Increase the performance and lifespan of your SSD

[John Doe]

Must be drunk, again...

--

Date: Sun, 20 Jan 2013 23:09:26 -0500
From: Yousuf Khan <bbbl67 spammenot.yahoo.com>
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:9.0) Gecko/20111222 Thunderbird/9.0.1
MIME-Version: 1.0
Newsgroups: alt.comp.hardware.pc-homebuilt
Subject: Re: Increase the performance and lifespan of your SSD
References: <kd9dje$i8m$1 dont-email.me> <sg9hf8do5uvfft3hjm4dgr6nu0tg7oj65m 4ax.com> <kdabvj$kuv$3 dont-email.me>
In-Reply-To: <kdabvj$kuv$3 dont-email.me>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
NNTP-Posting-Host: 108.162.170.180
Message-ID: <50fcbf72$1 news.bnb-lp.com>
X-Trace: news.bnb-lp.com 1358741362 108.162.170.180 (20 Jan 2013 23:09:22 -0400)
Organization: Send abuse or DMCA complaints to abuse bnb-lp.com
Lines: 43
X-Authenticated-User: vfa100
X-DMCA-Complaints: Send abuse or DMCA complaints to abuse bnb-lp.com
X-DMCA-Complaints: The subject line should contain only the 4 letters DMCA
Path: eternal-september.org!mx04.eternal-september.org!feeder.eternal-september.org!news.glorb.com!news.bnb-lp.com!not-for-mail
Xref: mx04.eternal-september.org alt.comp.hardware.pc-homebuilt:25924



Oh my god, did you just say that?!? You're obviously one of those who
only has enough information to hang yourself by. Windows, the operating
system, is the one that requires a reboot, Windows is the one that
treats pagefile.sys specially! It's not the SSD that cares, it's Windows!

The SSD itself has a layer of abstraction below the interface that
Windows or any other operating system isn't even aware of. This is
actually no different than the case with hard disks, because they too
used to have layers of abstraction below the OS which they used to
replace bad sectors with good ones, etc. But the SSD takes this several
steps further.

In the meantime, you've used nothing but bellicosity and ill-manners to
insult every person in this thread that tried to help you. All to cover
up your obvious lack of knowledge. What was the purpose of that?


Yes, "as far as you know", which you've demonstrated is not much. It's
Windows that makes it unmoveable, so any utility, like a defragger will
not be allowed to touch it, because Windows will prevent it. However,
Windows has no control over its location below the SATA interface level.
The SSD move it all over the place as it likes, whenever it is
necessary, but it doesn't tell Windows about any of it. That's because
Windows doesn't need to know, and it doesn't care how it's organized at
the physical level. At the logical level, Windows does care, but Windows
deals with the logical level, the SSD deals with it at the physical level.

Yousuf Khan

 

[Yousuf Khan]

It can happen to any SSD even now if you write to it faster than it's
garbage collector can keep up.

Which is also the same with HDD's, it's known as the disk queue length
within the OS. Basically anytime the apps are accessing the drive faster
than the drive can keep up, it'll go over 1.0 and you'll experience
slowdowns.

Yousuf Khan

 

[Loren Pechtel]

Which is also the same with HDD's, it's known as the disk queue length
within the OS. Basically anytime the apps are accessing the drive faster
than the drive can keep up, it'll go over 1.0 and you'll experience
slowdowns.

But HDDs continue to write data at the same rate no matter how big the
queue gets. If you're running it at 100% it matter not one bit if
that's for a second or a week.

A SSD run at 100% for an extended period will empty it's pool of
erased blocks and it's performance will nosedive.

 

[Yousuf Khan]

But HDDs continue to write data at the same rate no matter how big the
queue gets. If you're running it at 100% it matter not one bit if
that's for a second or a week.

The disk queue will keep getting larger. Even though the the disk
becomes 100% busy once it reaches a 1.0 disk queue value, the disk queue
can go upto 2.0, 5.0, 10.0, etc. The maximum speed at which the disk can
deliver data will remain the same throughout, once it's 100% busy,
that's as fast as it will ever go. But programs can keep piling it on,
and the share of time available to access the disk by each program
becomes smaller and smaller as the queue gets larger. So the larger the
queue, the more programs there are trying to access it simultaneously.

A SSD run at 100% for an extended period will empty it's pool of
erased blocks and it's performance will nosedive.

That too will be reflected in the disk queue length measurements. The
slower its maximum performance gets, the larger its disk queue length
will appear to be.

Anyways, it's pretty difficult to actually reach the limit of the SSD's
performance. I calculated from my first SSD that my average writes per
day were 20GB/day! Although I have seen an SSD show maximum 100%
business early on, that turned out to be a fixable bug, and I've not
seen an SSD max out to 100% business since then.

Yousuf Khan

 

[Charlie Hoffpauir]

Oh my god, did you just say that?!? You're obviously one of those who
only has enough information to hang yourself by. Windows, the operating
system, is the one that requires a reboot, Windows is the one that
treats pagefile.sys specially! It's not the SSD that cares, it's Windows!

The SSD itself has a layer of abstraction below the interface that
Windows or any other operating system isn't even aware of. This is
actually no different than the case with hard disks, because they too
used to have layers of abstraction below the OS which they used to
replace bad sectors with good ones, etc. But the SSD takes this several
steps further.

In the meantime, you've used nothing but bellicosity and ill-manners to
insult every person in this thread that tried to help you. All to cover
up your obvious lack of knowledge. What was the purpose of that?


Yes, "as far as you know", which you've demonstrated is not much. It's
Windows that makes it unmoveable, so any utility, like a defragger will
not be allowed to touch it, because Windows will prevent it. However,
Windows has no control over its location below the SATA interface level.
The SSD move it all over the place as it likes, whenever it is
necessary, but it doesn't tell Windows about any of it. That's because
Windows doesn't need to know, and it doesn't care how it's organized at
the physical level. At the logical level, Windows does care, but Windows
deals with the logical level, the SSD deals with it at the physical level.

Yousuf Khan

Just to make it clear, what you quoted at the very first (attributed
to me) was my comment, what followed (unattributed) was the OPs reply
to my comment.

 

[Loren Pechtel]

The disk queue will keep getting larger. Even though the the disk
becomes 100% busy once it reaches a 1.0 disk queue value, the disk queue
can go upto 2.0, 5.0, 10.0, etc. The maximum speed at which the disk can
deliver data will remain the same throughout, once it's 100% busy,
that's as fast as it will ever go. But programs can keep piling it on,
and the share of time available to access the disk by each program
becomes smaller and smaller as the queue gets larger. So the larger the
queue, the more programs there are trying to access it simultaneously.

You're misunderstanding--I'm saying data piled on at the rate the
drive can write it.

That too will be reflected in the disk queue length measurements. The
slower its maximum performance gets, the larger its disk queue length
will appear to be.

Anyways, it's pretty difficult to actually reach the limit of the SSD's
performance. I calculated from my first SSD that my average writes per
day were 20GB/day! Although I have seen an SSD show maximum 100%
business early on, that turned out to be a fixable bug, and I've not
seen an SSD max out to 100% business since then.

In the real world you're not going to run the drive out of blocks. I
was just saying in an extreme case you could and performance would
nosedive.

I just checked the stats on mine, my older drive is 21 months old and
reporting as 99% of life left according to the drive.

 

[DevilsPGD]

In the last episode of <[email protected]>,

But HDDs continue to write data at the same rate no matter how big the
queue gets. If you're running it at 100% it matter not one bit if
that's for a second or a week.

A SSD run at 100% for an extended period will empty it's pool of
erased blocks and it's performance will nosedive.

While true, even once it nosedives, a modern SSD will still be faster
than a typical HDD, so there's no real downside.

 

[John Doe]

In the last episode of <0ekpf8lpfh1sohhhaol1rlvbe3qb91dh4d 4ax.com>,


While true, even once it nosedives, a modern SSD will still be faster
than a typical HDD, so there's no real downside.

I guess that depends on what "modern" means. That certainly is
true if you believe the hype from manufacturers. They were
brainwashing people into thinking that SSDs had a mean time
between failure of over 1 million hours. It was even in their
advertising on Newegg. They probably knew they could hype some
morons into believing it. Everything is great until you have some
experience with it. You can always say that the most recent this
or that is the greatest...

You're just giddy because you finally got one, and now it's the
best thing since baked bread.

--

 

[DevilsPGD]

I guess that depends on what "modern" means. That certainly is
true if you believe the hype from manufacturers. They were
brainwashing people into thinking that SSDs had a mean time
between failure of over 1 million hours. It was even in their
advertising on Newegg. They probably knew they could hype some
morons into believing it. Everything is great until you have some
experience with it. You can always say that the most recent this
or that is the greatest...

Good straw-man attempt, but I wasn't talking about longevity, but
rather, write performance when the erased block pool expires.

With regards to longevity, I've only had a couple early generation ones
die so far, and they were heavily abused (including hosting VMs that
didn't realize they were on SSDs, and so may have defragmented or
otherwise optimized their drives internally)

You're just giddy because you finally got one, and now it's the
best thing since baked bread.

I was a very early adopter and was suitably displeased with how those
drives lived up to their expectations on lifespan. But performance was
still significantly better than rotational drives, making them
worthwhile.

Having used modern generations of SSDs as they've come out, I've been
very impressed with how well they handle, especially in terms of keeping
up with write performance, and I really can't imagine willingly building
a system that isn't based on SSD for OS, application and primary storage
these days, unless budget is a factor above all else.

 

[Yousuf Khan]

In the last episode of <[email protected]>,


While true, even once it nosedives, a modern SSD will still be faster
than a typical HDD, so there's no real downside.

I do see Loren's point though. Erasing a flash cell is extremely
time-consuming. I don't remember exactly what it is, but I had read at
one time that it may be measured in the milliseconds, rather than the
nanoseconds (i.e. million fold increase). That's why it's usually done
in the background, and new cells are substituted in during writes.

Yousuf Khan