need a program that...

N

Nastybutler187

New Member
measures the speed of your computer. i am doing a science fair project, and my question is if the speed of the computer will vary depending on the amount of % of hard drive space occupied (im only in 8th grade people, lol). so i need a program that i can run, and to see how fast my computer can open different programs.
 
Nastybutler187 said:
measures the speed of your computer. i am doing a science fair project, and my question is if the speed of the computer will vary depending on the amount of % of hard drive space occupied (im only in 8th grade people, lol). so i need a program that i can run, and to see how fast my computer can open different programs.
Click to expand...

Use a stopwatch.
 
Xelogen said:
That quote is going to go down in CF history :D
Click to expand...

Why?


I just didn't See why you would need a program for something so increadibly simple to do yourself.
 
Last edited:
flame1117 said:
Why?


I just didn't See why you would need a program for something so increadibly simple to do yourself.
Click to expand...
yeah i agree. its gunna go down as the most logical answer to have been posted lol
 
dude are you kidding me? this is most likely gonna come down to hundredths of a second, a stopwatch controlled by me would not work. Besides, we arent allowed to use humans (you know what i mean) in any part of the project. Thanks Geoff, i'll look into that.
 
Nastybutler187 said:
dude are you kidding me? this is most likely gonna come down to hundredths of a second, a stopwatch controlled by me would not work. Besides, we arent allowed to use humans (you know what i mean) in any part of the project. Thanks Geoff, i'll look into that.
Click to expand...


You just neeed nija-life reflexes like me.

and you should have been mroe specif then...

and my watch does to hundreth of a secend. 00.00.
 
Nastybutler187 said:
and my question is if the speed of the computer will vary depending on the amount of % of hard drive space occupied (im only in 8th grade people, lol).
Click to expand...

lol speed of computer and amount of hd space occupied, aren't those two totally different things? Did your teacher assign this or did you come up with the topic yourself?
 
i came up with the topic, but i will do tests on 2 programs: one that was downloaded before the addition of new programs, and one that was downloaded after, and see which one opens faster. the comp needs to look thru everything on your hard drive before it finds the program that you need, so i think it should take longer for the programs that were downloaded before the addition then it will take for the programs that were added after.
 
run a defrag, 3-4 times, make sure theres no red, then the system should run better no matter how much crap on the hdd
 
So your saying if you load a hd with "stuff" (being very un-spedific on purpose) it will slow down the speed at which it opens a program? I would agree if you put installed programs as the stuff on the drive.. but if you a hd 90% full with music, or movies (legal ones of coarse) It would not affect the speed the program opens. What you need to keep in mind when you install a program it also installs dll's and other things your computer will be running along side that opening program you will be speed testing. If your gonna load the hd, do it with data, not an installed program.

Also your worried about the time down to hundredths of a second, you should look into some other factors which will hinder the accuracy of the data. Those being... How long the computer has been on, the ambient temperature of the inside of the pc, and of the room. *if any other programs have been installed* Anyway hope that helps...

My hypothesis: The hd seek time on a loaded hard drive (data) compared to an empty hard drive will not effect the speed of the load to any *noticable* or significant amount.

Have fun :)
 
Trizoy said:
So your saying if you load a hd with "stuff" (being very un-spedific on purpose) it will slow down the speed at which it opens a program? I would agree if you put installed programs as the stuff on the drive.. but if you a hd 90% full with music, or movies (legal ones of coarse) It would not affect the speed the program opens. What you need to keep in mind when you install a program it also installs dll's and other things your computer will be running along side that opening program you will be speed testing. If your gonna load the hd, do it with data, not an installed program.

Also your worried about the time down to hundredths of a second, you should look into some other factors which will hinder the accuracy of the data. Those being... How long the computer has been on, the ambient temperature of the inside of the pc, and of the room. *if any other programs have been installed* Anyway hope that helps...

My hypothesis: The hd seek time on a loaded hard drive (data) compared to an empty hard drive will not effect the speed of the load to any *noticable* or significant amount.

Have fun :)
Click to expand...

when you say load it with data, do you mean like.. spam word documents? im not too good with computers, which is why im doing this, please be more specific. thanks
 
my question is if the speed of the computer will vary depending on the amount of % of hard drive space occupied (im only in 8th grade people, lol).
Click to expand...

dude are you kidding me? this is most likely gonna come down to hundredths of a second, a stopwatch controlled by me would not work. Besides, we arent allowed to use humans (you know what i mean) in any part of the project. Thanks Geoff, i'll look into that.
Click to expand...
Ok. Now not to rain on your parade or anything but there are a few technical issues that make the relationship between a the free-space and the "speed of the computer" an issue:
  1. Im going to assume by "speed of the computer" you mean "realworld performance" (because otherwise CPU clockspeed and HDD have very little to nothing in common
  2. Even if you keep the rest of the system constant (so that you can attempt to isolate the effect of the HDD on the system performance), drive caching, burst transfers and even fragmentation will affect the results
    • Suppose we have a HDD with 8MB of cache ... now one of the drive controller's 'jobs' is to keep that high speed cache filled ... so if by chance, you're reading data from the drive and it happens to be already cached then the response time will be on the nanosecond/microsecond scale versus the millisecond scale -- several orders of magnitude difference there -- and no way to determine if a drive IO is using that cache or not. Naturally the cache impacts read performance much more so that write performance
    • On the same lines as the cache, suppose you're doing a read-operation with a high cache hit ratio: after 8MB, the cache will have to be re-filled (technically this is inaccurate as the cache is on-the-fly-filled but wont matter too much as i explain the next issue) .. the issue of burst mode transfer now kicks in:
    • The general principle of harddrive transfers is that "if you want data from location1 then odds are you want data from location2, location3, location4 etc ... so what happens is the HDD, when you request data from location1, will automatically start feeding data from location2, location3 etc anyways.... now this is good if the drive isnt fragemented and the data is what you want .... if not then you have issues: [1] for non-cached transfers, the drive will have to kick out of burst mode and seek for the proper data if you dont actually need/want data from location2, location3 etc and [2] for cached transfers, (relating to the above), the cache, upon being "used up" will be filled with new data ... which isnt needed ... which means the cache will have to be re-filled. Granted all this occurs on a microsecond scale but it all adds up
    • Naturally drive fragmentation makes life difficult by negating the bonus of sequential operations (i.e., the drive head has to find the next data segment or has to go look for an empty sector) .. this is lag on a millisecond scale: and while you can defragment data to minimize this effect, you can never really get rid of it.
    • Now it seems that there are a few options that present decently normalized data: [1] small (i.e., smaller than the size of the cache) sequential read operations (unfortunately this will give you much higher than normal data) [2] large (i.e., few hundred MB or bigger) sequential read/write operation (which will then depend on drive geometry see next), (3) large random drive write operations (which, even for defragged drives, will tend to give much lower results than normal)
    • About drive geometry: for HDDs which run CAV (constant angular velocity), the inside tracks will yield higher performance and defragementation will attempt to arrange things near the center where possible so for the most part you can rest assured with good defragging however [1] when you repeat a test -- even if you defrag it again -- there's no guarantee the defrag operation will be identical (im thinking a drive image operation can negate this effect though) (2) while defragmentation attempts to arrange the data near the center on any given platter ... there are multiple platters ... and no way to determine which platter is currently in use and whether or not a platter change is needed (which is a milisecond scale operation)
  3. Naturally stuff like virtual memory, virus scanning and other background ops and services would need to be terminated
  4. If you were to test the performance of the drive by say, "opening a word document", you'd be measuring two things: the time it takes to start word, and the time to load the document -- with no way to determine the breakdown of latency between the two

and my watch does to hundreth of a secend. 00.00.
Click to expand...
Yes but not really. Its not accurate to 2 decimals.
 
You know they would have to be basically the same program, like the same amount of data that has to be loaded to get good results.
 
dosent look like this is gonna work out butler, why not try something else a little more simple...

How about testing how many frames per second different games have at different resolutions or something like that?

If your stuck on using your computer maybe someone else could recommend another test you could carry out.

Mike
 
How about testing how many frames per second different games have at different resolutions or something like that?
Click to expand...
FPS is bound by RAM, video, CPU, chipset and other "live" (aka run-time) constraints ... by the time you're playing the game, you're probably wont be dealing too much with virtual memory and caching
 
Maybe you could do a project on how relevant CPU clockspeeds actually are, I hear people all the time saying "Well my Intel processor is 3 ghz, and your AMD 64 is only 2 ghz, so my Intel is obviously better." That is a common mistake.
 
You must log in or register to reply here.
Back
Top