This is relevant - "Why solid-state disks are winning the argument".

In article , RJH
wrote:

This is definitely an issue if you have lots of writes. I don't know
how many "lots" is; it will depend on the type of disk, and this will
improve with time. However, modern SSDs can burn out in half an hour on
real-world systems with lots of writes. For write-once or write-few, of
course, they are OK.


Do you have any evidence of this 'half hour real world burnout' thesis?

there is none. ssds last a very long time even when being hammered and
certainly under normal use.

In article ,
Whisky-dave wrote:

... Unless your workload is very specifically single source, massive
capture, then you should be running SSDs. Even if you are not running
pure SSD, the case for tiered or hybrid storage makes itself.

SSDs are faster. They have way lower latency. They consume less power.
They take up less space.

True but the downside is they cost quite a bit more and aren't readily
avaible in the sizes some are used to.

they're available in up to 1 terabyte currently.

Which is also why they aren't yet used
in servers or for most as backup discs.

they are used in servers in some cases, but generally, the advantages
are mostly lost.

in a computer, however, the advantages are significant.

In article , Mayayana
wrote:

| I like to actually know what's going on before I say I know what's going
on.
|

A quick search found this, which seems to be a
relatively unbiased comparison between "fusion"
and caching:

http://www.pcmag.com/article2/0,2817,2412726,00.asp

It doesn't seem to be "sheer hokum".

correct. it's definitely not.

On the other
hand, do you want to spend hundreds of dollars extra
in order to have your most bloated software load
faster?

many people do.

Speed has been a dubious marketing scheme
ever since the late 90s.

nonsense.

I remember when Intel would
announce each new CPU with great fanfare. And 400 MHz
really was a big improvement over 300 MHz. But at some
point speed ceased to be a real issue. Nevertheless,
computer magazines would still describe the new 1,200 MHz
as "blazingly fast" while your "old" 1,000 MHz suddenly
became "good enough for email and web browsing". A
few months later the 1,200 was barely usable and the
1,350 was "screamingly fast".

what you're missing is that faster systems enable more capable
software. it also enables real-time editing of images, video, etc.

when i first started doing video, i had to run a render overnight, and
that was typically for a small portion of the video, not the entire
thing.

now i can edit in real time, and with far bigger assets.

Another version of the "fusion" approach, which has been
done for many years, is pre-loading. Bloatware such as
Libre Office, MS Office and, I think, Firefox, offer preloading
as a way to make their software seem fast. It's a bloated
mess but it responds instantly because the needed libraries
are already in RAM. Personally I don't find it stressful to wait
a few seconds once in a while. For most things my computer
(XP) responds instantly....and there's nothing faster than that.

preloading has little to do with fusion.

In my experience, a reasonably clean system is
fast and responsive by itself. Most things are instant
for me, using a standard hard disk. Much of the reason
for that is that most things don't have to be loaded
from disk in the first place. Most things are in RAM. (Which
is why most software loads faster the second time than
the first.) Some is pre-loaded. System files stay loaded.
So in many cases there isn't much that actually needs to
be loaded from disk. You might benefit a tiny bit from
having your 35 MB wedding photos in fast "fusion" storage,
but "fusion" won't know to do that until you've already
worked on the photos, so that's not going to be a relevant
effect.

the benefit of ssd is much, much, much more than 'a tiny bit'.

A few more recent programs I have are bloated --
Libre Office and PSP16, most notably. But those are
not things that I open and close a lot. Unfortunately,
a lot of current software is simply overproduced in
an attempt to always have new "features" with each
release.

customers want the features.

I find that PSP5 does 90% of what I want to
do and loads instantly.

photoshop loads instantly (under 1 second).

PSP16 takes about 100 MB RAM
just to sit there, takes several seconds to load, and opens
with a ridiculous and superfluous 3-part UI that includes
a file explorer window and another separate window with
a "quickie adjustment" selection of the normal editor tools.
It's also unstable.

why use something that's unstable? is it even supported on your
antiquated system?

Spending $250 for something like "fusion"
might help that to load, but load time is really only part
of the problem.

and it solves that part of the problem.

All of which is to say, speed is nice but one has to look
at it in context. It's not just a quantity where more is better.
The same issue happens with browsers. People want to know
what the fastest browser is. That's easy: It's the one that's
loading from the fastest website server and/or the smallest
webpage.

completely wrong. javascript engines and html rendering has greatly
improved in recent browsers.

In article , Martin Brown
wrote:

as the user uses the computer, commonly used files are moved to the ssd
and infrequently used files are moved to the hard drive, without the
user needing to do anything other than use the computer normally.

Would you care to explain how Apple's Fusion Drive differs from the SSD
cache technology that Intel introduced with SandyBridge Z86 chipset?

http://www.anandtech.com/show/4329/i...ching-review/2

Now called Intel "Smart Response Technology" - dreadful name...

as has been explained multiple times already, fusion is not a cache.

it's one logical volume that's actually tiered storage where files are
intelligently moved between ssd and the hard drive.

Fusion Drive was clearly the product of Trekkie inspired marketing men!

no.

On 2014-11-20, nospam wrote:
In article , Oregonian Haruspex
wrote:

I see no benefit to those new-fangled "hybrid" drives (really just a HD
with a bigger, smarter cache) because the rust will still be spinning
all the time.

hybrid drives are actually not that great and only slightly better than
a normal hd. it's basically a big cache for recently used files, which
may not be the ones that matter.

Yeah I know. Caching is a gamble anyway but especially when the OS and
the drive aren't talking to each other about it.

they don't need to. the drive cache holds recently accessed blocks with
the assumption they might be needed again.

As far as I understand it, the drive reads ahead and stores blocks in
the same sector, assuming that the OS might need them. Sequential
read-ahead. The OS is what stores the recently accessed blocks.

drives generally cache the entire file when one block is accessed
because there's an extremely high likelihood you'll be accessing the
entire file.

No they don't, because drives do not know the physical structure of the
filesystem nor what blocks in which a certain file is located! This is
the job of the OS. The drive just fetches blocks, and hopes that it
can fetch the right ones.

To my knowledge there are no drives existing that know about the
filesystems they contain.

the os tells the drive what to read and it's cached.

What blocks to read, not files. Those are only known to the OS.


Where is the technical documentation for Fusion? How do you know how it
works?

because i've read quite a bit about it.

here's apple's tech note:

Sorry, Apple notes don't count.
They're from Apple, remember?



--
When in doubt, use brute force.
-- Ken Thompson

| announce each new CPU with great fanfare. And 400 MHz
| really was a big improvement over 300 MHz. But at some
| point speed ceased to be a real issue. Nevertheless,
| computer magazines would still describe the new 1,200 MHz
| as "blazingly fast" while your "old" 1,000 MHz suddenly
| became "good enough for email and web browsing". A
| few months later the 1,200 was barely usable and the
| 1,350 was "screamingly fast".
|
| what you're missing is that faster systems enable more capable
| software. it also enables real-time editing of images, video, etc.
| when i first started doing video, i had to run a render overnight, and
| that was typically for a small portion of the video, not the entire
| thing.
| now i can edit in real time, and with far bigger assets.

And now I'm guessing your old dual CPU 6600MHz is
"good enough for email and web browsing"?

| Another version of the "fusion" approach, which has been
| done for many years, is pre-loading. Bloatware such as
| Libre Office, MS Office and, I think, Firefox, offer preloading
| as a way to make their software seem fast. It's a bloated
| mess but it responds instantly because the needed libraries
| are already in RAM. Personally I don't find it stressful to wait
| a few seconds once in a while. For most things my computer
| (XP) responds instantly....and there's nothing faster than that.
|
| preloading has little to do with fusion.

They're technically not the same thing, of course.
But both are methods to speed up access. Fusion
puts commonly used programs on an SSD. Preloading
loads the bulk those programs into RAM. (Which
I would expect to be a lot faster.) If most of the
needed libraries are already in RAM then the program
opens instantly. That's why MS Word and Internet
Explorer seem so fast. Windows loads them at startup.
So the theory of fusion makes sense. The value of it
is another story.

| photoshop loads instantly (under 1 second).

And much of it is probably preloaded, or maybe it's
giving you a GUI while it loads the rest. It may simply
be far less bloated than PSP, but I very much doubt
that, coming from the same company that's managed
to bloat a simple PDF reader into a major OS component.

| All of which is to say, speed is nice but one has to look
| at it in context. It's not just a quantity where more is better.
| The same issue happens with browsers. People want to know
| what the fastest browser is. That's easy: It's the one that's
| loading from the fastest website server and/or the smallest
| webpage.
|
| completely wrong. javascript engines and html rendering has greatly
| improved in recent browsers.

Yes, that's true. And it makes very little difference.
Your logic is a perfect example of how people get
affected by these trumped up "speed races" -- between
OSs, PCs, browsers, or whatever. Browser speed is
good marketing for browser companies and good filler
for the lapdog tech media, always short on relevant
content, but the issue is not supported by common
sense.

The current thinking among webmasters is that pages
need to load within 250ms or people will get impatient.
That's not easy to do when they're foisting 1/4 MB of
javascript libraries on people, in addition to images
and highly bloated HTML. Many commercial pages today
are actually not-so-small software programs.

So say there's a webpage that's loading 250 KB of
script and maybe FF processes that script in 120 ms,
Chrome in 110 ms and IE in 130 ms. You won't see
any difference. But what can make a big difference
is the crowding of the network and the speed of
servers. Many sites are loading files from up to a
dozen servers. If only one of them is slow or overwhelmed
the page will load more slowly. If you want
fast then try disabling script and blocking ads.

Back to the original point: Someone asks what's
the fastest browser. Let's say it's Chrome. The
tests I've seen depend on what webpage is loaded.
But let's just say Chrome is 10% faster overall. That
means that under the best conditions -- with you
having a fast connection, the network not being
bogged down, and the target servers being highly
efficient -- you *might* get the webpage displayed
25 ms faster in Chrome. That's not even discernable.
250 ms is pretty much the threshold of being
perceived as instant.

And then of course there's the elephant in the
room: What's the big hurry? Why do we need
webpages in 250ms? So we can rush to the next
one faster? With all the factors that go into
choosing a browser, speed should be very far
down the list. (By the way, Mario Andretti, I hope
you've shut off the Loonie Tunes vacuum cleaner
animation on your "dashboard" icons. That probably
loses 300 ms alone, which could cause your
Photoshop load to be reduced to an interminable
crawl of 1.3 seconds! At that speed you may as
well just skip it and and draw the damn picture.

On Thu, 20 Nov 2014 12:55:23 -0500, nospam
wrote:

In article , RJH
wrote:

This is definitely an issue if you have lots of writes. I don't know
how many "lots" is; it will depend on the type of disk, and this will
improve with time. However, modern SSDs can burn out in half an hour on
real-world systems with lots of writes. For write-once or write-few, of
course, they are OK.


Do you have any evidence of this 'half hour real world burnout' thesis?

there is none. ssds last a very long time even when being hammered and
certainly under normal use.

There is no sign of failure after half an hour but the article with
which I started this thread
http://www.theregister.co.uk/2014/11/07/storage_ssds/ says:

"On the whole, however, the statistics tend to confirm that after an
initial rough patch, SSDs have about the same reliability as
traditional magnetic disks. There are, of course, exceptions.

If you sit there and hammer a consumer SSD with high transactions
data loads all day long you will burn it out well before the
warranty expires.

Similarly, SSDs are a terrible place to do a bunch of log file
writes to; eleventy squillion crappy little sub-K writes will burn
out the SSDs in no time."
--

Regards,

Eric Stevens

In article , Rikishi42
wrote:

Where is the technical documentation for Fusion? How do you know how it
works?

because i've read quite a bit about it.

here's apple's tech note:

Sorry, Apple notes don't count.
They're from Apple, remember?

of course it counts. apple is who designed it. why would they lie about
it?

however, there is not extensive documentation about it but there is a
lot of third party testing, which confirm what apple said.

In article , Mayayana
wrote:

| announce each new CPU with great fanfare. And 400 MHz
| really was a big improvement over 300 MHz. But at some
| point speed ceased to be a real issue. Nevertheless,
| computer magazines would still describe the new 1,200 MHz
| as "blazingly fast" while your "old" 1,000 MHz suddenly
| became "good enough for email and web browsing". A
| few months later the 1,200 was barely usable and the
| 1,350 was "screamingly fast".
|
| what you're missing is that faster systems enable more capable
| software. it also enables real-time editing of images, video, etc.
| when i first started doing video, i had to run a render overnight, and
| that was typically for a small portion of the video, not the entire
| thing.
| now i can edit in real time, and with far bigger assets.

And now I'm guessing your old dual CPU 6600MHz is
"good enough for email and web browsing"?

6600 mhz?

| Another version of the "fusion" approach, which has been
| done for many years, is pre-loading. Bloatware such as
| Libre Office, MS Office and, I think, Firefox, offer preloading
| as a way to make their software seem fast. It's a bloated
| mess but it responds instantly because the needed libraries
| are already in RAM. Personally I don't find it stressful to wait
| a few seconds once in a while. For most things my computer
| (XP) responds instantly....and there's nothing faster than that.
|
| preloading has little to do with fusion.

They're technically not the same thing, of course.
But both are methods to speed up access.

that's about the extent of the similarity.

Fusion
puts commonly used programs on an SSD. Preloading
loads the bulk those programs into RAM. (Which
I would expect to be a lot faster.)

two totally different things, and one does not preclude the other
either. you could have a fusion drive *and* have the app cached, which
is what happens on a second launch.

If most of the
needed libraries are already in RAM then the program
opens instantly. That's why MS Word and Internet
Explorer seem so fast. Windows loads them at startup.

and if you aren't using those apps, the memory is wasted holding
libraries you aren't going to be using, making other apps slower.

So the theory of fusion makes sense. The value of it
is another story.

it's very valuable, and a very good compromise between fast but pricy
ssd versus affordable but slow hard drives.

| photoshop loads instantly (under 1 second).

And much of it is probably preloaded, or maybe it's
giving you a GUI while it loads the rest. It may simply
be far less bloated than PSP, but I very much doubt
that, coming from the same company that's managed
to bloat a simple PDF reader into a major OS component.

wrong on all counts. none of it is preloaded.

with photoshop on ssd, it loads in under 1 second from a cold boot
(nothing cached or preloaded). if you think apple is going to preload
adobe libraries, you're even more out of touch than i thought.

In article , Eric Stevens
wrote:

This is definitely an issue if you have lots of writes. I don't know
how many "lots" is; it will depend on the type of disk, and this will
improve with time. However, modern SSDs can burn out in half an hour on
real-world systems with lots of writes. For write-once or write-few, of
course, they are OK.

Do you have any evidence of this 'half hour real world burnout' thesis?

there is none. ssds last a very long time even when being hammered and
certainly under normal use.

There is no sign of failure after half an hour but the article with
which I started this thread
http://www.theregister.co.uk/2014/11/07/storage_ssds/ says:

"On the whole, however, the statistics tend to confirm that after an
initial rough patch, SSDs have about the same reliability as
traditional magnetic disks. There are, of course, exceptions.

so many exceptions that the claim is totally wrong.

the fact is that ssds are more reliable than hard drives.

some older ssds had issues, not because of flash exhaustion, but
because of firmware bugs in the controller.

when disk drives first came out many decades ago, they too had growing
pains, and even now, disk drives occasionally suffer from firmware
bugs, such as the one which affected seagate a couple of years ago.

If you sit there and hammer a consumer SSD with high transactions
data loads all day long you will burn it out well before the
warranty expires.

nope.

Similarly, SSDs are a terrible place to do a bunch of log file
writes to; eleventy squillion crappy little sub-K writes will burn
out the SSDs in no time."

nope.

see the link i posted about longetivity, where ssds were hammered and
sustained nearly a petabyte of writes. a petabyte is 1000 terabytes or
1 million gigabytes.

were you expecting to write log files that big? just how long do you
think that would take in normal use anyway?