm [6/6]
-*:pnc1:primeregmos.c consistent indentation and comments [4/4]
-*:pnc1:primeregmos.c data and output conform to specifications [6/6]
-*:pnc1:primeregmos.c make check runtime tests succeed [6/6]
-*:pnc1:primeregms.c performs proper argument checking [2/2]
-*:pnc1:primeregms.c no negative compiler messages [2/2]
-*:pnc1:primeregms.c implements only specified algorithm [6/6]
-*:pnc1:primeregms.c consistent indentation and comments [4/4]
-*:pnc1:primeregms.c data and output conform to specifications [6/6]
-*:pnc1:primeregms.c make check runtime tests succeed [6/6]
-*:pnc1:primeregoa.c performs proper argument checking [2/2]
-*:pnc1:primeregoa.c no negative compiler messages [2/2]
-*:pnc1:primeregoa.c implements only specified algorithm [6/6]
-*:pnc1:primeregoa.c consistent indentation and comments [4/4]
-*:pnc1:primeregoa.c data and output conform to specifications [6/6]
-*:pnc1:primeregoa.c make check runtime tests succeed [6/6]
-*:pnc1:primeregos.c performs proper argument checking [2/2]
-*:pnc1:primeregos.c no negative compiler messages [2/2]
-*:pnc1:primeregos.c implements only specified algorithm [6/6]
-*:pnc1:primeregos.c consistent indentation and comments [4/4]
-*:pnc1:primeregos.c data and output conform to specifications [6/6]
-*:pnc1:primeregos.c make check runtime tests succeed [6/6]
 </code>
-======Project: ALGORITHMS - PRIME NUMBER CALCULATION (pnc1)======
+What I will be looking for (for each file):
-=====Background=====
-In mathematics, a **prime** number is a value that is only evenly divisible by 1 and itself; it has just that one pair of factors, no others. Numbers that have divisibility/factors are classified as **composite** numbers.
-The number **6** is a **composite** number, as in addition to 1 and 6, it also has the factors of 2 and 3.
-The number **17**, however, is a **prime** number, as no numbers other than 1 and 17 can be evenly divided into it.
-=====Calculating the primality of a number=====
-As of yet, there is no quick and direct way of determining the primality of a given number. Instead, we must perform a series of tests to determine if it fails primality (typically by proving it is composite).
-This process incurs a considerable amount of processing overhead on the task, so much so that increasingly large values take ever-expanding amounts of time. Often, approaches to prime number calculation involve various algorithms, which offer various benefits (less time) and drawback (more complex code).
-Your task for this project is to implement a prime number program using the straightforward, unoptimized brute-force algorithm, which determines the primality of a number in a "trial by division" approach.
-=====Review of main algorithm: brute force (primereg)=====
-The brute force approach is the simplest to implement, although it does come at some cost, which hopefully you have realized by now having completed **pnc0**.
-To review the process of computing the primality of a number, we simply attempt to evenly divide all the potential factors between 2 and one less than the number into the number in question. If any one of them divides evenly, the number is **NOT** prime, but instead classified as composite (meaning made up of various parts, in this case those "parts" are factor pairs).
-Checking the **remainder** of an integer division indicates whether or not a division was clean (having 0 remainder indicates such a state).
-For example, with the number 11:
 <code>
-% 2 = 1 (2 is not a factor of 11)
+*:pnc2:primeALGO.c compiles cleanly, no compiler messages [4/4]
-% 3 = 2 (3 is not a factor of 11)
+*:pnc2:primeALGO.c implements only specified algorithm [8/8]
-% 4 = 3 (4 is not a factor of 11)
+*:pnc2:primeALGO.c consistent indentation throughout code [4/4]
-% 5 = 1 (5 is not a factor of 11)
+*:pnc2:primeALGO.c relevant how and why comments in code [7/7]
-% 6 = 5 (6 is not a factor of 11)
+*:pnc2:primeALGO.c code conforms to project specifications [4/4]
-% 7 = 4 (7 is not a factor of 11)
+*:pnc2:primeALGO.c implementation free from restrictions [13/13]
-% 8 = 3 (8 is not a factor of 11)
+*:pnc2:primeALGO runtime output conforms to specifications [4/4]
-% 9 = 2 (9 is not a factor of 11)
+*:pnc2:primeALGO make checkqty test times within reason [4/4]
-% 10 = 1 (10 is not a factor of 11)
+*:pnc2:primeALGO make checkrange test times within reason [4/4]
+*:pnc2:primeALGO make verifyall tests succeed [13/13]
 </code>
-Because none of the values 2-10 evenly divided into 11, we can say it passed the test: **11 is a prime number**
-On the other hand, take 119:
-<code>
-% 2 = 1 (2 is not a factor of 119)
-% 3 = 2 (3 is not a factor of 119)
-% 4 = 3 (4 is not a factor of 119)
-% 5 = 4 (5 is not a factor of 119)
-% 6 = 5 (6 is not a factor of 119)
-% 7 = 0 (7 is a factor of 119)
-% 8 = 7
-% 9 = 2
-% 10 = 9
-% 11 = 9
-% 12 = 11
-% 13 = 2
-...
-</code>
-Because, during our range of testing every value from 2-118, we find that 7 evenly divides into 119, it failed the test: 119 is **not** prime, but is instead a composite number.
-====algorithm====
-Some things to keep in mind on your implementation:
-  * you will want to use loops (no less than 2, no more than 2) for this program.
-    * a nested loop makes the most sense:
-      * an outer loop that drives the progression of each sequential number to be tested
-      * an inner loop that tests that current number to see if it has any factors
-  * you know the starting value and the terminating condition, so you have a clear starting and ending point to work with.
-  * I want you to use two **DIFFERENT** kind of loops in your programs. If you use a **for()** loop in your outer loop, I want you to use a **while()** or **do-while()** loop in your inner loop (and whatever combination you end up with).
-  * I do **NOT** want to see ambiguous, one-letter variables used in your implementation(s). Please use //meaningful// variable names.
-    * Some good examples of variable names would be:
-      * **number**: the number being tested
-      * **factor**: the value being divided into number to test for primality
-      * **step**: the rate by which some variable is changing
-      * **qty**: the count of the current tally of primes
-      * **max**: the maximum count we seek
-      * **start**: a value we are starting at
-      * **lower**: a lower bound
-      * **upper**: an upper bound
-      * see how much more readable and meaningful these are, especially as compared to **a**, **i**, **n**, **x**? You may even find it helps with debugging and understanding your code better.
-  * let the loops drive the overall process. Identify prime/composite status separate from loop terminating conditions.
-  * be mindful of the particular combination of optimizations you are implementing.
-  * your timing should start before the loop (just AFTER argument processing), and terminate immediately following the terminating newline outside the loops.
-=====prime algorithm optimizations=====
-To give us a decent appreciation of the subtleties of algorithm development in a theme of programs, I have identified the following optimizations that we will be implementing.
-For simplicity, I have encoded this information in the file name (and therefore resulting executable/binary) that will correspond to the indicated algorithm+optimizations.
-To break it down, all prime programs will be of the form:
-  * primeALG[O...]
-    * where each and every program starts with "prime"
-    * is immediately followed by a 3-letter (lowercase) abbreviation of the algorithm to be implemented (**reg**, for instance)
-    * and then is followed by 0 or more layered attributes describing the particular optimization that is applied (again, if any: **zero** or more).
-The optimizations we will be implementing in this project (and their naming values) include:
-  * **break on composite (b)** - once a tested number is proven composite, there is no need to continue processing: break out of the factor loop and proceed to the next number
-  * **mapping factors of 6 (m)** - it turns out that, aside from the initial primes of **2** and **3**, that **all** prime numbers fall to a +1 or -1 off a factor of six (there is an algorithm for this: **6a+/-1**). This optimization will utilize this property, only testing numbers +/-1 off of factors of 6 (how might this impact overall processing?)
-  * **odds-only checking (o)** - aside from **2**, **all** other prime numbers are odd. Therefore, there is zero need to perform a composite check on an even number, allowing us to focus exclusively on odd values (luckily, they seem to occur in a predictable pattern).
-  * **sqrt() trick (s)** - mathematically it has been shown that if a number has any evenly divisible factors, at least one half of that factor pair will occur by the square root point of the number being tested.
-  * **sqrt()-less square root approximation (a)** - **sqrt()**, a function in the math library, does an industrial strength square root calculation. We don't need that, merely a whole integer value corresponding to the approximate square root. Here we will implement our own logic to approximate square root, hopefully with a considerable performance impact.
-Unless specified in the encoded name, your algorithm should only implement the algorithm and optimization(s) specified.
-That is, if your program to implement is **primerego**, that means you are ONLY to implement the brute force algorithm and odds-only checking. **NO** break on composite, **NO** sqrt() trick, etc. We are establishing separate data points for analytical comparison.
-On the other hand, if your program to implement is **primeregbms**, that means your are implementing the "break on composite", "map traversal", and "sqrt() trick" optimizations.
-Some of these optimizations can co-exist easily (break + map, odd + sqrt()), others are partially compatible (map + odd can coexist in a certain form), while others are mutually exclusive (sqrt() and approximated square root conflict). So there are definitely a few combinations that are possible using this scheme.
-For this project, you'll be implementing the remaining (14) combinations of optimizations, giving us a full set of programs to analyze performance.
-=====Programs=====
-Those variants to implement are:
-  * the remainder of the viable double optimization combinations:
-    * **primeregmo.c**: map + odd traversal optimizations
-    * **primeregms.c**: map traversal + sqrt() trick
-    * **primeregma.c**: map treversal + approximated square root trick
-    * **primeregos.c**: odd traversal + sqrt() trick
-    * **primeregoa.c**: odd traversal + approximated square root trick
-  * all of the viable triple optimization combinations:
-    * **primeregbmo.c**: break + map + odd traversal
-    * **primeregbms.c**: break + map + sqrt() trick
-    * **primeregbma.c**: break + map + approximated square root trick
-    * **primeregbos.c**: break + odd + sqrt() trick
-    * **primeregboa.c**: break + odd + approximated square root trick
-    * **primeregmos.c**: map + odd traversal + sqrt() trick
-    * **primeregmoa.c**: map + odd traversal + approximated square root trick
-  * all of the viable quadruple optimizations combinations:
-    * **primeregbmos.c**: break + map + odd + sqrt() trick
-    * **primeregbmoa.c**: break + map + odd + approximated square root trick
-You should, if you haven't already, started to notice commonalities and patterns between the implementations. This should help you realize that although there are a number of distinct programs being produced, the actual development effort is realized to be far less.
-=====Command-Line Arguments=====
-To automate our comparisons, we will be making use of command-line arguments in our programs. As we have yet to really get into arrays, I will provide you some code that you can use that will allow you to utilize them for the purposes of this project.
-====header files====
-We don't need any extra header files to use command-line arguments, but we will need an additional header file to use the **atoi(3)** function, which we'll use to quickly turn the command-line parameter into an integer, and that header file is **stdlib.h**, so be sure to include it with the others:
-<code c>
-#include <stdio.h>
-#include <stdlib.h>
-</code>
-====setting up main()====
-To accept (or rather, to gain access) to arguments given to your program at runtime, we need to specify two parameters to the main() function. While the names don't matter, the types do.. I like the traditional **argc** and **argv** names, although it is also common to see them abbreviated as **ac** and **av**.
-Please declare your main() function as follows:
-<code c>
-int main(int argc, char **argv)
-</code>
-The arguments are accessible via the argv array, in the order they were specified:
-  * argv[0]: program invocation (path + program name)
-  * argv[1]: our maximum / upper bound
-  * argv[2]: reserved value, should still be provided and be a 1 for this project
-  * argv[3]: conditionally optional; represents lower bound
-  * argv[4]: conditionally optional; represents upper bound
-====Simple argument checks====
-While there are a number of checks we should perform, one of the first should be a check to see if the minimal number of arguments has been provided:
-<code c>
-    if (argc < 3)  // if less than 3 arguments (program_name + quantity + argv[2] == 3) have been provided
-    {
-        fprintf(stderr, "%s: insufficient number of arguments!\n", argv[0]);
-        exit(1);
-    }
-</code>
-Since argv[3] (lower bound) and argv[4] (upper bound) are conditionally optional, it wouldn't make sense to check for them in the overall count. But we can and do still want to stategically utilize **argc** to determine if an argv[3] or argv[4] is present.
-====Grab and convert max====
-Finally, we need to put the argument representing the maximum quantity into a variable.
-I'd recommend declaring a variable of type **int**.
-We will use the **atoi(3)** function to quickly convert the command-line arguments into **int** values:
-<code c>
-    max  = atoi (argv[1]);
-</code>
-And now we can proceed with the rest of our prime implementation.
-=====Timing=====
-Often times, when checking the efficiency of a solution, a good measurement (especially for comparison), is to time how long the processing takes.
-In order to do that in our prime number programs, we are going to use C library functions that obtain the current time, and use it as a stopwatch: we'll grab the time just before starting processing, and then once more when done. The total time will then be the difference between the two (end_time - start_time).
-We are going to use the **gettimeofday(2)** function to aid us in this, and to use it, we'll need to do the following:
-====header file====
-In order to use the **gettimeofday(2)** function in our program, we'll need to include the **sys/time.h** header file, so be sure to add it in with the existing ones:
-<code c>
-#include <stdio.h>
-#include <stdlib.h>
-#include <sys/time.h>
-</code>
-====timeval variables====
-**gettimeofday(2)** uses a **struct timeval** data type, of which we'll need to declare two variables in our programs (one for storing the starting time, and the other for the ending time).
-Please declare these with your other variables, up at the top of main() (but still WITHIN main()-- you do not need to declare global variables).
-<code c>
-    struct timeval time_start; // starting time
-    struct timeval time_end;   // ending time
-</code>
-====Obtaining the time====
-To use **gettimeofday(2)**, we merely place it at the point in our code we wish to take the time.
-For our prime number programs, you'll want to grab the start time **AFTER** you've declared variables and processed arguments, but **JUST BEFORE** starting the driving loop doing the processing.
-That call will look something like this:
-<code c>
-    gettimeofday(&time_start, 0);
-</code>
-The ending time should be taken immediately after all processing (and prime number output) is completed, and right before we display the timing information to STDERR:
-<code c>
-    gettimeofday(&time_end, 0);
-</code>
-=====Execution=====
-====specified quantity====
-Your program output should be as follows (given the specified quantity):
-<cli>
-lab46:~/src/discrete/pnc1$ ./primeregmo 24 1
-3 5 7 11 13 17 19 23 29 31 37 41 43 47 53 59 61 67 71 73 79 83 89
-.0001
-lab46:~/src/discrete/pnc1$
-</cli>
-The execution of the programs is short and simple- grab the parameters, do the processing, produce the output, and then terminate.
-====invalid lower bound====
-Here's an example that should generate an error upon running (based on project specifications):
-<cli>
-lab46:~/src/discrete/pnc1$ ./primeregos 32 1 0
-./primeregos: invalid lower bound
-lab46:~/src/discrete/pnc1$
-</cli>
-In this case, the program logic should have detected an invalid condition and bailed out before prime computations even began. No timing data is displayed, because exiting should occur even prior to that.
-====upper bound overriding quantity====
-As indicated above, there is potential interplay with an active quantity and upper bound values. Here is an example where upper bound overrides quantity, resulting in an early termination (ie upper bound is hit before quantity):
-<cli>
-lab46:~/src/cprog/pnc1$ ./primeregbms 128 1 7 23
-11 13 17 19 23
-.0001
-lab46:~/src/cprog/pnc1$
-</cli>
-Also for fun, I set the lower bound to 7, so you'll see computation starts at 7 (vs. the usual 2).
-=====Check Results=====
-If you'd like to compare your implementations, I rigged up a Makefile checking rule called "**make check**" which you can run to get a nice side-by-side comparison of your implementations.
-In order to work, you **MUST** be in the directory where your pnc1 binaries reside, and must be named as such (which occurs if you ran **make** to compile them).
-For instance (running on my implementation of the pnc1 programs, some output omitted to keep the surprise alive):
-<cli>
-lab46:~/src/discrete/pnc1$ make check
-=========================================================================================================================
-      qty   regmo  regbmo   regms   regma   regos   regoa  regmos  regmoa  regbms  regbma  regbos regbmos  regboa regbmoa
-=========================================================================================================================
-  0.0001  0.0001  0.0001  0.0001  0.0001  0.0001  0.0001  0.0001  0.0001  0.0001  0.0001  0.0001  0.0001  0.0001
-  0.0002  0.0002  0.0002  0.0001  0.0002  0.0001  0.0002  0.0001  0.0001  0.0001  0.0001  0.0001  0.0001  0.0001
-  0.0005  0.0003  0.0002  0.0002  0.0002  0.0002  0.0001  0.0001  0.0002  0.0001  0.0002  0.0001  0.0002  0.0002
-  0.0021  0.0011  0.0003  0.0003  0.0003  0.0003  0.0003  0.0002  0.0003  0.0003  0.0002  0.0002  0.0002  0.0002
-  0.0096  0.0039  0.0009  0.0009  0.0008  0.0008  0.0006  0.0006  0.0007  0.0006  0.0005  0.0004  0.0004  0.0004
-  0.0443  0.0158  0.0026  0.0025  0.0023  0.0021  0.0016  0.0015  0.0016  0.0015  0.0011  0.0010  0.0010  0.0009
-...
-  ------  ------  ------  ------  ------  ------  ------  ------  ------  ------  ------  ------  ------  ------
-=========================================================================================================================
- verify:     OK      OK      OK      OK      OK      OK      OK      OK      OK      OK      OK      OK      OK      OK
-=========================================================================================================================
-lab46:~/src/discrete/pnc1$
-</cli>
-If the runtime of a particular prime variant exceeds an upper runtime threshold (likely to be set at 1 second), it will be omitted from further tests, and a series of dashes will instead appear in the output.
-If you don't feel like waiting, simply hit **CTRL-c** (maybe a couple of times) and the script will terminate.
-I also include a validation check- to ensure your prime programs are actually producing the correct list of prime numbers. If the check is successful, you will see "OK" displayed beneath in the appropriate column; if unsuccessful, you will see "MISMATCH".
-Analyze the times you see... do they make sense, especially when comparing the algorithm used and the quantity being processed? These are related to some very important core Computer Science considerations we need to be increasingly mindful of as we design our programs and implement our solutions. Algorithmic complexity and algorithmic efficiency will be common themes in all we do.

Lab46 Wiki

User Tools

Site Tools

Differences

Page Tools