OT? Windows 7: Faster than XP?

[Yousuf Khan]

What *is* slow on this machine, is Paging. If something needs to page
in a significant way, the process needing the paging can experience
a significant dead time, of anywhere from 30 seconds to a minute.
It would seem that the Paging is done with random 4K blocks, causing
the hard drive grief. So the head on the disk gets a good workout.
I consider that to be more objectionable, than the WinXP boot behavior,
because it distracts from what you're trying to do. I can eliminate this
effect entirely, if I place the pagefile on a RAMDisk (the 6GB installed
test case for this machine, 2GB RAMDisk). That was smooth as can be.

I have moved my page file to an alternate drive, in fact, I moved it to
two alternate drives, half on one drive, half on the other, to get a bit
of interleaving going (don't know if any version of Windows interleaves
its page files). I found that there is not much paging going on during
startup though, so it doesn't help startup. Paging goes on much more
after certain programs like Thunderbird or Firefox are started.

Yousuf Khan

 

[Yousuf Khan]

Then it's time to reformat and reinstall XP.

That's the most back-assward suggestion yet. XP would've just kept
booting for that 17 minutes, without letting you do anything.

 

[Yousuf Khan]

Per Yousuf Khan:

The biggest improvement I got was when I converted the system
drive from hard disc to SDD.

That is exactly what is needed nowadays. I'm awaiting affordable 200GB
SSD's before I upgrade. That's what's necessary to be comfortable in Win
7 for me right now.

Yousuf Khan

 

[BillW50]

In Yousuf Khan typed:

I have moved my page file to an alternate drive, in fact, I moved it
to two alternate drives, half on one drive, half on the other, to get
a bit of interleaving going (don't know if any version of Windows
interleaves its page files). I found that there is not much paging
going on during startup though, so it doesn't help startup. Paging
goes on much more after certain programs like Thunderbird or Firefox
are started.

I have read reports of those who had done their homework and claims
there is no payoff doing it that way. I also use RAM instead of a drive
for a page file because it is over 20 times faster.

 

[BillW50]

In Yousuf Khan typed:

That is exactly what is needed nowadays. I'm awaiting affordable 200GB
SSD's before I upgrade. That's what's necessary to be comfortable in
Win 7 for me right now.

Yousuf Khan

I bought a SSD for Windows 7 and it cost me 50 bucks.

 

[Kernel]

| On 23/02/2012 12:43 PM, Kernel wrote:
| > Then it's time to reformat and reinstall XP.
|
| That's the most back-assward suggestion yet. XP would've just kept
| booting for that 17 minutes, without letting you do anything.

Do it this way and it won't boot for 17 minutes. Insert XP installation
disc.
Restart computer, boot to CD, F8, delete existing partition, create new one,
quick NTFS format, setup XP. When finished it will boot in about 45 secs
and shut down in about 10. Works for me.

 

[Char Jackson]

In Yousuf Khan typed:

I bought a SSD for Windows 7 and it cost me 50 bucks.

$50 for a 200GB SSD is a great price. I'm not seeing them that cheap.

 

[BillW50]

In

$50 for a 200GB SSD is a great price. I'm not seeing them that cheap.

No not for a 200GB SSD. When Windows 7 RC first came out in 2009, it
needed 16GB to install. And my 702s only had 8GB SSD. So I bought a 16GB
so I can install and run Windows 7 on one of them. But Windows 7 ended
up being such a lemon on a single core processor, that SSD now lives in
a junk drawer. As I never needed 16GB for Linux or Windows XP.

 

[Tim Slattery]

I have read reports of those who had done their homework and claims
there is no payoff doing it that way. I also use RAM instead of a drive
for a page file because it is over 20 times faster.

??? But the whole point of the paging file is to allow the VM system
to move less-used pages *out* of RAM so that more needed pages can be
moved in.

 

[BillW50]

In

??? But the whole point of the paging file is to allow the VM system
to move less-used pages *out* of RAM so that more needed pages can be
moved in.

That is the poor man's way of creating more RAM from disk space. But it
isn't free. It comes at a cost of many times slower performance. The
easiest thing one can do for is computer to boost speed is to install
more RAM. More RAM means less VM is needed. And if you have enough RAM
to run everything, VM just slows you down and you don't need it.

 

[(PeteCresswell)]

Per Char Jackson:

$50 for a 200GB SSD is a great price. I'm not seeing them that cheap.

When I was shopping mine, I got the impressing that SSDs come in
different flavors - with different reliabilities. I have no clue
as to the specifics.

I went for the Intel, which everybody seemed to agree was the
most reliable.

 

[BillW50]

In

Per Char Jackson:

When I was shopping mine, I got the impressing that SSDs come in
different flavors - with different reliabilities. I have no clue
as to the specifics.

I went for the Intel, which everybody seemed to agree was the
most reliable.

Yes, I hear very good things about Intel SSD. I also have had good luck
with Samsung SSD. There is two types of SSDs that I believe most don't
know about. They are the SLI and MLC SSDs. The SLI type is far, far,
better than MLC type. MLC type is half of the cost, slower, and lifespan
is about a tenth of what SLI types last.

 

[Paul]

In

Yes, I hear very good things about Intel SSD. I also have had good luck
with Samsung SSD. There is two types of SSDs that I believe most don't
know about. They are the SLI and MLC SSDs. The SLI type is far, far,
better than MLC type. MLC type is half of the cost, slower, and lifespan
is about a tenth of what SLI types last.

The other dimension of the problem, is time. MLC about three years ago,
would have a longer life than MLC now. They went from around 5000 write
cycles, to about 1000 in some cases. I would expect SLC to be affected
to some extend by geometry shrinkage, but haven't seen any comparison
information offered for them. The MLC went from two bits per cell to
three bits per cell, which really cranks the need for precise analog
levels and thresholds in each cell (four level to eight level).

As the geometry shrinks, the flash chips rely more and more on error
correction. The ECC used, needs more bits per block, to achieve the
same performance as the previous generation.

And this leads to the following two items, seen in recent days.

"The Bleak Future of NAND Flash Memory"
http://cseweb.ucsd.edu/users/swanson/papers/FAST2012BleakFlash.pdf

And a rebuttal
http://pcper.com/reviews/Editorial/NAND-Flash-Memory-Future-Not-So-Bleak-After-All

I'm still waiting for this stuff to come out. Then we can have
some competition, even if it takes these guys a while to
catch up.

http://en.wikipedia.org/wiki/Memristor

"In October 2011, the same team announced the commercial availability
of memristor technology within 18 months, as a replacement for Flash,
SSD, DRAM and SRAM."

"In April 2010, HP labs announced that they had practical memristors
working at 1 ns (~1 GHz) switching times and 3 nm by 3 nm sizes, with
electron/hole mobility of 1 m/s, which bodes well for the future of
the technology. At these densities it could easily rival the current
sub-25 nm flash memory technology."

Paul

 

[Kernel]

| In message <[email protected]>, Kernel
| []
| >If you want to boot XP in about 5 seconds, just install the Asrock 890 FX
| >Deluxe 4 mobo with instant boot. Yes, you read that correctly, 5 second
| >boot times with XP on that board with the Instant Boot Utility.
| >
| >
| How can a board make that difference - unless it has on-board SSD (even
| with that 5 seconds seems fast)? I assume we _are_ talking about full
| boot, not restore from something.
| --
|
There is no onboard SSD. How does the board do it? It's done with the
BIOS. The Asrock 890FX Deluxe 4 comes with a software suite, one
of the items offered is called INSTANT BOOT. They advertise 3 sec
boot times, mine takes about 7 so I split the difference. It's a mighty
fine motherboard...have two myself.

 

[Kernel]

http://www.zdnet.com/blog/hardware/asrock-4-second-instant-boot-for-windows/3012


| In message <[email protected]>, Kernel
| []
| >There is no onboard SSD. How does the board do it? It's done with the
| >BIOS. The Asrock 890FX Deluxe 4 comes with a software suite, one
| >of the items offered is called INSTANT BOOT. They advertise 3 sec
| >boot times, mine takes about 7 so I split the difference. It's a mighty
| >fine motherboard...have two myself.
| >
| >
| So how _does_ it do it, assuming we're talking about time from power-on
| to XP fully usable with nothing more to come? "It's done with the BIOS"
| I fear needs further explanation.
| --
| J. P. Gilliver. UMRA: 1960/<1985 MB++G.5AL-IS-P--Ch++(p)Ar@T0H+Sh0!:`)DNAf
|
| "I am entitled to my own opinion."
| "Yes, but it's your constant assumption that everyone else is also that's
so
| annoying." - Vila & Avon

 

[Kernel]

Oh no, not at all. The computer is completely off, and it boots from the
full off status.
It's truly amazing, boot Win XP in 5 seconds. I'd suggest you Google Asrock
Instant
Boot, or just click on this url:

http://www.asrock.com/feature/InstantBoot/index.asp

| In message <[email protected]>, Kernel
| >http://www.zdnet.com/blog/hardware/asrock-4-second-instant-boot-for-wind
| >ows/3012
| >
| >
| >| []
| >| So how _does_ it do it, assuming we're talking about time from power-on
| >| to XP fully usable with nothing more to come? "It's done with the BIOS"
| >| I fear needs further explanation.
| []
| So it doesn't actually do a _boot_, it does a _return from suspend_ (or
| possibly hibernate).
| --
| J. P. Gilliver. UMRA: 1960/<1985 MB++G.5AL-IS-P--Ch++(p)Ar@T0H+Sh0!:`)DNAf
|
| "I am entitled to my own opinion."
| "Yes, but it's your constant assumption that everyone else is also that's
so
| annoying." - Vila & Avon

 

[Patok]

Oh no, not at all. The computer is completely off, and it boots from the
full off status.
It's truly amazing, boot Win XP in 5 seconds. I'd suggest you Google Asrock
Instant
Boot, or just click on this url:

http://www.asrock.com/feature/InstantBoot/index.asp

No, it does not boot from off. it resumes from suspend, where it was
placed after a previous clean boot. *You* read the url again, since you
didn't get it (apparently).
And since it is from suspend, one needs to have the computer
connected to a UPS too, it seems.

| In message <[email protected]>, Kernel
| >http://www.zdnet.com/blog/hardware/asrock-4-second-instant-boot-for-wind
| >ows/3012
| >
| >
| >| []
| >| So how _does_ it do it, assuming we're talking about time from power-on
| >| to XP fully usable with nothing more to come? "It's done with the BIOS"
| >| I fear needs further explanation.
| []
| So it doesn't actually do a _boot_, it does a _return from suspend_ (or
| possibly hibernate).
| --
| J. P. Gilliver. UMRA: 1960/<1985 MB++G.5AL-IS-P--Ch++(p)Ar@T0H+Sh0!:`)DNAf
|
| "I am entitled to my own opinion."
| "Yes, but it's your constant assumption that everyone else is also that's
so
| annoying." - Vila & Avon

 

[BillW50]

In Patok typed:

Oh no, not at all. The computer is completely off, and it boots
from the full off status.
It's truly amazing, boot Win XP in 5 seconds. I'd suggest you
Google Asrock Instant
Boot, or just click on this url:

http://www.asrock.com/feature/InstantBoot/index.asp

No, it does not boot from off. it resumes from suspend, where it
was placed after a previous clean boot. *You* read the url again,
since you didn't get it (apparently).
And since it is from suspend, one needs to have the computer
connected to a UPS too, it seems.

In message <[email protected]>, Kernel
http://www.zdnet.com/blog/hardware/asrock-4-second-instant-boot-for-wind
ows/3012


message []
So how _does_ it do it, assuming we're talking about time from
power-on to XP fully usable with nothing more to come? "It's done
with the BIOS" I fear needs further explanation.
[]
So it doesn't actually do a _boot_, it does a _return from suspend_
(or possibly hibernate).

I didn't get that impression when I read that URL. Although perhaps you
are saying it stores a fresh boot and will use that copy to boot from
now on? If so, that is a lot like embedded Windows, Windows SteadyState,
a system sandbox, etc.

All which are very handy under some conditions. Like if you get malware,
it is totally gone after a reboot. Although all changes to the OS is
gone too. Like Windows updates, application installs, etc.

There was a guy on the eeeuser forum who stated his XP booted in 10
seconds. He did admit that he used a lite version of XP. Well he got my
interest since on a netbook especially, turning on and off a lot is very
useful to conserve battery life on the go. Although when I pressed him
further and I wanted to know more details, he claimed up.

I generally use BartPE from a SDHC or a CD. But from a SSD or a physical
hard drive I tried it once. And BartPE might actually boot in 10
seconds. While BartPE is great for some tasks, it doesn't make a great
general purpose OS.

 

[Patok]

In Patok typed:

I didn't get that impression when I read that URL. Although perhaps you
are saying it stores a fresh boot and will use that copy to boot from
now on? If so, that is a lot like embedded Windows, Windows SteadyState,
a system sandbox, etc.

No, not exactly. From what I understood from that page and comments
elsewhere, what it does is:

When you turn off / shut down, the installed software (it must be
installed) does a shutdown followed by a reboot, and /then/ stores the
fresh booted state into either suspend or hibernate. When you come back
to the computer next day, it resumes from there. So it is not the same
boot copy every time, but the most recent one.

And if this is a tower PC (not a laptop with batteries, if I
understand), then one needs constant power / UPS to be able to not have
lost power and state from suspend. Hibernate OTOH would be OK, it seems.

 

[BillW50]

In Patok typed:

No, not exactly. From what I understood from that page and comments
elsewhere, what it does is:

When you turn off / shut down, the installed software (it must be
installed) does a shutdown followed by a reboot, and /then/ stores the
fresh booted state into either suspend or hibernate. When you come
back to the computer next day, it resumes from there. So it is not
the same boot copy every time, but the most recent one.

And if this is a tower PC (not a laptop with batteries, if I
understand), then one needs constant power / UPS to be able to not
have lost power and state from suspend. Hibernate OTOH would be OK,
it seems.

Oh I see. So it isn't really saving you any time. Just rebooting while
being unintended and waiting for you to wake it up later. Fascinating
and clever, but doesn't really offer anybody anything who are happy with
standby and/or hibernation. Which in my case can last weeks or months at
a time without a reboot.