ReadyBoost causing excessive paging performance issue

[Guest]

I performed the 60 s system performance scan (I can't remember the exact name
but it was in the Control Panel) both with and without my ReadyBoost-capable
flash device connected to the PC. My device has 4GB of space, but my PC has
1 GB of RAM on it, so I set the device to use ~2.5 GB for ReadyBoost paging.

With the device connected, the report says that there is excessive paging,
and I have on average ~128 MB of RAM free.
When I disconnect the device, the report does not warn me about excessive
paging, and it says I have on average ~256 MB of RAM free.
However, I believe that the ReadyBoost is making a difference in the overall
performance because I can tell some programs boot faster with the device
connected than when it is not connected. However, this could just be my
imagination if the performance report is telling me I am better off without
the device.
So, basically, I would just like an explanation about this report because I
don't understand it. If I just uprgraded my RAM, to say 3GB, would it then
make little difference whether the ReadyBoost device is causing excessive
paging?

 

[Richard Urban]

Sounds to me like your performance scan is not ReadyBoost aware. I test by
using a stop watch when I load a large program. After the first start,
subsequent restarts of the program are about 75% faster.

ReadyBoost also does not do much if one has 2 gig of RAM, or more.

Also, as some tests have proven out - not all ReadyBoost capable USB thumb
drives are created equal. Some are 50% faster than others in performance. In
one test I read the least expensive USB thumb drive was tested to be the
fastest out of about a dozen tested.

--


Regards,

Richard Urban
Microsoft MVP Windows Shell/User
(For email, remove the obvious from my address)

Quote from George Ankner:
If you knew as much as you think you know,
You would realize that you don't know what you thought you knew!

 

[Steve Thackery]

Hang on a minute. ReadyBoost doesn't affect program loading times, does it?
I thought it was a kind of pre-buffer for the paging file, thus speeding up
the handling of page faults.

All NT-based versions of Windows load a program much faster the second and
subsequent times. The files stay in RAM until it is wanted by some other
process. This is a different mechanism from ReadyBoost. New to Vista,
Superfetch even tries to predict what you're going to want next, and
pre-loads it into RAM for you. Again, though, this is different from
ReadyBoost.

If' I've misunderstood something, please enlighten me!

Steve

 

[Richard Urban]

Maybe my basic test is wrong (probably is) but when I apply the stop watch,
with and without my USB thumb drive connected, I get a difference. Granted,
with 2 of RAM gig it is miniscule. With 1 gig of RAM it is somewhat more
substantial.

Yes. ReadyBoost is an extension of (add on to) the pagefile. It is a lot
faster that the hard drive. Yet RAM is faster than both. It just "appears"
to me that with the USB stick in the slot that subsequent reloads are faster
than without the stick (by X amount). That being said, I do not use ready
boost as it increases my startup time by 2-3 minutes as the ready boost
cache is getting populated.

Now a thought. At start up I am doing nothing with the computer. I am
making/drinking my coffee. Yet, Ready Boost cache is slowly filling. Filling
with what? I am not using the computer. Could it be that it is filling with
superfetch information? RAM overflow? It is nothing that I am doing that is
causing this yet it is cashing data from somewhere in anticipation of using
it how? If it is filling from the prefetch folder (programs that I used the
last few sessions) it could well speed up the programs loading as opposed as
working from the hard drive. But this I would think would effect the "first"
start of a program also, as well as later restarts.

Also, it has been reported that Ready Boost on some computers actually seems
to slow the system down. There are many variables and not many absolutes
regarding Ready Boost. Maybe the algorithm will be tweaked at a later point
(SP1).

--


Regards,

Richard Urban
Microsoft MVP Windows Shell/User
(For email, remove the obvious from my address)

Quote from George Ankner:
If you knew as much as you think you know,
You would realize that you don't know what you thought you knew!