Windows Experience Index

K

Ken Blake

If anyone has a few minutes and the equipment to do a quick
experiement I think the results might be interesting. (Unfortunately
I can't do the experiment myself yet or I would have just reported
what I found instead of begging for a favor)

"Ready Boost" is having Windows use a thumb drive for a fast cache.
http://en.wikipedia.org/wiki/ReadyBoost
I understand completely that this is NOT equivalent to installing an SSD.

If anyone has a mid-range system with Win7 with a mid-range SATA
drive and no SSD and one or two fast USB 3.0 8 or 16 gig thumb drives
then I would be creally curious if the WEI changes at all, whether a system
"feels even a little faster" and mostly whether from the sound the system is
hammering the hard drive less with this enabled when you test this with
zero, one and possibly two of these Ready Boost drives enabled.

My experience with ReadyBoost was that it was completely useless
except on systems with a very small amount of RAM.
 
S

s|b

"Ready Boost" is having Windows use a thumb drive for a fast cache.
http://en.wikipedia.org/wiki/ReadyBoost
I understand completely that this is NOT equivalent to installing an SSD.
I have an SSD and Intel SSD Toolbox says ReadyBoost is not found:
Apparently some USB 3.0 drives are twice the max speed of others.
Apparently USB 2.0 thumb drives are slow enough that Windows
may not even bother using them for Ready Boost.
I have a USB flash drive (SanDisk 64 GB), but I can't seem to use
ReadyBoost because of the SSD:
I'm wondering whether for twenty dollars a couple of the fast thumb
drives would make enough of a difference that I could tell without my
stopwatch and whether it would hammer the head of the drive enough
less that I'd be able to notice. I[m hoping dual fast thumb drives may
be enough.

If anyone is in a position to do a test and report back what their
configuration was and what they saw and even what might be a good
brand and model of fairly inexpensive thumb drives then perhaps I
and others can benefit.
I bought my SanDisk on eBay for about 58 euro.

<http://www.ebay.co.uk/itm/SanDisk-E...80-/230805643836?ssPageName=ADME:L:OC:BE:3160>

I could try ReadyBoost on my mother's laptop, but what software should I
use to test the speed?
 
S

s|b

My experience with ReadyBoost was that it was completely useless
except on systems with a very small amount of RAM.
Aha. Looks like I won't be testing...
 
B

Bill Simpson

s|b said:
Aha. Looks like I won't be testing...
I'm not trying to promote ReadyBoost. I want to be clear about that.

But thumb drives vary in speed by more than 80x, depending on brand, etc.

A Google search for 'speed usb thumb drive' turned up a nice table
http://usbspeed.nirsoft.net/
telling that the Sandisk Cruzer Glide I just bought has about 20 MB reads
and 6 MB writes. That is far from some of the claimed 80 MB values that
some manufacurers state, but also far from the 2 MB lower limit where
ReadyBoost supposedly just ignores the thumb drive. Unfortunately I
think it also turned out that my thumb is USB 2.0, not USB 3.0 that I
thought it was supposed to be, but the package has no hint of which it is.

But that table also doesn't really seem to indicate whether this is for
short
bursts of data or for large blocks. Thumb drives have approximately zero
seek time, unlike rotating drives, which favors the thumb for lots of small
bursts, like the zillion seeks at startup, but the rotating drive will do
lots
better when you ask to read that 50 gigabyte defragmented file in a single
stream.

A Google search for 'win7 test hard drive speed' turns up all sorts of
what look like easy to use and possibly safe utilities, some apparently
built into Windows, that will tell you your hard drive speed.

On the up side, trying out ReadyBoost seems to be as simple as erasing
anything you have on the drive, unplugging it, plugging it back in and
choosing "Speed up my system" when Win7 asks you what you want
to do with this new hardware. And when you are tired of this you then
unplug it and everything is back the way it was before you started.
Compare that to the cost and the nightmare it would take to try
migrating everything to an SSD and then migrating it all back to your
rotating drive when done.

All I wanted to know were hardware specs and whether a pair of really
fast USB 3.0 thumb drives, so they operate in parallel and double the
speed, might hammer the heads on my drive less than it otherwise would.

So if one or a few people do not have an SSD, have a midrange system,
have one or two fast USB 3.0 thumb drives with read/write speeds that
are at least a few times faster than the measured speed of your rotating
drive and want to take a few minutes for what looks like should be a
safe and quick test and then report all the numbers then I think we might
all learn a little bit. And I'd send you a cookie if I could figure out a
way
to do it.

Again, I was not expecting my system to be ten times faster just by
plugging in a couple of $10 thumb drives, all I was really hoping for
was that it might hammer the heads on the drive a lot less.

Thanks
 
S

s|b

But thumb drives vary in speed by more than 80x, depending on brand, etc.

A Google search for 'speed usb thumb drive' turned up a nice table
http://usbspeed.nirsoft.net/
telling that the Sandisk Cruzer Glide I just bought has about 20 MB reads
and 6 MB writes.
I downloaded usbdeview-x64.zip and ran the EXE. My SanDisk Extreme USB
3.0 Flash Drive 64GB:

Write speed: 140.63 MB/Sec
Read speed: 169.81 MB/Sec
On the up side, trying out ReadyBoost seems to be as simple as erasing
anything you have on the drive, unplugging it, plugging it back in and
choosing "Speed up my system" when Win7 asks you what you want
to do with this new hardware. And when you are tired of this you then
unplug it and everything is back the way it was before you started.
Compare that to the cost and the nightmare it would take to try
migrating everything to an SSD and then migrating it all back to your
rotating drive when done.
One would be mad to migrate back from SSD to PATA/SATA. ;-)
All I wanted to know were hardware specs and whether a pair of really
fast USB 3.0 thumb drives, so they operate in parallel and double the
speed, might hammer the heads on my drive less than it otherwise would.

So if one or a few people do not have an SSD, have a midrange system,
have one or two fast USB 3.0 thumb drives with read/write speeds that
are at least a few times faster than the measured speed of your rotating
drive and want to take a few minutes for what looks like should be a
safe and quick test and then report all the numbers then I think we might
all learn a little bit. And I'd send you a cookie if I could figure out a
way to do it.
I thought I could try this on my mother's laptop, but then I realized:
she hasn't got USB 3.0 ports. In fact, nobody I know (besides myself)
has USB 3.0 ports. Pity... :-(
 
G

Gene E. Bloch

I downloaded usbdeview-x64.zip and ran the EXE. My SanDisk Extreme USB
3.0 Flash Drive 64GB:

Write speed: 140.63 MB/Sec
Read speed: 169.81 MB/Sec


One would be mad to migrate back from SSD to PATA/SATA. ;-)


I thought I could try this on my mother's laptop, but then I realized:
she hasn't got USB 3.0 ports. In fact, nobody I know (besides myself)
has USB 3.0 ports. Pity... :-(
Try a USB2 to USB3 converter.

....No, no, no, that's a very bad joke.

Most laptops today seem to come with USB3 ports, and most modern
motherboards do as well.

My MB has two USB3 ports on the back, and I have added a set of four
ports up front by getting a PCIx card. More modern boards than mine also
have headers on the MB for front-panel USB3; mine doesn't. Of course,
none of this will help for an existing laptop.

Actually, a laptop with an Express Card port might be adaptable for
USB3, but that AFAICT, that port is uncommon on current laptops.

Here's one such card at Amazon US:
http://tinyurl.com/a4raxg4
 
K

Ken Blake

On the up side, trying out ReadyBoost seems to be as simple as erasing
anything you have on the drive, unplugging it, plugging it back in and
choosing "Speed up my system" when Win7 asks you what you want
to do with this new hardware. And when you are tired of this you then
unplug it and everything is back the way it was before you started.

No argument from me. There's no downside to trying it. But as I said,
my experience with it (on several different computers and with several
different thumb drives) has been that it did nothing discernable,
unless the RAM was under 2GB.
 
S

s|b

No argument from me. There's no downside to trying it. But as I said,
my experience with it (on several different computers and with several
different thumb drives) has been that it did nothing discernable,
unless the RAM was under 2GB.
Were some of those thumb drives USB 3.0?
 
B

Bill Simpson

s|b said:
But thumb drives vary in speed by more than 80x, depending on brand, etc.

A Google search for 'speed usb thumb drive' turned up a nice table
http://usbspeed.nirsoft.net/
telling that the Sandisk Cruzer Glide I just bought has about 20 MB reads
and 6 MB writes.
I downloaded usbdeview-x64.zip and ran the EXE. My SanDisk Extreme USB
3.0 Flash Drive 64GB:

Write speed: 140.63 MB/Sec
Read speed: 169.81 MB/Sec [snip]
I think I have figured it out.

My "fast thumb drive" is 1/6 the speed of my rotating drive I was trying
to speed up. I had no idea until I measured it. That sounds hopeless.

Your "fast thumb drive" is 8x times the speed of mine and 3/2 the speed
of my rotating drive and a pair of your thumb drives might be 3x the
speed of my rotating drive, not counting seek time and driver
performance questions. That might show a little promise.

So I'll go buy a pair of your fast thumb drives and try this again.

Google
regedit readyboost
turns up a few bits of information I was not aware of.

I thought there might be some edits to get closer to just turning off the
rotating drive for minimal use.

If anyone gets more concrete data then please post it.
Thanks for the information from everyone.
 
P

Paul

s|b said:
Were some of those thumb drives USB 3.0?
I have a Lexar S73, which is a cheap USB3. I got it for around $20
or so. It does only 45MB/sec on USB3, but when plugged into a USB2
port, it actually does 35MB/sec. Whereas one of my hard drive USB2
enclosures, does slightly less than that. So to me, the new wave
of USB3 drives are actually interesting. They've managed to get
a bit more from a USB2 port, than a USB2 device can.

The writes are probably still shy of being flat out, but they're
getting there.

This is much better than a "Sandisk Ultra 8GB" USB2 flash I got at a Staples,
which works in spurts, and only managed 4MB/sec or so. Since I bought
them on the same day, from two different stores, the comparison between
them is truly pathetic. Sandisk should be ashamed of themselves.
Both USB flash devices cost around the same, but the Lexar S73
happens to have 4x the capacity, and ~8X the speed. I think
the worst part, is Sandisk would use the word "Ultra" to describe
the product. Marketing *fail*.

And you can do better on USB3 devices plugged into USB3 ports.
Patriot makes some fat-body USB3 flash keys, which have
somewhere around 100MB/sec or so. The fat-body is needed to
support more flash channels, which is how those things
get their speed. So if you want something faster, look
for a key which is physically bigger.

The USB3 keys, work in both USB3 and USB2. Since all my hardware
is USB2, I've been testing my USB3 key in only USB2 computers.
The 45MB/sec quoted above, is the rated speed, and I couldn't
test that. I still don't have any USB3 ports.

*******

In terms of booting from USB3 on USB3, very few chipsets have
native USB3 ports on the Southbridge. Intel delivered theirs only
very recently. And AMD didn't put it on all their product lines.
Many motherboards, the USB3 comes from a NEC, Asmedia, or other
brand of chip, and when that happens, it might not support booting
via those ports. It really depends on the BIOS design. In the past,
the rule of thumb was, that only ports directly on the Southbridge,
had good coverage at boot time. (Storage chips are OK, because they
ship with Extended INT 0x13 bios code modules, but things like USB3
chips, I don't think they come with any support code the motherboard
maker can use. That's why they'd be a problem.) There would be
no problem booting from a USB3 flash stuffed into a USB2 socket
which is actually hosted by the Southbridge. So if you notice
some permutations and combinations don't work for booting,
that could be why - they're hosted by two different chips.
And BIOS designs have never been that good, at supporting
all hardware on board, to the same level. And that's why
"integrated ports" on the Southbridge, are so valuable.

Paul
 
R

R. C. White

Hi, Bill.

Does your computer have the hardware to support USB 3.0? Maybe I missed it,
but I don't see where you've said that it does.

We can plug a 3.0 drive into a 2.0 port and it will work fine - at 2.0
speed - because 3.0 is backward-compatible with 2.0. But to get 3.0 speeds
you must have a motherboard with 3.0 ports, or an add-in circuit board with
such ports. And, of course, the BIOS and drivers must support 3.0, too.

My 2-month-old MSI 990FXA-GD90 mobo has both 3.0 and 2.0 ports. The 3.0
ports are distinguished by the BLUE "tongues" inside the connectors. Some
USB 3.0 devices, such as my Lexar thumb drive and Seagate GoFlex external
HDD, also have blue; maybe they all do. I've not run any speed tests.

As a side comment, I see no reason for new computers to have USB 2.0 ports
since USB 3.0 devices are backwards compatible, while 2.0 ports are not
"forward compatible". Why have four 2.0 ports and two 3.0 ports; why not
just have all six ports 3.0? Is there a significant cost difference to the
motherboard maker?

RC
--
R. C. White, CPA
San Marcos, TX
(e-mail address removed)
Microsoft Windows MVP (2002-2010)
Windows Live Mail 2012 (Build 16.4.3505.0912) in Win8 Pro


"Bill Simpson" wrote in message

s|b said:
But thumb drives vary in speed by more than 80x, depending on brand, etc.

A Google search for 'speed usb thumb drive' turned up a nice table
http://usbspeed.nirsoft.net/
telling that the Sandisk Cruzer Glide I just bought has about 20 MB reads
and 6 MB writes.
I downloaded usbdeview-x64.zip and ran the EXE. My SanDisk Extreme USB
3.0 Flash Drive 64GB:

Write speed: 140.63 MB/Sec
Read speed: 169.81 MB/Sec [snip]
I think I have figured it out.

My "fast thumb drive" is 1/6 the speed of my rotating drive I was trying
to speed up. I had no idea until I measured it. That sounds hopeless.

Your "fast thumb drive" is 8x times the speed of mine and 3/2 the speed
of my rotating drive and a pair of your thumb drives might be 3x the
speed of my rotating drive, not counting seek time and driver
performance questions. That might show a little promise.

So I'll go buy a pair of your fast thumb drives and try this again.

Google
regedit readyboost
turns up a few bits of information I was not aware of.

I thought there might be some edits to get closer to just turning off the
rotating drive for minimal use.

If anyone gets more concrete data then please post it.
Thanks for the information from everyone.
 
P

Paul

R. C. White said:
Hi, Bill.

Does your computer have the hardware to support USB 3.0? Maybe I missed
it, but I don't see where you've said that it does.

We can plug a 3.0 drive into a 2.0 port and it will work fine - at 2.0
speed - because 3.0 is backward-compatible with 2.0. But to get 3.0
speeds you must have a motherboard with 3.0 ports, or an add-in circuit
board with such ports. And, of course, the BIOS and drivers must
support 3.0, too.

My 2-month-old MSI 990FXA-GD90 mobo has both 3.0 and 2.0 ports. The 3.0
ports are distinguished by the BLUE "tongues" inside the connectors.
Some USB 3.0 devices, such as my Lexar thumb drive and Seagate GoFlex
external HDD, also have blue; maybe they all do. I've not run any speed
tests.

As a side comment, I see no reason for new computers to have USB 2.0
ports since USB 3.0 devices are backwards compatible, while 2.0 ports
are not "forward compatible". Why have four 2.0 ports and two 3.0
ports; why not just have all six ports 3.0? Is there a significant cost
difference to the motherboard maker?

RC
--
R. C. White, CPA
San Marcos, TX
(e-mail address removed)
Microsoft Windows MVP (2002-2010)
Windows Live Mail 2012 (Build 16.4.3505.0912) in Win8 Pro
It's to do with the state of the chipsets providing the feature.

Very few chipsets have included USB3. The ones that do, may not
do it on all ports.

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

Z77 Panther Point USB Rev 3.0, 4 Ports and
USB Rev 2.0, 10 ports

And that would be considered a good case. On a motherboard like
mine, the Southbridge only has USB2 ports. If the manufacturer
wanted USB3 ports, they'd use a NEC USB3 chip (2 ports) or
an Asmedia chip. Some of the latest add-on chips, may have
four USB3 ports, helping to reduce the cost of adding that
many ports.

On the AMD side...

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

A75 Hudson-D3 USB Rev 3.0, 4 Ports and
USB Rev 2.0, 10 ports
USB Rev 1.1 2 ports [really?]

So when the two major chipset supplies limit themselves
when it comes to integrated chipsets, as a motherboard
manufacturer you don't have a say in the matter.

A typical "well-tuned" motherboard design then, ends up
with four USB3 ports, either by using two NEC chips,
a single add-on four port USB3 chip, or by using the
USB3 ports on the Southbridge. In the case of the AMD A75,
you might need to purchase an FM2 socket motherboard, to
get the integrated USB3. If you picked up an AM3 socket
motherboard, the chipsets are different, and then to get
USB3 means an add-on chip instead.

For best BIOS support (i.e. booting from USB3), it
helps if the USB3 ports are integrated in the Southbridge
of the chipset. Generally, BIOS writers support *everything*
on a Southbridge. Support for add-on chips, is more of a joke.

Paul
 
R

R. C. White

Hi, Paul.

As I mentioned, my MSI 990FXA-GD80 mobo has 4 USB 3.0 ports. Two are on the
backplane, right between the audio jacks and two USB 3.0 ports. The other
two are on a bracket provided to be mounted in an empty slot in the back of
the case. This bracket has a cable long enough (about 18") to reach the
front of the mobo where it can plug into a flat blue 20-pin connector, right
next to the six SATA 3 (6 Gb/s) ports. In my small tower case, the spacing
between that socket and my HDDs is so awkward that the plug bent some pins
in the socket while I was trying to plug it in and I had to replace the new
mobo.

Here is my new mobo (I had a typo in the number in my earlier post):
http://www.msi.com/product/mb/990FXA-GD80.html

While I've built my own computers for the past 20 years, I'm a retired
accountant, not a techie, and don't really know much about chipsets and
such. But I can plug parts and components together, like a hi-tech Lego
set. ;^}

RC
--
R. C. White, CPA
San Marcos, TX
(e-mail address removed)
Microsoft Windows MVP (2002-2010)
Windows Live Mail 2012 (Build 16.4.3505.0912) in Win8 Pro


"Paul" wrote in message
R. C. White said:
Hi, Bill.

Does your computer have the hardware to support USB 3.0? Maybe I missed
it, but I don't see where you've said that it does.

We can plug a 3.0 drive into a 2.0 port and it will work fine - at 2.0
speed - because 3.0 is backward-compatible with 2.0. But to get 3.0
speeds you must have a motherboard with 3.0 ports, or an add-in circuit
board with such ports. And, of course, the BIOS and drivers must support
3.0, too.

My 2-month-old MSI 990FXA-GD90 mobo has both 3.0 and 2.0 ports. The 3.0
ports are distinguished by the BLUE "tongues" inside the connectors. Some
USB 3.0 devices, such as my Lexar thumb drive and Seagate GoFlex external
HDD, also have blue; maybe they all do. I've not run any speed tests.

As a side comment, I see no reason for new computers to have USB 2.0 ports
since USB 3.0 devices are backwards compatible, while 2.0 ports are not
"forward compatible". Why have four 2.0 ports and two 3.0 ports; why not
just have all six ports 3.0? Is there a significant cost difference to
the motherboard maker?

RC
It's to do with the state of the chipsets providing the feature.

Very few chipsets have included USB3. The ones that do, may not
do it on all ports.

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

Z77 Panther Point USB Rev 3.0, 4 Ports and
USB Rev 2.0, 10 ports

And that would be considered a good case. On a motherboard like
mine, the Southbridge only has USB2 ports. If the manufacturer
wanted USB3 ports, they'd use a NEC USB3 chip (2 ports) or
an Asmedia chip. Some of the latest add-on chips, may have
four USB3 ports, helping to reduce the cost of adding that
many ports.

On the AMD side...

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

A75 Hudson-D3 USB Rev 3.0, 4 Ports and
USB Rev 2.0, 10 ports
USB Rev 1.1 2 ports [really?]

So when the two major chipset supplies limit themselves
when it comes to integrated chipsets, as a motherboard
manufacturer you don't have a say in the matter.

A typical "well-tuned" motherboard design then, ends up
with four USB3 ports, either by using two NEC chips,
a single add-on four port USB3 chip, or by using the
USB3 ports on the Southbridge. In the case of the AMD A75,
you might need to purchase an FM2 socket motherboard, to
get the integrated USB3. If you picked up an AM3 socket
motherboard, the chipsets are different, and then to get
USB3 means an add-on chip instead.

For best BIOS support (i.e. booting from USB3), it
helps if the USB3 ports are integrated in the Southbridge
of the chipset. Generally, BIOS writers support *everything*
on a Southbridge. Support for add-on chips, is more of a joke.

Paul
 
R

R. C. White

Hi, again, Paul.

Another typo - in the second sentence. Should say "...between the audio
jacks and two USB 2.0 ports", not 3.0 this time. :>{ Sorry 'bout that.

RC

"R. C. White" wrote in message

Hi, Paul.

As I mentioned, my MSI 990FXA-GD80 mobo has 4 USB 3.0 ports. Two are on the
backplane, right between the audio jacks and two USB 3.0 ports. The other
two are on a bracket provided to be mounted in an empty slot in the back of
the case. This bracket has a cable long enough (about 18") to reach the
front of the mobo where it can plug into a flat blue 20-pin connector, right
next to the six SATA 3 (6 Gb/s) ports. In my small tower case, the spacing
between that socket and my HDDs is so awkward that the plug bent some pins
in the socket while I was trying to plug it in and I had to replace the new
mobo.

Here is my new mobo (I had a typo in the number in my earlier post):
http://www.msi.com/product/mb/990FXA-GD80.html

While I've built my own computers for the past 20 years, I'm a retired
accountant, not a techie, and don't really know much about chipsets and
such. But I can plug parts and components together, like a hi-tech Lego
set. ;^}

RC
--
R. C. White, CPA
San Marcos, TX
(e-mail address removed)
Microsoft Windows MVP (2002-2010)
Windows Live Mail 2012 (Build 16.4.3505.0912) in Win8 Pro


"Paul" wrote in message
R. C. White said:
Hi, Bill.

Does your computer have the hardware to support USB 3.0? Maybe I missed
it, but I don't see where you've said that it does.

We can plug a 3.0 drive into a 2.0 port and it will work fine - at 2.0
speed - because 3.0 is backward-compatible with 2.0. But to get 3.0
speeds you must have a motherboard with 3.0 ports, or an add-in circuit
board with such ports. And, of course, the BIOS and drivers must support
3.0, too.

My 2-month-old MSI 990FXA-GD90 mobo has both 3.0 and 2.0 ports. The 3.0
ports are distinguished by the BLUE "tongues" inside the connectors. Some
USB 3.0 devices, such as my Lexar thumb drive and Seagate GoFlex external
HDD, also have blue; maybe they all do. I've not run any speed tests.

As a side comment, I see no reason for new computers to have USB 2.0 ports
since USB 3.0 devices are backwards compatible, while 2.0 ports are not
"forward compatible". Why have four 2.0 ports and two 3.0 ports; why not
just have all six ports 3.0? Is there a significant cost difference to
the motherboard maker?

RC
It's to do with the state of the chipsets providing the feature.

Very few chipsets have included USB3. The ones that do, may not
do it on all ports.

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

Z77 Panther Point USB Rev 3.0, 4 Ports and
USB Rev 2.0, 10 ports

And that would be considered a good case. On a motherboard like
mine, the Southbridge only has USB2 ports. If the manufacturer
wanted USB3 ports, they'd use a NEC USB3 chip (2 ports) or
an Asmedia chip. Some of the latest add-on chips, may have
four USB3 ports, helping to reduce the cost of adding that
many ports.

On the AMD side...

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

A75 Hudson-D3 USB Rev 3.0, 4 Ports and
USB Rev 2.0, 10 ports
USB Rev 1.1 2 ports [really?]

So when the two major chipset supplies limit themselves
when it comes to integrated chipsets, as a motherboard
manufacturer you don't have a say in the matter.

A typical "well-tuned" motherboard design then, ends up
with four USB3 ports, either by using two NEC chips,
a single add-on four port USB3 chip, or by using the
USB3 ports on the Southbridge. In the case of the AMD A75,
you might need to purchase an FM2 socket motherboard, to
get the integrated USB3. If you picked up an AM3 socket
motherboard, the chipsets are different, and then to get
USB3 means an add-on chip instead.

For best BIOS support (i.e. booting from USB3), it
helps if the USB3 ports are integrated in the Southbridge
of the chipset. Generally, BIOS writers support *everything*
on a Southbridge. Support for add-on chips, is more of a joke.

Paul
 
S

s|b

This is much better than a "Sandisk Ultra 8GB" USB2 flash I got at a Staples,
which works in spurts, and only managed 4MB/sec or so. Since I bought
them on the same day, from two different stores, the comparison between
them is truly pathetic. Sandisk should be ashamed of themselves.
Both USB flash devices cost around the same, but the Lexar S73
happens to have 4x the capacity, and ~8X the speed. I think
the worst part, is Sandisk would use the word "Ultra" to describe
the product. Marketing *fail*.
I've always been happy about the performance of my SanDisk flash drives.
I used USBDeview to test their speed.

SanDisk U3 Cruzer Micro USB 2.0 (512 MB)
Write speed: 3.39 MB/Sec
Read speed: 17.12 MB/Sec

SanDisk Ultra Cruzer U3 Titanium USB 2.0 (4 GB)
Write speed: 9.25 MB/Sec
Read speed: 27.87 MB/Sec

SanDisk Extreme Contour USB 2.0 (16 GB)
Write speed: 10.89 MB/Sec
Read speed: 28.28 MB/Sec

SanDisk Extreme USB 3.0 (64 GB)
Write speed: 128.21 MB/Sec
Read speed: 169.81 MB/Sec

"Ultra" may have been a poor choice of words, "Extreme" is right on the
top.
And you can do better on USB3 devices plugged into USB3 ports.
Patriot makes some fat-body USB3 flash keys, which have
somewhere around 100MB/sec or so. The fat-body is needed to
support more flash channels, which is how those things
get their speed. So if you want something faster, look
for a key which is physically bigger.
My SanDisk Extreme Contour (USB 2.0) is slightly bigger than my SanDisk
Extreme (USB 3.0) and the latter seems to have no problem reaching
speeds (read/write) over 100 MB/Sec.
 
P

Paul

s|b said:
I've always been happy about the performance of my SanDisk flash drives.
I used USBDeview to test their speed.
SanDisk Extreme USB 3.0 (64 GB)
Write speed: 128.21 MB/Sec
Read speed: 169.81 MB/Sec

"Ultra" may have been a poor choice of words, "Extreme" is right on the
top.
What a relief!

At least one of their products doesn't suck.

Paul
 
S

s|b

What a relief!

At least one of their products doesn't suck.
Maybe you just used the wrong software to test read/write speed.

<https://en.wikipedia.org/wiki/USB_flash_drive#File_transfer_speeds>

<quote>

File transfer speeds

USB flash drives usually specify their read and write speeds in
megabytes per second (MB/s); read speed is usually faster. These speeds
are for optimal conditions; real-world speeds are usually slower. In
particular, circumstances that often lead to speeds much lower than
advertised are transfer (particularly writing) of many small files
rather than a few very large ones, and mixed reading and writing to the
same device. In a typical well-conducted review[35] of a number of
high-performance USB 3.0 drives, a drive that could read large files at
68MB/s and write at 46MB/s, could only manage 14MB/s and 0.3MB/s with
many small files. When combining streaming reads and writes the speed of
another drive, that could read at 92MB/s and read at 70MB/s, was 8MB/s.
These differences differ radically from one drive to another; some
drives could write small files at over 10% of the speed for large ones.
The examples given are chosen to illustrate extremes.

</quote>

I tested the same flash drive (SanDisk Extreme USB 3.0 64 GB) with
CrystalDiskMark. Portable version can be downloaded here:
<http://sourceforge.jp/projects/crystaldiskmark/downloads/46482/CrystalDiskMark3_0_2e.zip/>

Results were very different:
<https://imageshack.us/a/img43/5674/crystaldiskmarkextreme.png>

| -----------------------------------------------------------------------
| CrystalDiskMark 3.0.2 x64 (C) 2007-2013 hiyohiyo
| Crystal Dew World : http://crystalmark.info/
| -----------------------------------------------------------------------
| * MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]
|
| Sequential Read : 181.855 MB/s
| Sequential Write : 163.584 MB/s
| Random Read 512KB : 139.438 MB/s
| Random Write 512KB : 26.071 MB/s
| Random Read 4KB (QD=1) : 11.575 MB/s [ 2826.0 IOPS]
| Random Write 4KB (QD=1) : 10.094 MB/s [ 2464.4 IOPS]
| Random Read 4KB (QD=32) : 9.816 MB/s [ 2396.4 IOPS]
| Random Write 4KB (QD=32) : 4.872 MB/s [ 1189.5 IOPS]
|
| Test : 1000 MB [S: 42.3% (25.2/59.6 GB)] (x5)
| Date : 2013/03/09 17:47:10
| OS : Windows 7 Home Premium Edition SP1 [6.1 Build 7601] (x64)

An explanation can be found in Help (F1) > How to Use

<quote>

Test Types

All : All Test (Seq, 512K, 4K, 4K QD32)
Seq : Sequential Read/Write Test (Block Size = 1024KB)
512K : Random Read/Write Test (Block Size = 512KB)
4K : Random Read/Write Test (Block Size = 4KB)
4K QD32 : Random Read/Write Test (Block Size = 4KB, Queue Depth = 32)
for NCQ&AHCI

</quote>

So maybe you need to test again...
 
P

Paul

s|b said:
Maybe you just used the wrong software to test read/write speed.
The crappy Sandisk I bought, has visibly defective behavior.
It will not transfer data, at a consistent rate of speed.
My other flash keys are much more consistent in how they
behave.

It does this, no matter where in the flash storage space
it is writing. It isn't a bad block problem.

It's a controller problem. The controller is the small
chip on the left.

http://upload.wikimedia.org/wikipedia/commons/e/e2/Usbkey_internals.jpg

Even if I do a sustained transfer, the best way possible
("dd", with block size to suit flash), it still behaves like crap.
With "dd", you can do a transfer for testing, that is aligned
with the flash storage format inside.

Any manufacturer who makes USB keys, should review their behavior
in the lab, before selling them. It's obvious in this case, that
nobody at Sandisk cares.

Paul
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top