SSD Tweaks

[AkG]

IIRC the 640 and 1TB models should be just above 2mb/s on 4k writes (and I may be wrong...been awhile since I looked at the wee little 'uns). Nothing terrible with those numbers, but a little low.

[Perineum]

Thanks a lot AkG

[freeloader1969]

Don't move the pagefile off the SSD.

Support and Q&A for Solid-State Drives - Engineering Windows 7 - Site Home - MSDN Blogs

[gingerbee]

no thanks I will keep it this way no need to write to it at all if not need and have the page file on the ssd or hdd makes no speed difference. all this states is that 7 has optimized it pagefile about damn time. great info thanks for the link. i just run enough ram

[freeloader1969]

Originally Posted by gingerbee

no thanks I will keep it this way no need to write to it at all if not need and have the page file on the ssd or hdd makes no speed difference. all this states is that 7 has optimized it pagefile about damn time. great info thanks for the link. i just run enough ram

You could have 64 gigs of ram in your system and Windows would still use the pagefile. Of course, if you have 8 gigs or more, you can try disabling the pagefile altogether.

[AkG]

Two ways of looking at this issue.

For a slight increase of the pg file "speed" (or "less bad" decrease in speed vs RAM) you are burning up cell life. Not a very good trade-off IMHO.

For a slight decrease in the pg file "speed" you can drop it to a wd balck 640gb (cheap) or vraptor 150 (also cheap'ish) and save your ssd cells for actual important things!

Guess which one I chose and recommend ;)

The MS Blog is technically correct...but it is too narrowly focused. ALL usage of the pg file is slow compared to ram. Why shorten the life of your ssd for a smaller pause? And in all likely hood it was looking at it from a workstation / corp enviro POV where the ssd is just another fungible good. You are much better off moving it off the c drive and sticking in 6, 8, 12 or 16gb of ram into your system so the pg file is accessed VERY infrequently and when it is its not burning out your SSD on next to useless writes.

YMMV

[lowfat]

So far I've only ever heard of one SSD that have had its cells burned out. And that was because he ran Iometer doing 4k writes for 2 weeks straight. Every other failure I've seen has been to controller failure. Personally I think people are taking the NAND cell life to seriously. Do you seriously expect to use the same SSD as your boot drive 3-4 years down the road? Besides when using SSDS w/ the JM602 I have never taken my page file off the SSD. Even after a year my X25-M (which I thrash w/ writes) is still going strong. My 30GB Vertex also lasted a year as a boot drive, well until it had a controller failure.

[AkG]

While I personally change SSDs faster then some people change their wardrobe....yes, I actually DO expect most consumers to use their expensive kit for many years to come. Why wouldnt they?
Why reduce the life of the drive for no noticeable real world gain? Its not like its hard to swap the pg file over to a HDD. So why not? Of course my POV is: RAM > pg file. YMMV lowfat. :)

[enzo matrix]

I have found what lowfat has said interesting. And I would like to see some kind of evidence that either proves modern SSDs can die within an amount of time that easily or proof contrary to that idea.

Also, on an earlier linked page was this:
Should the pagefile be placed on SSDs?
Yes. Most pagefile operations are small random reads or larger sequential writes, both of which are types of operations that SSDs handle well.
In looking at telemetry data from thousands of traces and focusing on pagefile reads and writes, we find that
Pagefile.sys reads outnumber pagefile.sys writes by about 40 to 1,
Pagefile.sys read sizes are typically quite small, with 67% less than or equal to 4 KB, and 88% less than 16 KB.
Pagefile.sys writes are relatively large, with 62% greater than or equal to 128 KB and 45% being exactly 1 MB in size.
In fact, given typical pagefile reference patterns and the favorable performance characteristics SSDs have on those patterns, there are few files better than the pagefile to place on an SSD.

This makes sense to me. They don't mention the effects of it though. It seems to indicate the pagefile may not be nearly as dangerous as suggested considering that reads on a drive do not wear the drive out. It also makes sense that if one reboots rarely, then the wear would be even further reduced.

Any opinions or facts?

[lowfat]

I am going to add one more point here form something I have recently discovered. I have two Windows 7 machines that use SSD's as their boot drives at the moment. Machine 'A' has 2GB of ram, and a 40GB X25-V. The machine is a file server and rarely gets used besides ripping an occasional bluray. Machine 'B' has 9GB of ram and a 30GB X25-E. It gets frequently used however the only writes I really do to the SSD is to rip DVD's. So maybe 20-30GB of its writes are from that. The X25-V maybe has a month longer of use on it. Both of which are still rather young (3-4 months of use). Prefetch and Superfetch are completely disabled on both machines. Neither machine has has any games or large applications on the boot drive.

Now to my point. Both systems have the pagefile on the drive. Machine 'A' which only has 2GB of ram has written nearly 2TB to the SSD according to CrystalDiskInfo. While machine 'B' has written only 250GB to it. Remember that 20-30GB of that was from ripping DVD's. So if you are worried about keeping the page file on the SSD, just make sure you have enough ram on the machine to not cause excessive paging.

One other thing. Most writes done to the SSD from a page file should be sequential writes, which according to Intel cause a lot less wear to the cells.