ReadyBoost Performance Improvement?

  • Thread starter Thread starter cvp
  • Start date Start date
C

cvp

After being prompted on several occasions to donate my USB devices to
improving my system performance, I decided to try it out yesterday.

The machine I tried it on has 2GB memory (approx 50% in use during tests)
and I used a 1GB Simpletech USB device.
In theory I expected improvements in sleep, sleep to hibernate, hibernate,
and resume, with unobservable hits (either way) to performance while
running,

I was a little surprised to find that sleep and resume performance was
degraded (presumably the resume because it recreates the cache on the device
after each boot (why?).

Has anyone else noticed any measurable improvements? What are the
circumstances that cause the improvement?
 
Drop down to about 512MB of ram, then you'll get your improvement. According
to a Tomshardware test, which doesn't cover every possible scenario, but
gives a good indicator. 1GB and higher doesn't see much (if any) gain with
ready boost

http://www.tgdaily.com/2007/02/08/analysis_vista_ready_boost/

Maybe someone else has some tangible benchmark that can be run showing a 2GB
system getting a performance gain out of readyboost, but it appears to be
limited to under powered systems
 
ReadyBoost moves your pagefile (a file that mimics RAM) from the hard drive
to the flash drive, giving it faster access. So, if you have 2 GB of RAM and
only use 50%, you're not going to hit the pagefile much. If you drop to 1 GB
and are using 95-100% of your RAM, your pagefile will be busy, and you will
notice a difference in speed.

I don't notice much, if any, improvement on a 2 GB system, but I do on a 512
(which was unbearable, so it got upgraded to 1.5 GB!) and a 1 GB.
 
I think the key here is that pagefile usage on a system with 2GB is going to
be rare anyway. As long has you have unused physical RAM, Vista has very
little use for the pagefile. Since Readyboost is only designed to speed up
pagefile access, it makes sense to me that you won't see any improvement in
overall system performance unless you were using the pagefile to some extent
to begin with.

cvp said:
After being prompted on several occasions to donate my USB devices to
improving my system performance, I decided to try it out yesterday.

The machine I tried it on has 2GB memory (approx 50% in use during tests)
and I used a 1GB Simpletech USB device.
In theory I expected improvements in sleep, sleep to hibernate, hibernate,
and resume, with unobservable hits (either way) to performance while
running,
Click to expand...

I think this is because it can't be sure that the flash device didn't get
removed and information changed while Vista was away.
I was a little surprised to find that sleep and resume performance was
degraded (presumably the resume because it recreates the cache on the
device after each boot (why?).
Click to expand...

I purchased a cheap USB stick to hold a donated a 512mb 60x SD card early in
my Vista Usage. If it weren't for appearing in the list of drives, I would
have probably long ago forgotten it's there. Since my total investment is
about $25, I guess I don't really care if it's helping or not.
 
hi

cvp said:
After being prompted on several occasions to donate my USB devices to
improving my system performance, I decided to try it out yesterday.

The machine I tried it on has 2GB memory (approx 50% in use during tests)
and I used a 1GB Simpletech USB device.
In theory I expected improvements in sleep, sleep to hibernate, hibernate,
and resume, with unobservable hits (either way) to performance while
running,

I was a little surprised to find that sleep and resume performance was
degraded (presumably the resume because it recreates the cache on the
device after each boot (why?).

Has anyone else noticed any measurable improvements? What are the
circumstances that cause the improvement?
Click to expand...

My 'windows experience index' went from 4.8 to 5.0 after
I put my Disgo 1gig stick in. I took it out again and the syetem went
back to 4.8. So I leave it in now.
 
thanatos to vista

long live xp :-)


Thanatos said:
hi



My 'windows experience index' went from 4.8 to 5.0 after
I put my Disgo 1gig stick in. I took it out again and the syetem went
back to 4.8. So I leave it in now.
Click to expand...
 
Interesting concept, Dale. Well, it might be interesting if I could get the
same performance out of 512MB with a flash device, as I do with 2GB.
Unfortunately, I don't have any systems with memory sticks less than 1GB, so
I can't even test it if I wanted to. I guess the conclusion is that with 2GB
I shouldn't expect an improvement. Well, OK, but I didn't expect it to
degrade performance. I guess I'll just give the whole idea a miss!
 
Dustin,

So if I retest with a workload with a working set over 2GB, I might see a
difference. Hmmm...
Well, I might try again, but I think I'd need a larger flash drive to match
the pagefile size.
As I replied to someone else, it was disappointing to see degradation and no
improvement. I think I'd rather spend any cash I had for that larger flash
drive on more memory.
 
So, as I replied above, the money for a pagefile size flash device would be
better spent on more memory.
 
bye

Thanatos said:
hi



My 'windows experience index' went from 4.8 to 5.0 after
I put my Disgo 1gig stick in. I took it out again and the syetem went
back to 4.8. So I leave it in now.
Click to expand...
 
Physical RAM is always prefered over ReadyBoost. If you are buying a flash
drive for exclusive use with ReadyBoost, it's a mistake. You will get MANY
times improvement by adding more RAM. You don't want Windows to access the
pagefile, either on a flash drive or a hard drive. It does that when it's
out of fast physical memory. ReadyBoost is faster than a hard drive, but
still doesn't even come close to physical RAM.

If you have a chance and need more performance, get more RAM.

--
Dustin Harper
(e-mail address removed)
http://www.vistarip.com

--
 
You could always use the Maxmem option to set your system to unly use
X-amount of the ramy ou have available. Under 2k\Xp it was a boot.ini option
where you added /Maxmem=512

Under Vista, you run msconfig, it the Avance tab and choose maxmem.

http://images1.filecloud.com/395728/advance_config.jpg

Reboot and you'll have yourself a crippled system, just ripe for testing
 
In message <#[email protected]> "Dustin Harper"
Physical RAM is always prefered over ReadyBoost. If you are buying a flash
drive for exclusive use with ReadyBoost, it's a mistake. You will get MANY
times improvement by adding more RAM. You don't want Windows to access the
pagefile, either on a flash drive or a hard drive. It does that when it's
out of fast physical memory. ReadyBoost is faster than a hard drive, but
still doesn't even come close to physical RAM.
Click to expand...

Of course you can get a 2GB flash drive for under $30, going from 1GB to
2GB of RAM on my notebook would be over $150.

RAM is much better, no doubt, but ReadyBoost has some value.
 
In message <[email protected]> "Dale White"
Drop down to about 512MB of ram, then you'll get your improvement. According
to a Tomshardware test, which doesn't cover every possible scenario, but
gives a good indicator. 1GB and higher doesn't see much (if any) gain with
ready boost

http://www.tgdaily.com/2007/02/08/analysis_vista_ready_boost/

Maybe someone else has some tangible benchmark that can be run showing a 2GB
system getting a performance gain out of readyboost, but it appears to be
limited to under powered systems
Click to expand...

Alternatively, if you have a lot of RAM, but devote a large chunk to
virtual machines, you may see an improvement while the VMs are running.

One of my laptops has 1GB, even without a VM running it's noticeably
faster, but that's likely because the laptop has a 4300rpm drive.

On my other laptop, with a 7200rpm drive, doesn't really make a
difference unless I fire up a VM or two and use 256MB-512MB of my
memory.

Your mileage may vary.
 
DevilsPGD said:
In message <[email protected]> "Dale White"


Alternatively, if you have a lot of RAM, but devote a large chunk to
virtual machines, you may see an improvement while the VMs are running.
Click to expand...

That's what I have found, too. With 2GB of Ram, it is the only time
I have ever *noticed* ReadyBoost making a difference.


-Michael
 
In message <[email protected]> "MICHAEL"
That's what I have found, too. With 2GB of Ram, it is the only time
I have ever *noticed* ReadyBoost making a difference.
Click to expand...

VMs, or anything else that uses a ton of memory -- System performance
while running a prime95 stress test, for example.

Now, how often do you need to optimize for performance while running an
app that uses most/app of your available RAM?

You might not -- But then again, you might do image editing. Me, I'm a
software guy, I spend much of my day in a virtual machine which runs an
app that needs a chunk of RAM.

As soon as Microsoft sees fit to give Action Pack subscribers access to
the 64-bit release, I'll grab a couple more GB of RAM and not have to
care.
 
Back
Top