On Sun, 2004-10-24 at 10:46 +0200, Jan Kesten wrote: > > Buffers: Relatively temporary storage for raw disk blocks > ~ shouldn't get tremendously large (20MB or so) > ~ Cached: in-memory cache for files read from the disk (the > ~ pagecache). Doesn't include SwapCached Yes. > Cached instead is a cache for files read from disk. You notice it's > funtion if you access a file. First time you access it, it takes some > time since it is loaded. If you close it and then reopen it, this is > much faster: > > teefix:~# time cp linux-2.6.8.1.tar /dev/null > real 0m10.456s > user 0m0.049s > sys 0m1.408s > teefix:~# time cp linux-2.6.8.1.tar /dev/null > real 0m0.975s > user 0m0.034s > sys 0m0.911s > > After doing much disk access: > > teefix:~# time cp linux-2.6.8.1.tar /dev/null > real 0m10.474s > user 0m0.036s > sys 0m1.433s > teefix:~# time cp linux-2.6.8.1.tar /dev/null > real 0m1.007s > user 0m0.036s > sys 0m0.936s > > Hope I'm not too wrong in my explanation - if drop me a mail :-) > You are correct. What the tests above show is that the CPU time spent is always relatively the same (user & sys). What is changing is "real" time which is actual wall time. So the question you need to ask is... if the time spent on the CPU isn't changing why is it taking me 10 seconds to cp this file the first time but only one second on the following try? It's because a process blocks on I/O. If a process needs data from disk then it is taken off the CPU and put to sleep while the I/O request is being handled. In the meantime another process is placed on the CPU. This is referred to as a "context switch". There is no reason why a process should sit idle on the CPU while it waits for some disk I/O request. This forms the basis for "multitasking". So what these tests demonstrate is that if we run the command in succession we see the affect of pagecache after the initial disk I/O takes place. The pagecache has shaved nearly 9 seconds off the time which is a big deal. We can also see that as the disk is accessed more and more the pages that once sat in pagecache have been replaced by more recent ones. -- Eric Gaumer <gaumerel@ecs.fullerton.edu>
Attachment:
signature.asc
Description: This is a digitally signed message part