Testing
Once the project computer is assembled, various tests will be conducted to
monitor the effect of computer modifications on the speed of mailto:SETI@Home operation. Different settings and components will be tweaked and replaced, and the results will be used to create a list of optimal conditions for SETI@Home. This list will be posted at setisciencefair.blogspot.com so other SETI@Home users can emulate it with their PC for maximum signal analysis.
One of the most important and delicate tests will be that of overclocking the computer, discussed briefly in the “The Computer” section. Overclocking is easy to try, but hard to master, and its nuances are discussed in the “Overclocking” section.
The importance of RAM (Random Access Memory) will also be examined. It is know that SETI@Home can accomplish its task with as little as sixteen megabytes of RAM, but the program will be tested with one 512 megabyte DDR2 module, then with two of these modules running in tandem (for a total capacity of one gigabyte). This would seem an irrelevant test, since SETI@Home, as mentioned, will only implement sixteen of these 512 or 1,024 (one gigabyte) megabytes, but since DDR2 is specifically designed to run in pairs of modules, writing efficiency and speed increase with two modules.
To record information streaming from the CPU, the RAM must run at least close to its interlocutor’s speed. If the CPU runs at 4.01 gigahertz and the RAM only records at 433 megahertz, the RAM will not remain up-to-date, and the system will either crash, or the BIOS (Basic Input/Output System), will lower the processor speed. Because of this, the RAM’s write speed will be overclocked simultaneously with the CPU. This, also, is discussed further in the “Overclocking” section. It is likely that many tweaks and changes will have little or no effect on mailto:SETI@Home running speed. This will not be a failure, but a crucial part of discovering which modifications are useful and which are not; extreme precision and useless data are facets of computing experiments; many of the performance differences that are detected will probably be small, and largely irrelevant to average SETI@Home volunteers. The publishing of these findings will be useful for extreme speed optimization and interesting from a conceptual perspective. Also, with more powerful computers in the years to come, certain small performance differences may become more evident or fall under greater scrutiny.
monitor the effect of computer modifications on the speed of mailto:SETI@Home operation. Different settings and components will be tweaked and replaced, and the results will be used to create a list of optimal conditions for SETI@Home. This list will be posted at setisciencefair.blogspot.com so other SETI@Home users can emulate it with their PC for maximum signal analysis.
One of the most important and delicate tests will be that of overclocking the computer, discussed briefly in the “The Computer” section. Overclocking is easy to try, but hard to master, and its nuances are discussed in the “Overclocking” section.
The importance of RAM (Random Access Memory) will also be examined. It is know that SETI@Home can accomplish its task with as little as sixteen megabytes of RAM, but the program will be tested with one 512 megabyte DDR2 module, then with two of these modules running in tandem (for a total capacity of one gigabyte). This would seem an irrelevant test, since SETI@Home, as mentioned, will only implement sixteen of these 512 or 1,024 (one gigabyte) megabytes, but since DDR2 is specifically designed to run in pairs of modules, writing efficiency and speed increase with two modules.
To record information streaming from the CPU, the RAM must run at least close to its interlocutor’s speed. If the CPU runs at 4.01 gigahertz and the RAM only records at 433 megahertz, the RAM will not remain up-to-date, and the system will either crash, or the BIOS (Basic Input/Output System), will lower the processor speed. Because of this, the RAM’s write speed will be overclocked simultaneously with the CPU. This, also, is discussed further in the “Overclocking” section. It is likely that many tweaks and changes will have little or no effect on mailto:SETI@Home running speed. This will not be a failure, but a crucial part of discovering which modifications are useful and which are not; extreme precision and useless data are facets of computing experiments; many of the performance differences that are detected will probably be small, and largely irrelevant to average SETI@Home volunteers. The publishing of these findings will be useful for extreme speed optimization and interesting from a conceptual perspective. Also, with more powerful computers in the years to come, certain small performance differences may become more evident or fall under greater scrutiny.
posted by Zak Rogoff at 3:54 PM
0 Comments:
Post a Comment
<< Home