Windows 7 Versus Ubuntu Linux

0 Comment

A. Project Plan The purpose of this project is to determine the fastest computer operating system for solving mathematical equations. Which operating system, Windows 7 or Ubuntu Linux, performs the best when it comes to pure number crunching? In order to test this question, the researcher ran a Java based application that benchmarks the ability of each operating system to perform a mathematical operation across three different data types and a second, more complex, mathematical operation.

The Java based application, Java Micro Benchmark, performs each mathematical operation eight times per test and the researcher ran the test five times for each operating system. The mathematical operations testsed by Java Micro Benchmark are integer arithmetic, long integer arithmetic, double-precision floating point arithmetic and trigonometric arithmetic. This method gave the researcher data for forty tests per mathematical operation for each operating system. The resulting data is then compared to determine which operating system is able to process mathematical operations the fastest.

Research into which operating system performs better is extremely relevant in today’s technological environment. Every task that a computer performs is a mathematical operation. An area that is highly affected by the performance of the operating system is computer programming. When a computer programmer writes a program, the program must be converted from a programming language into a language the computer can understand, a process called compiling. Compiling is very hardware intensive and completely dependent on mathematical operations. The faster a computer is able to perform mathematical tasks, the faster compiling will run as well.

Knowing which operations system performs better can save software developers time, which in turn saves money in development. These savings can then be passed on to the consumer, as well as leaves more time for programmers to develop new programs. A1. Literature Review Windows 7 by Microsoft is a proprietary, closed source operating system based on Microsoft’s NT platform. Windows 7 was released in October of 2009. Ubuntu on the other hand is an open sourced, community driven operating system. As with all Linux operating systems, Ubuntu is based on the Unix operating system.

Ubuntu was first released in October 2004, with its latest release, version 11. 04, being released in April 2011. In the computer world, it has been highly debated on which operating system is the best. When looking at an operating system, there are many variables that go into deciding what would be the proper operating system for different needs. In this study, the researcher focuses on pure performance in regards to mathematical operations, rather than other factors such as aesthetics, usability, supported programs, etc. The focus on mathematical operations directly correlates to the operating systems ability to compile programs faster.

Michael Larabel from the website Phoronix. com did a benchmark comparison of Ubuntu and Windows 7 in August of 2010. Larabel tested several benchmarks including graphics, Hard Disk Read/Write performance and RSA encryption performance (Larabel, 2010). The benchmark test I was most interested in was that of the RSA encryption test. Encryption of any sort is very processor intensive and correlates well to the experiment demonstrated in this paper. Larabel concluded that Ubuntu was more than twice as fast than Windows 7 in his benchmark. On the other hand, in February of 2009 the website TuxRadar. om performed some benchmarking tests of its own. Like Larabel, TuxRadar performaed several benchmarks that tested several aspects of the operating system. We will focus on the benchmark that tests compute-intensive tasks.

TuxRadar’s results showed Windows 7 edging out Ubuntu by a slight margin running a 64-bit OS and Windows 7 32-bit outperformed Ubuntu 32-bit by a large margin (TuxRadar. com, 2009). A2a. Experimental Design Steps Prepare Computer for Test: For this experiment I used the same computer for all tests to avoid any conflicts in data due to hardware. The computer used is a laptop with an Intel i7 2630QM processor clocked at 2. Ghz and capable of a turbo up to 2. 8 Ghz. The processor has 4 cores with hyperthreading allowing for 8 threads. The laptop is also equipped with 8 GB of DDR3 RAM and a Solid State hard drive. The rest of the specifications of the laptop are irrelevant to the benchmarking tests. I split the hard drive into two partitions allowing installation of both operating systems on the same laptop. On the first hard drive partition, Windows 7 Professional 64-bit was installed. On the second hard drive partition, Ubuntu Linux 11. 04 64-bit was installed. To ensure validity of test, Sun Java version 1. 6. _26 was installed in both operating systems.

Also installed in both operating systems was the benchmarking tool Java Micro Benchmark version 05. 061. Networking and all programs were disabled to ensure no running programs would adversely affect the results. Perform Benchmarking Tests On Windows 7: The computer was booted into the Windows 7 operating system and the Java Micro Benchmark program was started. Once started, Java Micro Benchmark was set to run its Desktop benchmark, which performs four different mathematical calculations eight times. The benchmark test runs without user input and when finished displays the results.

The results are given for each mathematical operation for each of its runs and also give an overall index score for all tests. The results are recorded. This process is repeated five times giving the researcher a total of forty tests for each mathematical operation. Perform Benchmarking Tests On Ubuntu: The computer is booted in the Ubuntu operating system and Java Micro Benchmark was started. Java Micro Benchmark was again set to run its Desktop benchmark and the test was started. The test was repeated five times to give the researcher forty results for each mathematical operation. A2b.

Reasoning This method of research was designed to test the ability of Windows 7 and Ubuntu operating systems to perform mathematical operations. Java Micro Benchmark was written to test mathematical operations in order to help computer programmers write the most efficient programs. The same computer was performed on all tests to alleviate any hardware concerns that would affect the results. The same version of Sun Java and Java Micro Benchmark was used in both operating systems to standardize the testing environment and rule out any software discrepancies. The only variable between the two esting environments was the operating system itself. A2c. Data Collection Quantitative data should be collected after each test run.

The data collected is averaged for each mathematical operation for each operating system and put into a graph to display results. A2d. Tools * Laptop computer with hard drive partitioned into two partitions * Windows 7 64-bit installed on one partition * Ubuntu 64-bit installed on second partition * Sun Java version 1. 6. 0_26 installed in both operating systems * Java Micro Benchmark installed in both operating systems A3. Variables Independent Variable: Operating System, Windows 7 or Ubuntu

The independent variable, the operating system, is predicted to influence the results of the dependent variables benchmarks scores. Dependent Variable: Benchmark scores in operations per second The dependent variable, the benchmarks scores, is the observed result of the performance of the operating systems. Controlled Variables: Same computer used in all tests, same version of Sun Java used on both operating systems, the same version of Java Micro Benchmark used in all tests, all other software on operating systems disabled, networking turned off on both operating systems.

All controlled variables were kept constant during the experiment. A4. Threat Reduction to Internal Validity All variables are to be kept constant except for the operating system itself. Same versions of testing software are to be used. The exact same computer is to be used for all tests to avoid any hardware discrepancies. The benchmarking tests are to performed five times on each operating system to give an overall average of performance. A5. Hypothesis The Ubuntu operating system will perform better in processing mathematical operations.

This hypothesis was developed after researching benchmarking results from various sources and from personal experience using and compiling programs in both operating systems. B. Data Collection For each test run in Java Micro Benchmark, the researcher collected the results. The results included benchmarks categories for integer arithmetic, long integer arithmetic, double-precise floating point arithmetic, and trigonometric arithmetic. Each benchmark test produced eight results for each benchmark category. An overall index score was also produce for each test performed.

These results were recorded and the test was performed a total of five times for each operating system. After testing was complete, each benchmark category was averaged to give a final result. The overall index scores were also averaged to give a complete score for each operating system. Benchmark Results Table: Measurements in Millions of Operations Per Second Operating System| Integer Arithmetic| Long Integer Arithmetic| Double-Precision Arithmetic| Trigonometric Arithmetic| Overall Index Score| Windows 7| 647. 91| 211. 63| 472. 1| 2. 44| 2668. 14| Ubuntu Linux| 690. 1| 321. 86| 477. 68| 2. 53| 2984. 37| (Table 1) The first row of the table lists the benchmarking categories tested and the Overall Index Score. The first column lists the operating systems tested. B1. Appropriate Methods The data collected in the table above was collected directly from benchmarking application after every test. Once all tests were completed, the data was averaged to get the final results. C. Results (Chart 1) The chart represents the operating systems’ performance in calculating integer arithmetic, represented in millions of operations per second. Chart 2)

The chart represents the operating systems’ performance in calculating long integer arithmetic, represented in millions of operations per second. (Chart 3) The chart represents the operating systems’ performance in calculating double-precision floating point arithmetic, represented in millions of operations per second. (Chart 4) The chart represents the operating systems’ performance in calculating trigonometric arithmetic, represented in millions of operations per second. (Chart 5) The chart represents the overall index score given by the benchmark application. D.

Conclusion At conclusion of the experiment, the researcher was surprised to find that, although Ubuntu was the clearly faster at computing mathematical operations, Windows 7 performed exceptionally well. In three of the four benchmarks, integer arithmetic, double-precision arithmetic, and trigonometric arithmetic, Windows 7 was fairly close in performance to Ubuntu. Ubuntu was clearly dominant in the area of long integer arithmetic. One thing to consider is that during high volume calculations, like those needed to compile a computer program, those small differences can add up. D1.

Based upon the findings in this experiment, it is evident that Ubuntu is out performs Windows 7 in mathematical calculations and therefore the initial hypothesis can be accepted. D2. Experimental design was a key factor in this experiment because it required the researcher to consider all variables that could have adverse effects on the results of the experiment. The researcher was required to ask a question, and then through scientific inquiry, answer that question. Without proper planning of the experimental design, the results of the inquiry could be favor one independent variable over the other.

In this experiment, all variables were ruled out except the variable for which the experiment was designed to test; whether Windows 7 or Ubuntu were faster at mathematical operations. D3. Replication Science experiments should be replicated because an experiment conducted in the same way, with same conditions should provide similar results. By other researchers duplicating the experiment and concluding the same results, the original results concluded from this experiment can be justified as confirmed.

Furthermore, replication of the experiment proves it as a valid experiment. If other researchers find that the experiment is flawed, then those flaws need to corrected for the experiment to remain valid. E. Citations Larabel, Michael (2010) “Workstation Benchmarks: Windows 7 vs. Linux” Retrieved from http://www. phoronix. com/scan. php? page=article&item=ubuntu_win7_ws&num=1 TuxRadar. com (2009) “Benchmarked: Ubuntu vs Vista vs Windows 7” Retrieved from http://www. tuxradar. com/content/benchmarked-ubuntu-vs-vista-vs-windows-7

Leave a Reply

Your email address will not be published. Required fields are marked *