Make your SSD Speed Really Offensive

Hello Playmates Since I enquired a few posts back about Pagefile relocation etc.. I've been tinkering about with different configurations etc to speedup My Corsair Force 3. Well yesterday there was a Exhibition at SciTech (Perth Australia) and Corsair were actually there promoting their Newest range (Blah Blah Blah) Anyway in short heres how to do it. Set the page file custom size to 1024 and the maximum size to whatever amount of physical RAM you have installed, so for instance I have 16GB so Min /Max is 1024-16384 disable Hiberfil.sys / disable Disk Defragmentation from within services (Just turning off the scheduler wont cut it) and leave everything else Alone. I did this, and I now have 100GB Free on my 120GB SSD and the size for Windows is just 11.1 GB reboot speed, from shutdown to desktop, is a tiny 17 seconds Also everything else opens in the Blink of an Eye

 

DVD+R,

So the only thing you changed was the pagefile? Interesting. What was the performance like before the change?

 

With leaving the SSD alone after a clean installation of Windows 7 Professioanl x64-bit with SP1 Cold Start was around 15 seconds from turning on the PC first thing in the Morning, I have shaved 8 seconds off of this Cold start time, I'm only using 1.86GB of RAM of the 16GB I have installed 1268 of that is cached and only 1% CPU usage and Windows Experience Index has jumped from 7.4 to 7.9

 

Maybe you could get even better performance with a much smaller pagefile (minimum setting) and a setting of Small Memory Dump.

 

I have, tried 16MB, however that reduced the Windows Experince Index, even though this is an Old extra originally used on Vista, It still gives a good reading on W7

 

I don't think you should be dinking with the page file's size. Just let Windows manage it. Microsoft has spent years and $millions in research tweaking the PF and with Windows 7, it is a very mature, well refined feature. I know for a fact I am not smarter than all those engineers or their super-computers they use during design and testing.

The only time I recommend dinking with the PF size is when free disk space is limited, or you have a very small amount of RAM - and for the latter, I recommend increasing the PF.

It makes no sense to make the PF smaller. Why on Earth would you take resources away from Windows then think it will perform better? It is incorrect to assume a small or little PF will force Windows to use more of the (faster) RAM. Windows does not work that way. Windows will use any resource it can find. And that is a good thing, not bad. A small PF simply means needed data will NOT be readily available for the CPU to crunch and it will have to go into yet another wait state.

So what? Too much emphasis is being put on start times. In regards to performance, start times, mean NOTHING!!!!!! So what if it takes 3 or 4 minutes to boot? Assuming there are no errors (typically with drivers), long boot times generally mean you have several things loading at start. And again, so what? As long as performance is good once all has loaded, that's what you are looking for.

I say again, fast boots time are NOT an indication of performance. Faster boot times do NOT mean better! Period.

And certainly you cannot compare performance of two machines by the boot times, except if using the exact same hardware, HW configurations, software and software configurations, and about the only time you see that is with two consecutive machines rolling off the assembly line. But 10 minutes after first powered on, each machine will become unique as users are created, security and networking are setup and user programs installed.

Avira does not load in the same time as AVG and that does not load in the same time as MSE. Windows Firewall loads at a different speed than Comodo. Joe loads Secunia PSI, Fred loads SuperAnti-spyware. Jane loads SpeedFan, Jill loads CoreTemp. John has an external drive attached, Dave has two monitors. And no one has started "computing" yet.

If these are programs you regularly use during your computing sessions, it is much quicker to get them loaded at boot (where super and prefetch and can do their things) than to load each one individually later. And note, it takes several sessions for super and prefetch to "learn" you.

Unless you have advanced Computer Science degrees and years of expertise in programming operating systems and analyzing data management statistics, leave the Page File alone!

 

Mark Russinovich works for Microsoft now.

http://blogs.technet.com/b/markrussinovich/archive/2008/11/17/3155406.aspx

Do you have links to your opinions?

 

If you are going to take the time to fully analyze your computer usage with the several advanced analysis tools Mark created and suggested, then by all means, tweak the Page File.

For my opinions? Which ones, but generally, no. My opinions are just that. But for the facts I present, most are right there ready for Bing Google to find, or I try to rationalize with common knowledge, but I also do have links to some. Sadly, there are many many tricks and tweaks from XP days and even back to Win95 days that some folks still believe apply today. It is just not so.

I used ALWAYS recommend a fixed PF size with XP. But not with Windows 7 and today's monster drives with little excuse for not having lots of free space.

The option to set a fixed size is there for Windows experts. If you are a Windows expert, then go for it. You can always uncheck. But see this first, http://lifehacker.com/5426041/understanding-the-windows-pagefile-and-why-you-shouldnt-disable-it

 

This sums up Mark's analysis and it's what I've always used.

For 16 GB of RAM the calculation above will always be negative. If you only want a Small Memory Dump (that's all I want) then a pagefile of a few hundred MB is by far large enough.

Interesting. Your link suggests to use Mark's formula.

 

Good God! Had a Bad day have we Bill, with all that ranting Read again where I got the information to do what I did, I somewhat doubt Corsair technicians would be feeding their customers misleading, and negative advice

 

Judging by this, it's also relevant to me that your also extremely critical on the pagefile scenario so which is it, leave it or adjust it, make up your mind

 

Interesting. In another thread I reported the Win7 Index (Disk Data Transfer rate) increased from 5.9 to 7.7 by increasing the amount of free space in the Win7 partition. So free space does matter. As does pagefile size.

 

Dear Brain,

And, for this reason I suggest 25% free space. This free space also allows the Pagefile.sys to grow from minimum to maximum size and then still leaving some more free space for Windows performance and TRIM to work on SSDs.

Best regards,

 

I agree free space is needed.

If I had 16 GB RAM I'd use a 300 MB pagefile, 600 MB maximum. Different from what you are using.

 

Lets put this to bed shall we, I've looked up this from Intel and it says it all:

http://www.intel.com/support/ssdc/hpssd/sb/CS-029623.htm#5


So regardless of what we've been told, it seems the only thing to do is Buy it/ install it/ Forget about it

 

DVD+R,

Can you see a section on the amount of free space needed in a partition on a SSD? I can't. Wish I could.

 

Is this for Windows 7 or all types of Windows. Windows 8 has lower Pagefile.sys requirement than Windows 7. Each one is different.

Best regards,

 

The objective data is from Mark's blog. It needs to be read in its entirety to understand what the pagefile does.

 

Mark is no god. He is one of the employees of Microsoft. Different people have different opinion and proofs.

Take for example, Windows 7 sets the Pagefile.sys to 16MB minimum and the maximum is 1 and 1/2 times the total RAM memory. Where as Windows 8 sets the maximum Pagefile.sys to half the total RAM memory.

Now look at the OPs, where Corsair, suggests the minimum 1GB Pagefile.sys and maximum to the total RAM memory.

These are experienced engineers with their opinion and proofs.

Best regards,

 

But at least Mark has given a reasoned argument for managing the pagefile. Where is the reasoned argument for a pagefile being 1.5 times the amount of RAM or half the amount of RAM?

Another quote....

 

Try this Brian http://www.overclockers.com/ssds-trim-disk-space/

 

It makes no difference how big it is as long as it's big enough to hold the data. If your system never puts information in it it won't matter if it's 250mb or 16GB.

If your system puts 1GB of data into it it won't matter if it's 1GB or 100GB.

The system itself only uses about 250mb in case it crashes so you get page dumps with information on what caused the crash.

Otherwise it should only ever use it when the system is low on RAM. If your system is never low on RAM you should never hit the page.

So why turn it off?

Because some programs will page "early" - they'll do it when they don't need to. Sometimes this is so that if the program crashes you don't lose everything but most of the time it's just poor memory management.

If you have 8GB or 12GB or 16GB I think there's actually an article where Microsoft says it's safe to turn it off - though, as I have, they warn you that you'll lose pagedumps. (I believe MS draws the line at 12GB.)

There are a ton of "formulas" for how much RAM:agefile you need. It's silly.

If you've got an 8GB pagefile and you're swapping 8GB of RAM onto disk and back your system is going to be crawling and it's not going to matter.

The only time you should do that is if you're going ot be rendering tons of video content or editing massive files in photoshop. At that point you have no choice.

Otherwise, for most users, default settings are fine and if you've got 8+ GB of RAM you can feel free to turn it off. You will likely not run out of RAM.

 

Thanks. Wow, talk about conflicting information.

Are you 2 hours behind me? UTC+8?

 

changed my page file from 2GB to OP's recommendations and noticed no difference in speed, only extra space usage.

i would rather have slightly slower read speeds and more space rather than maximum speed and lower available disk space so i left put it back to 2GB (not that i noticed a speed decrease).

 

tree,

I'm confused. The OP used a 1 GB pagefile so if you reduced yours from 2 GB to 1 GB you should have had more free space. The 16 GB maximum should not have changed your free space. Am I on the wrong track?