Mailing List Archive
tlug.jp Mailing List tlug archive tlug Mailing List Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]Re: [tlug] Possible command to boost to laptop performance
- Date: Mon, 22 Nov 2010 13:16:19 +0900
- From: "Stephen J. Turnbull" <stephen@example.com>
- Subject: Re: [tlug] Possible command to boost to laptop performance
- References: <4CCF948D.6020707@example.com> <AANLkTi=oqGNRxz20OwaJ=yx10_G0txFE7ThSMZDw5HJX@example.com> <4CD15049.8080503@example.com> <AANLkTimUVaxD8XZFz8dLKx4hdmsfgujTRvdM-L1RqD_J@example.com> <87oca5j0b2.fsf@example.com> <4CE92F8A.3080600@example.com>
Dave M G writes: > After having played with it a while, the speed improvement I'm seeing is > only to do with switching programs. Sometimes, when the memory usage is > high, it can take a minute (seriously) to switch between Netbeans and > Firefox. I can't hear the hard drive swapping, since it's a very quiet > HD, but I assume that's what's happening. Ah, I think I see what's happening, then. If you clear the cache first, then probably this is the story: 1. Linux mmaps the program's executable pages. 2. Linux starts it running, which because of mmap means that it swaps in executable code on demand (ie, if the next instruction isn't in memory, the page gets swapped in). 3. Because the program is running, its pages get high priority for optimistic caching in the background -- later, the probability that its pages are already in memory goes way up. But if the cache is already full, pages get swapped in only on demand, which is much slower than if they're already in memory. Still, if this is the story, as the # of pages actually read by the process goes up, the cache ought to go down so that the running process needs to swap less. Over time (a dozen switches or so) memory ought to end up mostly populated with frequently used code from your two most frequently used applications. So if you use only those two apps for a while, the switch time ought to go down to a couple of seconds. Does that fit you experience? If not, I suspect there's a bug in the memory allocation algorithm. P.S. Dave, you ought to talk to a business consultant about your budget issues. It sounds like you're working quite a bit but you're not able to afford the equipment to work efficiently. (I know it's fun to root your phone, but you should be able to buy a Galaxy for your job and play with your HT-03A, right?) While I can't promise you that a consultant can help you with the cash flow, many times they can, at least enough to let you concentrate on work rather than infrastructure. I don't know any *good but affordable* ones offhand, but it might be possible to get you some *free and possibly useful* advice from our MBAs-in-training.
- References:
- [tlug] Possible command to boost to laptop performance
- From: Dave M G
- Re: [tlug] Possible command to boost to laptop performance
- From: Marty Pauley
- Re: [tlug] Possible command to boost to laptop performance
- From: Edward Middleton
- Re: [tlug] Possible command to boost to laptop performance
- From: Josh Glover
- Re: [tlug] Possible command to boost to laptop performance
- From: Stephen J. Turnbull
- Re: [tlug] Possible command to boost to laptop performance
- From: Dave M G
Home | Main Index | Thread Index
- Prev by Date: [tlug] Rooting the Android phone rocks! (Also: screw Docomo!)
- Next by Date: Re: [tlug] Rooting the Android phone rocks! (Also: screw Docomo!)
- Previous by thread: Re: [tlug] Possible command to boost to laptop performance
- Next by thread: Re: [tlug] Possible command to boost to laptop performance
- Index(es):
Home Page Mailing List Linux and Japan TLUG Members Links