I hear that alot, but it's not the ram use that's the issue. it's the time spent caching all that crap to ram when it's only going to be cleared again when I actually want to do something. Caching should look ahead when it suspects sequential access, and nothing more. If I wanna preload stuff, I'll preload stuff. If I/O performance is terrible in some application, I'll ask the developers why they haven't thought about using a caching daemon. Otherwise, stop your incessant caching and let me use my computer, thanks.
I use regcleaner to remove windows services. I don't recall how I got it, but I'm pretty sure it was adware/spyware free and it works a treat. Kill any processes you don't want to load [such as the realplayer and QT daemons, adobe loader, apple update, jsupdate etc] and then remove them using the 'startup' tab on regcleaner. You may have to do this once every few months, but it has been well worth it. You can get the same functionality by browsing and modifying your registry yourself, but that's just icky.
The main difficulty I've found with getting windows to boot fast is to stop it trying to cache your whole hard disk to ram, while at the same time creating a giant page file. I've seen "sort of" ways to do that, but most of them seem too dangerous, difficult or obscure to me. In fact, computers with less ram and smaller hard drives seem to load windows a little faster than those with more.
If you're running windows in a VM, however, you can probably let it view your local disks as a network disk. That way, there's little caching and no paging. I've not tried that, but it would be interesting to know if anyone has. BTW, unless you're a heavy gamer without PP [multicore or multiprocessor], you'd probably get better performance and more Linux-y functionality with a VM-ed windows anyway.
Hmm, I don't know. Debian boots so fast on my machine [20 seconds] I never bother with hibernate anymore. XP takes nearly a minute to get to the desktop.
Or no office. Unless you're doing accounting or something, who needs word, powerpoint or excel? And who would pay $75 for stuff like that anyway? LaTeX doesn't do spreadsheets, but it makes powerpoint look oh-so 1995.
Great idea. Microsoft's OEM licensing requires a separate production line for machines if they don't want to pay the cpu-tax, but this might get around that.
They're using virtual machines anyway, so there's no issue with running windows on one. I'm not sure what the deal is with VMs on SPARCs, though. But it's true, it's easier and cheaper to switch to *nix and remove the need for more ram in the process.
Actually, it would be nifty if NVIDIA/ATI offered a mobile DX10 solution for an otherwise perfect gaming machine. And the fact that they haven't done so doesn't mean they couldn't. The way I read it, it looks like you're suggesting the author is blaming Dell for not providing DX10 hardware support.
It might be just as terrible that we can deny these young people their sexuality and/or independence for so long. It's a serious breach of freedom. Views are mixed and even I'm not sure where I stand, physical age is a pretty bad way to judge the sort of maturity needed for romantic relationships, I've met 25 year olds that shouldn't be allowed to marry, but I don't see anyone suggesting we raise the bar to prevent that. What can you do?
Is that a strong enough argument for introducing moving parts into an otherwise solid state system? I know that wasn't what was said above, but does this fix the issues introduced by that?
Diamond makes a fantastic semiconductor, and it's more heat tolerant than silicon. The drawback, however, is that it's not easy to make diamond slivers:P
There are a lot of applications where taking advantage of a large number of threads is a fantastic way to increase performance. However, there is a lot of lower-level stuff that massively multicore CPUs should look into. Vector and matrix operations, FFTs, and single instruction multiple data [especially SIMD multiple compare, where many possibilities can be checked at once, eg for regexes, which is similar to the method d-wave uses for solving NP-complete problems with quantum computing] are the sort of instructions we are talking about- really using all those processors together. When the big CPU people focus on the sort of algorithms that crop up in scientific computing and database access, we should see the real power of top-down multicore design.
Heh, Ruby for web apps. That makes about as much sense as using PHP for doing the sort of abstract manipulations Lisp was created for. Or using candles to toast bread. You could do it, and I'm sure people do every morning in cubicles around the world, but I don't really think that's what it's made for.
Slashdot Mirror
I hear that alot, but it's not the ram use that's the issue. it's the time spent caching all that crap to ram when it's only going to be cleared again when I actually want to do something. Caching should look ahead when it suspects sequential access, and nothing more. If I wanna preload stuff, I'll preload stuff. If I/O performance is terrible in some application, I'll ask the developers why they haven't thought about using a caching daemon. Otherwise, stop your incessant caching and let me use my computer, thanks.
I use regcleaner to remove windows services. I don't recall how I got it, but I'm pretty sure it was adware/spyware free and it works a treat. Kill any processes you don't want to load [such as the realplayer and QT daemons, adobe loader, apple update, jsupdate etc] and then remove them using the 'startup' tab on regcleaner. You may have to do this once every few months, but it has been well worth it. You can get the same functionality by browsing and modifying your registry yourself, but that's just icky. The main difficulty I've found with getting windows to boot fast is to stop it trying to cache your whole hard disk to ram, while at the same time creating a giant page file. I've seen "sort of" ways to do that, but most of them seem too dangerous, difficult or obscure to me. In fact, computers with less ram and smaller hard drives seem to load windows a little faster than those with more. If you're running windows in a VM, however, you can probably let it view your local disks as a network disk. That way, there's little caching and no paging. I've not tried that, but it would be interesting to know if anyone has. BTW, unless you're a heavy gamer without PP [multicore or multiprocessor], you'd probably get better performance and more Linux-y functionality with a VM-ed windows anyway.
Hmm, I don't know. Debian boots so fast on my machine [20 seconds] I never bother with hibernate anymore. XP takes nearly a minute to get to the desktop.
don't most mayors manage their cities in EA/Maxis SimCity? Those mayors have had that data available since the mid nineties.
It's for cataloging purposes, as far as I understand. Something to do with sorting large databases by month first. [it's pretty stupid though.]
Or no office. Unless you're doing accounting or something, who needs word, powerpoint or excel? And who would pay $75 for stuff like that anyway? LaTeX doesn't do spreadsheets, but it makes powerpoint look oh-so 1995.
Great idea. Microsoft's OEM licensing requires a separate production line for machines if they don't want to pay the cpu-tax, but this might get around that.
They're using virtual machines anyway, so there's no issue with running windows on one. I'm not sure what the deal is with VMs on SPARCs, though. But it's true, it's easier and cheaper to switch to *nix and remove the need for more ram in the process.
Actually, it would be nifty if NVIDIA/ATI offered a mobile DX10 solution for an otherwise perfect gaming machine. And the fact that they haven't done so doesn't mean they couldn't. The way I read it, it looks like you're suggesting the author is blaming Dell for not providing DX10 hardware support.
It might be just as terrible that we can deny these young people their sexuality and/or independence for so long. It's a serious breach of freedom. Views are mixed and even I'm not sure where I stand, physical age is a pretty bad way to judge the sort of maturity needed for romantic relationships, I've met 25 year olds that shouldn't be allowed to marry, but I don't see anyone suggesting we raise the bar to prevent that. What can you do?
Is that a strong enough argument for introducing moving parts into an otherwise solid state system? I know that wasn't what was said above, but does this fix the issues introduced by that?
Diamond makes a fantastic semiconductor, and it's more heat tolerant than silicon. The drawback, however, is that it's not easy to make diamond slivers :P
There are a lot of applications where taking advantage of a large number of threads is a fantastic way to increase performance. However, there is a lot of lower-level stuff that massively multicore CPUs should look into. Vector and matrix operations, FFTs, and single instruction multiple data [especially SIMD multiple compare, where many possibilities can be checked at once, eg for regexes, which is similar to the method d-wave uses for solving NP-complete problems with quantum computing] are the sort of instructions we are talking about- really using all those processors together. When the big CPU people focus on the sort of algorithms that crop up in scientific computing and database access, we should see the real power of top-down multicore design.
Heh, Ruby for web apps. That makes about as much sense as using PHP for doing the sort of abstract manipulations Lisp was created for. Or using candles to toast bread. You could do it, and I'm sure people do every morning in cubicles around the world, but I don't really think that's what it's made for.