• Welcome to Tux Reports: Where Penguins Fly. We hope you find the topics varied, interesting, and worthy of your time. Please become a member and join in the discussions.

How I Got Full Windows XP Installed Under 2GB!

B

BillW50

Flightless Bird
In news:hiqnvm$kf2$1@reader1.panix.com,
the wharf rat typed on Fri, 15 Jan 2010 21:50:46 +0000 (UTC):
> In article <hiqmeq$mc1$1@news.eternal-september.org>,
> BillW50 <BillW50@aol.kom> wrote:
>>
>> True, but that takes 100,000 rewrites per cell to kill it. Writing
>> 100MB
>
> Not really. Given a sufficiently large population of cells one
> cell will fail for every certain number of writes. The chances of any
> particular cell failing on any particular write are about 1 in 100000,
> but the chances of any one of the cells in the array failing are much
> larger. And they're still dependent on the number of writes, so SOME
> cells will INEVITABLY fail.

So? In time a hard drive will suffer a bad sector or two. They are
marked as bad and life moves on. The same on a flash drive. No big deal.

--
Bill
Asus EEE PC 702G8 ~ 2GB RAM ~ 16GB-SDHC
Windows XP SP2
 
T

the wharf rat

Flightless Bird
In article <hiqp5i$8nk$1@news.eternal-september.org>,
BillW50 <BillW50@aol.kom> wrote:
>
>So? In time a hard drive will suffer a bad sector or two. They are
>marked as bad and life moves on. The same on a flash drive. No big deal.
>

The difference is related to capacity. A bad block on a terabyte
hard drive is less signifcant than a bad block on a 32MB flash device.
 
B

BillW50

Flightless Bird
In news:hisivf$ean$1@reader1.panix.com,
the wharf rat typed on Sat, 16 Jan 2010 14:37:35 +0000 (UTC):
> In article <hiqp5i$8nk$1@news.eternal-september.org>,
> BillW50 <BillW50@aol.kom> wrote:
>>
>> So? In time a hard drive will suffer a bad sector or two. They are
>> marked as bad and life moves on. The same on a flash drive. No big
>> deal.
>
> The difference is related to capacity. A bad block on a terabyte
> hard drive is less signifcant than a bad block on a 32MB flash device.

Boy are you digging deep here. 32MB flash drives? What good is a fully
functional 32MB flash drive good for anyway? Yeah you can run DOS from
it, but not much good for anything else. I have 12 flash drives. Eight
of them have 1GB or more each.

--
Bill
Asus EEE PC 702G8 ~ 2GB RAM ~ 16GB-SDHC
Windows XP SP2
 
T

the wharf rat

Flightless Bird
In article <hism72$prb$1@news.eternal-september.org>,
BillW50 <BillW50@aol.kom> wrote:
>>
>> The difference is related to capacity. A bad block on a terabyte
>> hard drive is less signifcant than a bad block on a 32MB flash device.
>
>Boy are you digging deep here. 32MB flash drives? What good is a fully
>functional 32MB flash drive good for anyway? Yeah you can run DOS from
>it, but not much good for anything else. I have 12 flash drives. Eight
>of them have 1GB or more each.

How many blocks are on a terabyte disc? How many blocks are on a 1GB
flash drive? Is a bad block on a 1Gb flash drive a relative loss of 1000
times more storage capacity? Or is that also strictly a Linux problem?
 
B

BillW50

Flightless Bird
In news:hisske$r25$2@reader1.panix.com,
the wharf rat typed on Sat, 16 Jan 2010 17:22:22 +0000 (UTC):
> In article <hism72$prb$1@news.eternal-september.org>,
> BillW50 <BillW50@aol.kom> wrote:
>>>
>>> The difference is related to capacity. A bad block on a terabyte
>>> hard drive is less signifcant than a bad block on a 32MB flash
>>> device.
>>
>> Boy are you digging deep here. 32MB flash drives? What good is a
>> fully functional 32MB flash drive good for anyway? Yeah you can run
>> DOS from it, but not much good for anything else. I have 12 flash
>> drives. Eight of them have 1GB or more each.
>
> How many blocks are on a terabyte disc? How many blocks are on a 1GB
> flash drive? Is a bad block on a 1Gb flash drive a relative loss of
> 1000 times more storage capacity? Or is that also strictly a Linux
> problem?

I wouldn't know without checking. Here I have my three 16GB flash drives
right here. Here is what CHKDSK reports for the three.

1) 15,880,752 KB total disk space. 1,985,094 blocks.
2) 15,672,112 KB total disk space. 1,959,014 blocks.
3) 15,664,088 KB total disk space. 1,958,011 blocks.

Baring anything funny going on, like a secret hidden partition on them
or anything. They all should be the same size. I use #1 the most lately
and it had as far as I know, its first bad block which I fixed with
CHKDSK /r (after 8 months of everyday use). So I am assuming the other
two have suffered more than one bad block. Each block is 8K. #2 and #3
have been used for two years now. #1 and #2 are bootable.

So there you go. From the worst to the best there is like a 220kb
(27,083 blocks) difference. Do I miss them? Nope! Am I worried? Nope!
Should I be? I don't see any reason why I should. As I see the very same
thing on comparing the same size hard drives as well.

Remember all hard drives have bad blocks on them even brand new! They
are there when manufactured. I assume the very same thing is true of
flash drives as well. The reason why you can't see the number of bad
blocks like you could decades ago, is because the hardware hides them.
They have done this because they were getting too many returns. So the
only way to know if they are there is through the total disk space. Well
for the average user anyway. ;-)

--
Bill
Asus EEE PC 702G8 ~ 2GB RAM ~ 16GB-SDHC
Windows XP SP2
 
M

~misfit~

Flightless Bird
Somewhere on teh intarwebs BillW50 wrote:
> In news:hiqjra$o0b$2@reader1.panix.com,
> the wharf rat typed on Fri, 15 Jan 2010 20:40:10 +0000 (UTC):
>> In article <hiq7mg$dcp$1@news.eternal-september.org>,
>> BillW50 <BillW50@aol.kom> wrote:
>>>
>>> I average 100MB to 150MB writes per day of writes to SLC flash
>>> drives. At that rate, it would take 4,000 years to hit 100,000
>>> writes per cell.
>>
>> But writes aren't evenly distributed across cells. The exact pattern
>> depends on free space available, the balancing algorithms, and the
>> patterns of the load. Also, MTBF doesn't mean that ALL those cells
>> will last that long. It's a statistic; you'll see plenty of failures
>> before that magic number is reached.
>
> No problem. As that is what wear leveling takes care of. I have been
> using flash drives (SSD) for running Windows for two years now and no
> problems yet.

So you've been using the term 'flash drive' to refer to an SSD from the
start of this thread?

That's pretty confusing as they're two different things with fairly well
established nomenclature.
--
Shaun.

"Give a man a fire and he's warm for the day. But set fire to him and he's
warm for the rest of his life." Terry Pratchet, 'Jingo'.
 
B

BillW50

Flightless Bird
In news:hitld1$hn0$1@news.eternal-september.org,
~misfit~ typed on Sun, 17 Jan 2010 13:25:01 +1300:
> Somewhere on teh intarwebs BillW50 wrote:
>> In news:hiqjra$o0b$2@reader1.panix.com,
>> the wharf rat typed on Fri, 15 Jan 2010 20:40:10 +0000 (UTC):
>>> In article <hiq7mg$dcp$1@news.eternal-september.org>,
>>> BillW50 <BillW50@aol.kom> wrote:
>>>>
>>>> I average 100MB to 150MB writes per day of writes to SLC flash
>>>> drives. At that rate, it would take 4,000 years to hit 100,000
>>>> writes per cell.
>>>
>>> But writes aren't evenly distributed across cells. The exact
>>> pattern depends on free space available, the balancing algorithms,
>>> and the patterns of the load. Also, MTBF doesn't mean that ALL
>>> those cells will last that long. It's a statistic; you'll see
>>> plenty of failures before that magic number is reached.
>>
>> No problem. As that is what wear leveling takes care of. I have been
>> using flash drives (SSD) for running Windows for two years now and no
>> problems yet.
>
> So you've been using the term 'flash drive' to refer to an SSD from
> the start of this thread?
>
> That's pretty confusing as they're two different things with fairly
> well established nomenclature.

How do you figure? From the outside they might look different. But from
the inside, they are all the same. As they use the same technology with
the same SLC / MLC chips for the drive. And the same rules of longevity
and wear leveling also apply. Speaking of which, have you ever thought
about how much writing is evolved for 100,000 writes per cell? For a 4GB
for example, you would have to write 400TB worth to reach the expected
lifespan. That is a lifetime or more worth the way I use computers.

And the problem that I had solved to get the full Windows XP to fit
inside of less than 2GB SSD, was to move the Program Files folder over
to a SD flash drive. Sounds easy, except Windows sees the SD flash as a
removable drive. Plus all of the many registry entries are thinking that
programs are in the C:/Program Files folder. And many programs that you
wish to later install will refuse to install on a removable drive
anyway.

All of these problems disappear, if you mount the SD flash in the
Program Files folder. As the Windows registry is happy, as nothing as
far as it is concern has changed. And Windows and the applications are
happy, as they don't see the SD drive as a removable drive anymore. And
unlike previous hacks to pull this off, which requires registry hacking
and slower I/O performance. This one even a non-computer geek can pull
it off very easily.

Thus I believe this information is a gold mine for those with 2GB and
4GB SSD running Windows XP. As free space can get very limited really
fast if you are not careful. And this allows for a lot of more breathing
room. ;-)

--
Bill
Gateway M465e ('06 era) - Windows XP SP3
 
T

the wharf rat

Flightless Bird
In article <hit182$v58$1@news.eternal-september.org>,
BillW50 <BillW50@aol.kom> wrote:
>
>I wouldn't know without checking. Here I have my three 16GB flash drives
>right here. Here is what CHKDSK reports for the three.
>
>1) 15,880,752 KB total disk space. 1,985,094 blocks.
>2) 15,672,112 KB total disk space. 1,959,014 blocks.
>3) 15,664,088 KB total disk space. 1,958,011 blocks.

PHYSICAL blocks. Not logical blocks... Sigh.


Anyway, the reason wear matters is because solid state storage
devices have fewer physical blocks. Losing one matters a lot more.
 
B

BillW50

Flightless Bird
In news:hivnqq$c8b$2@reader1.panix.com,
the wharf rat typed on Sun, 17 Jan 2010 19:18:50 +0000 (UTC):
> In article <hit182$v58$1@news.eternal-september.org>,
> BillW50 <BillW50@aol.kom> wrote:
>>
>> I wouldn't know without checking. Here I have my three 16GB flash
>> drives right here. Here is what CHKDSK reports for the three.
>>
>> 1) 15,880,752 KB total disk space. 1,985,094 blocks.
>> 2) 15,672,112 KB total disk space. 1,959,014 blocks.
>> 3) 15,664,088 KB total disk space. 1,958,011 blocks.
>
> PHYSICAL blocks. Not logical blocks... Sigh.

Well a Google search for Adata 16GB SDHC data sheet should pull the
information up.

> Anyway, the reason wear matters is because solid state storage
> devices have fewer physical blocks. Losing one matters a lot more.

It doesn't matter to me. What matters to me is how much of the total
space disappears over its lifetime. After all, that is the real
bottom-line, right? And the 12 flash drives I have hasn't even dropped
1% of the original capacity yet. It isn't even close.

And let's say after 20 years of heavy use the capacity drops to a
whopping 10% less capacity than they were when brand new. Then I still
wouldn't care. Since 20 years from now, 16GB flash drives will be a dime
a dozen by then. ;-)

--
Bill
Gateway M465e ('06 era) - Windows XP SP3
 
T

the wharf rat

Flightless Bird
In article <hivp9a$4po$1@news.eternal-september.org>,
BillW50 <BillW50@aol.kom> wrote:
>
>bottom-line, right? And the 12 flash drives I have hasn't even dropped
>1% of the original capacity yet. It isn't even close.
>

How many of those are in service as virtual memory or scratch
fiel space for an operating system installation?

>20 years from now, 16GB flash drives will be a dime a dozen by then. ;-)

Sure. And in 2001 we'll all be driving flying cars.
 
M

~misfit~

Flightless Bird
Somewhere on teh intarwebs BillW50 wrote:
> In news:hitld1$hn0$1@news.eternal-september.org,
> ~misfit~ typed on Sun, 17 Jan 2010 13:25:01 +1300:
>> Somewhere on teh intarwebs BillW50 wrote:
>>> In news:hiqjra$o0b$2@reader1.panix.com,
>>> the wharf rat typed on Fri, 15 Jan 2010 20:40:10 +0000 (UTC):
>>>> In article <hiq7mg$dcp$1@news.eternal-september.org>,
>>>> BillW50 <BillW50@aol.kom> wrote:
>>>>>
>>>>> I average 100MB to 150MB writes per day of writes to SLC flash
>>>>> drives. At that rate, it would take 4,000 years to hit 100,000
>>>>> writes per cell.
>>>>
>>>> But writes aren't evenly distributed across cells. The exact
>>>> pattern depends on free space available, the balancing algorithms,
>>>> and the patterns of the load. Also, MTBF doesn't mean that ALL
>>>> those cells will last that long. It's a statistic; you'll see
>>>> plenty of failures before that magic number is reached.
>>>
>>> No problem. As that is what wear leveling takes care of. I have been
>>> using flash drives (SSD) for running Windows for two years now and
>>> no problems yet.
>>
>> So you've been using the term 'flash drive' to refer to an SSD from
>> the start of this thread?
>>
>> That's pretty confusing as they're two different things with fairly
>> well established nomenclature.
>
> How do you figure?

<sigh>

Google "Flash drive".

Then Google "SSD".

I don't know why you feel the urge to write all this other stuff below.
--
Shaun.

"Give a man a fire and he's warm for the day. But set fire to him and he's
warm for the rest of his life." Terry Pratchet, 'Jingo'.

> From the outside they might look different. But
> from the inside, they are all the same. As they use the same
> technology with the same SLC / MLC chips for the drive. And the same
> rules of longevity and wear leveling also apply. Speaking of which,
> have you ever thought about how much writing is evolved for 100,000
> writes per cell? For a 4GB for example, you would have to write 400TB
> worth to reach the expected lifespan. That is a lifetime or more
> worth the way I use computers.
> And the problem that I had solved to get the full Windows XP to fit
> inside of less than 2GB SSD, was to move the Program Files folder over
> to a SD flash drive. Sounds easy, except Windows sees the SD flash as
> a removable drive. Plus all of the many registry entries are thinking
> that programs are in the C:/Program Files folder. And many programs
> that you wish to later install will refuse to install on a removable
> drive anyway.
>
> All of these problems disappear, if you mount the SD flash in the
> Program Files folder. As the Windows registry is happy, as nothing as
> far as it is concern has changed. And Windows and the applications are
> happy, as they don't see the SD drive as a removable drive anymore.
> And unlike previous hacks to pull this off, which requires registry
> hacking and slower I/O performance. This one even a non-computer geek
> can pull it off very easily.
>
> Thus I believe this information is a gold mine for those with 2GB and
> 4GB SSD running Windows XP. As free space can get very limited really
> fast if you are not careful. And this allows for a lot of more
> breathing room. ;-)
 
M

M.I.5¾

Flightless Bird
"Charlie Hoffpauir" <invalid@invalid.com> wrote in message
news:a5avk5p0tsmiifho5d4i1qduc6fgkhiint@4ax.com...
> On Thu, 14 Jan 2010 14:52:02 -0800 (PST), "flamer
> die.spam@hotmail.com" <die.spam@hotmail.com> wrote:
>
>>flash drives are NOT made for this application. constant read/writing
>>to the drive will kill it very quickly as they are limited to the
>>number of re-write cycles and have a limited number of years data
>>retention.
>>
>>Flamer.
>
> How quickly is "quickly"? A check with wikipedia says modern flash
> drives are rated for 1 million write/erase cycles per cell. I don't
> know how he uses his drive, but that far exceeds my expected remaining
> lifetime (I'm 70). Now data retention is a limitation, with only 10
> years.... I do expect to be around for longer than that. But I doubt
> I'd still be using the same netbook for that long.

You trust Wikipedia?

Modern FLASH memory can be rated for 1 million cycles, but only if they have
a wear levelling controller. The actual memory unit itself is usually rated
for a average of a few thousand write/erase cycles, with one block having a
substantially longer life to act as the housekeeping block. The likely
level of writing and erasing that you are proposing will use the erase/write
cycles up of the shortest lived cells relatively quickly. The situation is
exacerbated because flash memory is not bit (or even byte) writeable.
Changing so much as one bit requires the erasure and rewrite of an entire
block.
 
M

M.I.5¾

Flightless Bird
"BillW50" <BillW50@aol.kom> wrote in message
news:hiq7mg$dcp$1@news.eternal-september.org...
> In news:hiof7h$5a1$1@news.eternal-september.org,
> mike typed on Thu, 14 Jan 2010 17:08:19 -0800:
>> Charlie Hoffpauir wrote:
>>> On Thu, 14 Jan 2010 14:52:02 -0800 (PST), "flamer
>>> die.spam@hotmail.com" <die.spam@hotmail.com> wrote:
>>>
>>>> flash drives are NOT made for this application. constant
>>>> read/writing to the drive will kill it very quickly as they are
>>>> limited to the number of re-write cycles and have a limited number
>>>> of years data retention.
>>>>
>>>> Flamer.
>>>
>>> How quickly is "quickly"? A check with wikipedia says modern flash
>>> drives are rated for 1 million write/erase cycles per cell. I don't
>>> know how he uses his drive, but that far exceeds my expected
>>> remaining lifetime (I'm 70). Now data retention is a limitation,
>>> with only 10 years.... I do expect to be around for longer than
>>> that. But I doubt I'd still be using the same netbook for that long.
>>
>> This is a very complex issue. Most of what's going on is trade secret
>> and unknowable by mere mortals.
>> Reading is not a problem. Writing is the problem.
>>
>> I thought that moving the program files folder would be ok, because
>> it's all read stuff and the volatile stuff would be stored in the
>> registry on the HD. But when I checked it out, I found over a hundred
>> .ini files that store volatile stuff like most recently used files.
>>
>> If it gets written once when you close the program, you're probably
>> safe. If you're copying 10,000 files and the MRU entry gets updated
>> for every file, you've got a problem.
>>
>> Some programs may also store temporary files or caches in their
>> install directory.
>>
>> Depending on the wear leveling algorithm, if any, it may matter a LOT
>> whether the flash drive is mostly empty of mostly full.
>>
>> You'll have to google it, but there are some tools designed for
>> embedded windows that dramatically reduce writes to the drive. Not
>> at all clear if they would work the way you've set up your system. There
>> was a long thread a few months ago about longevity of SSD. Think it was
>> in comp.sys.ibm.pc.hardware.storage.
>>
>> A million sounds like a big number until you start talking about how
>> fast a computer can run through it.
>
> Well for starters, I use 100,000 per cell write cycles for my SLC flash
> drive figures, which are more conservative. And if I wrote 4GB per hour
> for 24 hours per day on a 4GB flash drive, it would take 11 years to wear
> it out.
>
> MTBF for a SLC flash is 227 years, 7 times longer than hard drives.
>

MTBF does not include wear out due to use. MTBF refers to the expected life
of the memory if it is read only

> I average 100MB to 150MB writes per day of writes to SLC flash drives. At
> that rate, it would take 4,000 years to hit 100,000 writes per cell.
>

A lot depends on how they are written. If there is a wear levelling
controller (likely), then small writes will make the memory last longer than
it is really entitled to.

But the reality is that all the figures quoted are average figures. The
deviation from the mean is substantial indeed for FLASH memory, and
individual examples can have extremely short lives. And I am talking memory
from reputable suppliers not shonky made in China stuff.

> So unless you are buying inferior MLC flash drives which the real cheap
> ones are only good for 5,000 writes per cell, there is little to worry
> about. But even if you do, they are so cheap, so what if you replace them
> once a year or so?
>

Not a problem if you don't need the data that you can't recover from them
once they fail.
 
M

M.I.5¾

Flightless Bird
"M.I.5¾" <no.one@no.where.NO_SPAM.co.uk> wrote in message
news:4b544656$1_1@glkas0286.greenlnk.net...
>
> "BillW50" <BillW50@aol.kom> wrote in message
> news:hiq7mg$dcp$1@news.eternal-september.org...
>> In news:hiof7h$5a1$1@news.eternal-september.org,
>> mike typed on Thu, 14 Jan 2010 17:08:19 -0800:
>>> Charlie Hoffpauir wrote:
>>>> On Thu, 14 Jan 2010 14:52:02 -0800 (PST), "flamer
>>>> die.spam@hotmail.com" <die.spam@hotmail.com> wrote:
>>>>
>>>>> flash drives are NOT made for this application. constant
>>>>> read/writing to the drive will kill it very quickly as they are
>>>>> limited to the number of re-write cycles and have a limited number
>>>>> of years data retention.
>>>>>
>>>>> Flamer.
>>>>
>>>> How quickly is "quickly"? A check with wikipedia says modern flash
>>>> drives are rated for 1 million write/erase cycles per cell. I don't
>>>> know how he uses his drive, but that far exceeds my expected
>>>> remaining lifetime (I'm 70). Now data retention is a limitation,
>>>> with only 10 years.... I do expect to be around for longer than
>>>> that. But I doubt I'd still be using the same netbook for that long.
>>>
>>> This is a very complex issue. Most of what's going on is trade secret
>>> and unknowable by mere mortals.
>>> Reading is not a problem. Writing is the problem.
>>>
>>> I thought that moving the program files folder would be ok, because
>>> it's all read stuff and the volatile stuff would be stored in the
>>> registry on the HD. But when I checked it out, I found over a hundred
>>> .ini files that store volatile stuff like most recently used files.
>>>
>>> If it gets written once when you close the program, you're probably
>>> safe. If you're copying 10,000 files and the MRU entry gets updated
>>> for every file, you've got a problem.
>>>
>>> Some programs may also store temporary files or caches in their
>>> install directory.
>>>
>>> Depending on the wear leveling algorithm, if any, it may matter a LOT
>>> whether the flash drive is mostly empty of mostly full.
>>>
>>> You'll have to google it, but there are some tools designed for
>>> embedded windows that dramatically reduce writes to the drive. Not
>>> at all clear if they would work the way you've set up your system. There
>>> was a long thread a few months ago about longevity of SSD. Think it was
>>> in comp.sys.ibm.pc.hardware.storage.
>>>
>>> A million sounds like a big number until you start talking about how
>>> fast a computer can run through it.
>>
>> Well for starters, I use 100,000 per cell write cycles for my SLC flash
>> drive figures, which are more conservative. And if I wrote 4GB per hour
>> for 24 hours per day on a 4GB flash drive, it would take 11 years to wear
>> it out.
>>
>> MTBF for a SLC flash is 227 years, 7 times longer than hard drives.
>>
>
> MTBF does not include wear out due to use. MTBF refers to the expected
> life of the memory if it is read only
>
>> I average 100MB to 150MB writes per day of writes to SLC flash drives. At
>> that rate, it would take 4,000 years to hit 100,000 writes per cell.
>>
>
> A lot depends on how they are written. If there is a wear levelling
> controller (likely), then small writes will make the memory last longer
> than it is really entitled to.
>
> But the reality is that all the figures quoted are average figures. The
> deviation from the mean is substantial indeed for FLASH memory, and
> individual examples can have extremely short lives. And I am talking
> memory from reputable suppliers not shonky made in China stuff.
>

I should also point out that failure of one cell in a memory will often
render the whole memory useless. The effects vary depending on the
controller or the internal housekeeping, but the most common failure mode is
the controller or housekeeping logic renders the memory read only.
 
M

M.I.5¾

Flightless Bird
"BillW50" <BillW50@aol.kom> wrote in message
news:hiqmeq$mc1$1@news.eternal-september.org...
> In news:hiqm5g$23m$1@reader1.panix.com,
> the wharf rat typed on Fri, 15 Jan 2010 21:19:44 +0000 (UTC):
>> In article <hiqkid$9e7$1@news.eternal-september.org>,
>> BillW50 <BillW50@aol.kom> wrote:
>>>
>>> Not from what I understand. It isn't a hard failure, but a soft one.
>>> As
>>
>> Writing data to flash requires applying an electrical charge. The
>> individual cells "trap" that charge and the gate (cell) resists the
>> passage of the sensor current. The ability to store that charge
>> degrades with each write cycle. Eventually the cell becomes unable
>> to store enough charge to provide the required resistance and reads
>> permanently high - always shows 1.
>
> True, but that takes 100,000 rewrites per cell to kill it. Writing 100MB
> per day on a 4GB would take 4,000 years to rewrite every cell 100,000
> times.
>

No. It takes an *average* of 100,000 rewrites. It only requires one cell
with a substantially shorter life to kill the memory. The memory logic
usually (but not always) locks the memory in read only mode once a faulty
cell is detected.
 
M

M.I.5¾

Flightless Bird
"BillW50" <BillW50@aol.kom> wrote in message
news:hiqp5i$8nk$1@news.eternal-september.org...
> In news:hiqnvm$kf2$1@reader1.panix.com,
> the wharf rat typed on Fri, 15 Jan 2010 21:50:46 +0000 (UTC):
>> In article <hiqmeq$mc1$1@news.eternal-september.org>,
>> BillW50 <BillW50@aol.kom> wrote:
>>>
>>> True, but that takes 100,000 rewrites per cell to kill it. Writing
>>> 100MB
>>
>> Not really. Given a sufficiently large population of cells one
>> cell will fail for every certain number of writes. The chances of any
>> particular cell failing on any particular write are about 1 in 100000,
>> but the chances of any one of the cells in the array failing are much
>> larger. And they're still dependent on the number of writes, so SOME
>> cells will INEVITABLY fail.
>
> So? In time a hard drive will suffer a bad sector or two. They are marked
> as bad and life moves on. The same on a flash drive. No big deal.
>

Not so. If a bad sector on hard drive is detected, the drive logic simply
maps a spare sector to cover the hole and the drive carries on.

Not so on a FLASH memory chip, a bad cell will be detected and render the
whole memory read only (usually). There are FLASH memory chips with spare
blocks to cover failures but I have never seen them in anything other than
very high end equipment.
 
B

BillW50

Flightless Bird
In news:hj0q6h$ej1$2@reader1.panix.com,
the wharf rat typed on Mon, 18 Jan 2010 05:05:21 +0000 (UTC):
> In article <hivp9a$4po$1@news.eternal-september.org>,
> BillW50 <BillW50@aol.kom> wrote:
>>
>> bottom-line, right? And the 12 flash drives I have hasn't even
>> dropped 1% of the original capacity yet. It isn't even close.
>
> How many of those are in service as virtual memory or scratch
> fiel space for an operating system installation?

It varies, hard drives have the same problem.

>> 20 years from now, 16GB flash drives will be a dime a dozen by then.
>> ;-)
>
> Sure. And in 2001 we'll all be driving flying cars.

Not very knowledgeable about this computer stuff, are you? Back in the
mid '80's it was easy to get 32KB flash drives dirt cheap. In the late
'90's, it was true for 32MB flash drives. A thousand times more storage.
Today, it is easy to purchase 32GB flash drives. That is another
thousand times more storage just 10 years ago. So you are acting foolish
and believe that we won't have cheap 32TB flash drives in a decade or
two, eh?

"People who say it cannot be done should not interrupt the people who
are doing it." -- Anonymous

--
Bill
Gateway M465e ('06 era) - Windows XP SP3
 
B

BillW50

Flightless Bird
In news:4b544b3f$1_1@glkas0286.greenlnk.net,
M.I.5¾ typed on Mon, 18 Jan 2010 11:52:26 -0000:
> "BillW50" <BillW50@aol.kom> wrote in message
> news:hiqp5i$8nk$1@news.eternal-september.org...
>> In news:hiqnvm$kf2$1@reader1.panix.com,
>> the wharf rat typed on Fri, 15 Jan 2010 21:50:46 +0000 (UTC):
>>> In article <hiqmeq$mc1$1@news.eternal-september.org>,
>>> BillW50 <BillW50@aol.kom> wrote:
>>>>
>>>> True, but that takes 100,000 rewrites per cell to kill it. Writing
>>>> 100MB
>>>
>>> Not really. Given a sufficiently large population of cells one
>>> cell will fail for every certain number of writes. The chances of
>>> any particular cell failing on any particular write are about 1 in
>>> 100000, but the chances of any one of the cells in the array
>>> failing are much larger. And they're still dependent on the number
>>> of writes, so SOME cells will INEVITABLY fail.
>>
>> So? In time a hard drive will suffer a bad sector or two. They are
>> marked as bad and life moves on. The same on a flash drive. No big
>> deal.
>
> Not so. If a bad sector on hard drive is detected, the drive logic
> simply maps a spare sector to cover the hole and the drive carries on.
>
> Not so on a FLASH memory chip, a bad cell will be detected and render
> the whole memory read only (usually). There are FLASH memory chips
> with spare blocks to cover failures but I have never seen them in
> anything other than very high end equipment.

I don't know where you are getting this nonsense! But I just had a bad
block on my Adata 16GB SDHC Class 6 flash drive just the other day. No
problem, marked as bad and everything is fine. And how do you explain my
three different capacities with my three 16GB Adata flash drives? That
is if they are not being marked as bad, then why is there a difference
in capacity between them? This is reported by CHKDSK.

1) 15,880,752 KB total disk space. 1,985,094 blocks.
2) 15,672,112 KB total disk space. 1,959,014 blocks.
3) 15,664,088 KB total disk space. 1,958,011 blocks.

"People who say it cannot be done should not interrupt the people who
are doing it." -- Anonymous

--
Bill
Gateway M465e ('06 era) - Windows XP SP3
 
B

BillW50

Flightless Bird
In news:hj1bi9$eqj$1@news.eternal-september.org,
~misfit~ typed on Mon, 18 Jan 2010 23:01:40 +1300:
> Somewhere on teh intarwebs BillW50 wrote:
>> In news:hitld1$hn0$1@news.eternal-september.org,
>> ~misfit~ typed on Sun, 17 Jan 2010 13:25:01 +1300:
>>> Somewhere on teh intarwebs BillW50 wrote:
>>>> In news:hiqjra$o0b$2@reader1.panix.com,
>>>> the wharf rat typed on Fri, 15 Jan 2010 20:40:10 +0000 (UTC):
>>>>> In article <hiq7mg$dcp$1@news.eternal-september.org>,
>>>>> BillW50 <BillW50@aol.kom> wrote:
>>>>>>
>>>>>> I average 100MB to 150MB writes per day of writes to SLC flash
>>>>>> drives. At that rate, it would take 4,000 years to hit 100,000
>>>>>> writes per cell.
>>>>>
>>>>> But writes aren't evenly distributed across cells. The exact
>>>>> pattern depends on free space available, the balancing algorithms,
>>>>> and the patterns of the load. Also, MTBF doesn't mean that ALL
>>>>> those cells will last that long. It's a statistic; you'll see
>>>>> plenty of failures before that magic number is reached.
>>>>
>>>> No problem. As that is what wear leveling takes care of. I have
>>>> been using flash drives (SSD) for running Windows for two years
>>>> now and no problems yet.
>>>
>>> So you've been using the term 'flash drive' to refer to an SSD from
>>> the start of this thread?
>>>
>>> That's pretty confusing as they're two different things with fairly
>>> well established nomenclature.
>>
>> How do you figure?
>
> <sigh>
>
> Google "Flash drive".
>
> Then Google "SSD".

Actually, you are the one who needs to use Google. Here let me help you.

http://stackoverflow.com/questions/555458/whats-the-difference-between-a-flash-drive-and-an-ssd

Now solid state drive stores folders and files in solid-state memory
(either Flash / SRAM / DRAM). And flash does not require constant power
and is non-volatile while SRAM and DRAM are volatile. Also when we talk
about SSD drives here, we are generally talking about the flash based
ones. The very same type of memory used in flash drives.

The reason why we don't talk much about SRAM and DRAM based solid state
drives here in this newsgroup is that they are not really practical for
our hard drive replacements. Although a few of us do take advantage of
DRAM / SRAM based SSD in the form of a RAM disk. My drive R on my
computers are actually RAM based drives. Not too useful for storing
anything long term on them do to being so volatile though. Think you got
all of this straight yet?

> I don't know why you feel the urge to write all this other stuff
> below.

Not very knowledgeable about this computer stuff, are you Shaun? Most of
it just flies right over your head, eh?

And I am sorry that you are so darn confused about the terms SSD and
flash drive. But to put it into simple terms that a simpleton could
understand, they are one in the same. Thus flash based SSD and flash
drive are one in the same.

>> From the outside they might look different. But
>> from the inside, they are all the same. As they use the same
>> technology with the same SLC / MLC chips for the drive. And the same
>> rules of longevity and wear leveling also apply. Speaking of which,
>> have you ever thought about how much writing is evolved for 100,000
>> writes per cell? For a 4GB for example, you would have to write 400TB
>> worth to reach the expected lifespan. That is a lifetime or more
>> worth the way I use computers.
>> And the problem that I had solved to get the full Windows XP to fit
>> inside of less than 2GB SSD, was to move the Program Files folder
>> over to a SD flash drive. Sounds easy, except Windows sees the SD
>> flash as a removable drive. Plus all of the many registry entries
>> are thinking that programs are in the C:/Program Files folder. And
>> many programs that you wish to later install will refuse to install
>> on a removable drive anyway.
>>
>> All of these problems disappear, if you mount the SD flash in the
>> Program Files folder. As the Windows registry is happy, as nothing as
>> far as it is concern has changed. And Windows and the applications
>> are happy, as they don't see the SD drive as a removable drive
>> anymore. And unlike previous hacks to pull this off, which requires
>> registry hacking and slower I/O performance. This one even a
>> non-computer geek can pull it off very easily.
>>
>> Thus I believe this information is a gold mine for those with 2GB and
>> 4GB SSD running Windows XP. As free space can get very limited really
>> fast if you are not careful. And this allows for a lot of more
>> breathing room. ;-)

--
Bill
Gateway M465e ('06 era) - Windows XP SP3
 
B

BillW50

Flightless Bird
In news:hj2691$hg2$1@news.eternal-september.org,
BillW50 typed on Mon, 18 Jan 2010 11:37:26 -0600:
> In news:hj1bi9$eqj$1@news.eternal-september.org,
> ~misfit~ typed on Mon, 18 Jan 2010 23:01:40 +1300:
>> Somewhere on teh intarwebs BillW50 wrote:
>
>> I don't know why you feel the urge to write all this other stuff
>> below.
>
> Not very knowledgeable about this computer stuff, are you Shaun? Most
> of it just flies right over your head, eh?

Thus if you bothered to read it and actually understood it. You would
realize that SSD and flash drive are one in the same. You seem to
believe that SSD (solid state drive) and flash drive are different. They
actually are not. Windows might see them as different, thus adds to your
confusion.

But Windows doesn't see SSD and flash drive as different per se, but how
it connects to the computer as different. As both flash SSD and flash
drives are the same. Although Windows sees some solid state drives as
removable. And the point of the extra verbiage was to trick Windows in
seeing a removable flash drive (aka SSD) as a fixed flash drive (aka
SSD). But I guess you were not smart enough to figure that all out and
thus adds to your confusion. <sigh>

--
Bill
Gateway M465e ('06 era) - Windows XP SP3
 
You must log in or register to reply here.
Top