-
We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.
How to CORRECTLY optimize your SSD for windows 7
Page 6 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
- Status
Or just use Everything. Because it hooks up into the NTFS journal, it finds everything instantly, including that file you just downloaded seconds ago, some obscure log file, a windows DLL, etc. Correct me, but I don't think there is a similar tool out there (every other "quick search thing" I've used requires indexing and can't find very new files). For that reason it ONLY works on Windows NTFS volumes, nothing else.
I only set Windows Search to index content (ie My Documents) and use it for content search. It doesn't even do a particularly good job of THAT, so most of time I end up using grep or equivalent to find content that I need.
Last edited:
Just downloaded and tested...holy freakin! ridiculously fast and exactly the kind of functionality I find to be productive in a search tool (the context menu options, etc).
And freeware? Now how the heck is it that Windows 7 couldn't ship with this kind of functionality?
Also, for you Hibernate + SSD users out there, checkout the the powercfg option. When you run an elevated command window you can set the size of the hibernate file as a percentage of your ram, as little as 50%.
^ this line sets your hibernate file to 50% your ram size...for me it reduced my 6GB hibernate file to 4GB on my SSD. (I have 8GB of ram)
taltamir
Lifer
http://www.voidtools.com/
I am curious, does anyone use google desktop? how is that in terms of searching?
I used google desktop but not anymore - it works but I don't really care for content indexing...I just need to know where the file is located.
Hacp
Lifer
I still prefer the windows 7 index. Hit the windows button, type in four letters and hit enter. 99% of the time, it'll bring up the file or program you want.
Google desktop hates my system for some reason. Don't know how to explain it, but it either causes program incompatibilities, doesn't index what it is supposed to, or even crashes.
I've tried a whole lot of other indexing tools - Copernic, X1, a random thing called Archivarius, even corporate grade Isys-Search, but really nothing beats Everything and manual content searching (i.e. grep or equivalent). I typically content search things like source code and XML files, and none of the programs above except for grep handles that well. Windows search is semi-useful for content searching.
I've tried a whole lot of other indexing tools - Copernic, X1, a random thing called Archivarius, even corporate grade Isys-Search, but really nothing beats Everything and manual content searching (i.e. grep or equivalent). I typically content search things like source code and XML files, and none of the programs above except for grep handles that well. Windows search is semi-useful for content searching.
For kicks, enable ETP and try the remote search. Even remote searching is almost as fast -- and supports getting the files too.
taltamir
Lifer
nope, they are exactly the same. I can verify that if you add a spindle drive, windows 7 automatically enables defragging only for that drive. And it only runs the service when a defrag is actually initiated (service is off when you bootup).
Win7 seems to know what it is doing in this regards.
How does Win7 implicitly know that a given drive is an SSD vs. spindle?
Is it just a look-up table type of a deal? Checks online to some big database that is warehousing all the drive ID's?
What I've read it's 1). That was some post on the win7 engineering blogs, no idea how to find that one though, but it also sounds like the most sensible solution.
The thing about the random access check was one thing they considered back then with those horrible first gen JMicron drives so they could not do some optimisations they usually did for SSDs, not sure how that worked out.
The thing about the random access check was one thing they considered back then with those horrible first gen JMicron drives so they could not do some optimisations they usually did for SSDs, not sure how that worked out.
GlacierFreeze
Golden Member
AHCI has to be enabled in BIOS (before OS install). It's "on" for all drives.
And how many damn games do you have? lol
And how many damn games do you have? lol
AdamK47
Lifer
Which newer Intel drivers would that be?
AdamK47
Lifer
660+. Legally installed and rangring from years 1988 to today.
Dorkenstein
Diamond Member
Is even a light defragmentation/optimization bad for an older SSD? I have a Supertalent Ultra Drive ME 64gb (before the renaming) and I was curious. I'm going to start saving for a bigger/better/badder SSD soon.
taltamir
Lifer
an SSD does not give the OS direct access to where the data is located. It uses a conversion table.
the OS says "file.iso is of such size that it takes 12 sectors to fill, SSD, place them in order from sector 120 to 132"
SSD says "sure OS, done", but it lies, in fact what it did was place them in random locations of its choice (spread across the various chips in RAID0 like configuration for speed, and shuffled to ensure wear leveling), it then created entries in its virtual table that only it sees and the OS doesn't that say:
Virtual Address (what os sees) -> Real address (where the SSD placed it)
120 -> 200
121 -> 1258
122 -> 2579
123 -> 3479
124 -> 4724
125 -> 5524
126 -> 6244
127 -> 6379
128 -> 7782
129 -> 8239
130 -> 8552
131 -> 9421
132 -> 9613
Those "real addresses" could even shuffle on occasion during normal operation to ensure wear leveling (especially pronounced if you are running low on free space in the drive)... meaning that 120 could then be changed to point to 7821 (and its data is moved to said location). However, that occurrence is not directly related to why defragmenting it is bad.
Defrag works on the assumption that you have a spindle disk that benefits from addresses being in order, and that the address given is the real location.
If, for example, you are writing two files at once, you could end up with addresses (as far as the OS is concerned; that is, real address for spindle, virtual for SSD) such as:
120: file 1
121: file 1
122: file 2
123: file 1
124: file 1
125: file 2
etc... defrag says, uh oh, lets condense it and put all the chunks of file 1 in order, and all the chunks of file 2... so lets move file 1 to addresses 130 through 133. and file 2 to address 134 and 135 to make both sequential... this reduces seek times to read said files in a magnetic media... but in an SSD it just shuffles data around pointlessly because the address the OS has for it is not the actual physical location where it is stored (not to mention that SSDs read faster when in parallel, aka, non sequential), wasting your limited amount of writes (10,000 per cell before it cannot be written to anymore, but can still be read from). And since the old addresses are not trimmed by defrag, it also harms your performance.
This is why defrag is bad for SSDs. that being said, if you defragged an SSD its not the end of the world. Stop doing it and performance will recover naturally via trim (or you can run the intel toolbox or wiper.exe or whatever your drive came with). Also, so you lost a few writes, your drive still has plenty. The damage done to an SSD is not severe, but it does harm it. The benefits are non existent, defrag only benefit spindle disks and doesn't benefits SSDs at all.
As for "light optimization"... it entirely depends on the specific exact method of "optimization" in question. Some are total scams, some give insignificant improvements, some can slightly improve or slightly worsen your performance depending on what you do with it and require you to do your own benchmarks... It is a question of specific case by case
Last edited:
Good explanation of the logical block addressing system.
The question I had about defrag was if there are any filesystem-related benefits to doing so. Assume that a defrag pass simply "shuffles data around" on the SSD without improving or decreasing device-level performance. I understand typical filesystem fragments as comprised of a header, block data, and a pointer to the next fragment (sort of like a LinkedList). Are filesystem requests faster with the file in one contiguous (logical) block as opposed to hundreds or thousands of (logical) fragments, because fewer pointers need to be processed?
The question I had about defrag was if there are any filesystem-related benefits to doing so. Assume that a defrag pass simply "shuffles data around" on the SSD without improving or decreasing device-level performance. I understand typical filesystem fragments as comprised of a header, block data, and a pointer to the next fragment (sort of like a LinkedList). Are filesystem requests faster with the file in one contiguous (logical) block as opposed to hundreds or thousands of (logical) fragments, because fewer pointers need to be processed?
TastesLikeChicken
Lifer
I read the following earlier today and it contains a passage relevant to your question:
http://www.eecs.berkeley.edu/~laurak/iotafs.pdf
Maybe that answers why defragging a SSD really doesn't provide any significant gains?
- Status
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 25K
-
-
-
