Are you ready for ReadyBoost?

WarMace

WarMace

Gawd
Joined
Aug 25, 2004
Messages
963
Do you have a Flash drive that you have tested on Windows Vista RTM (v6000 ONLY)? If so please document your experience here including the model of flash drive. Hopefully we can gradually document a list of compatible flash drives.

ReadyBoost is a feature of Vista and promises great performance improvements. It is the use of a fast flash drive for improving small random I/O.
(Not to be confused with ReadyDrive, which involves hybrid hard drives.)

Key FAQ's can be found Tom Archer's Blog:
http://blogs.msdn.com/tomarcher/archive/2006/06/02/615199.aspx

and in Jim Allchin's Vista team site posting:
http://windowsvistablog.com/blogs/windowsvista/archive/2006/11/20/windows-readyboost.aspx

the key requirement a flash drive must meet is described in Tom Archers blog as "2.5MB/sec throughput for 4K random reads and 1.75MB/sec throughput for 512K random writes"

You can get your flash drive's speed score using the following commands in vista while running the command prompt AS ADMINISTRATOR:
Read test: winsat disk –read –ran –ransize 4096 –drive f (f is the drive letter without colon)
Write test: winsat disk –write –ran –ransize 524288 –drive f (this one takes a few min, the size is mesured in bytes)

More thorough testing info can be found thanks to MixManSC
MixManSC said:
Here is a tool and real documentation that I got directly from MS on Readyboost in case any of you are interested.

http://www.sl-digital.com/mixmansc/ReadyBoostKit.zip
Click to expand...


I myself have tested 4 flash drives:
Memorex Mini TravelDrive U3 (1GB and 512MB) - 4.24 MB/s (4K)read; 1.72 MB/s (512K)write, this score is .03 MB/s to slow. This drive passed in Beta, but fails in the RTM version. Plus U3 does not work in vista, and e-mails to U3 so far have been useless, with one reply telling me its a issue with Microsoft. A no-go until they get in shape. UPDATE: after a format, this drive passes....barely.

TravelDrive Original 512MB and 256MB - Fails as well. :(
 
SanDisk Cruzer Micro 1GB = No issues

SanDisk Cruzer Micro 4GB = No issues

I use both at various times, sometimes just parts of the capacity, sometimes the full space on the drive.

What I'm really waiting on is some way of actually benchmarking the performance differences between not only different brands of USB drives, but exactly how it benefits Vista in day-to-day operation.

So far, all I've found spread across the Net is "Plug it in and it helps" but I can't find anything that specifically says why it helps and specifically how. Normal benchmarking apps currently available simply don't do any good in this yet, so I'm hoping there will be some way to check.

Also, WinSAT doesn't seem to take ReadyBoost into account during it's performance rating tests. I ran and re-ran the tests after changing some hardware and video drivers, the ratings are always the same.

So if ReadyBoost is supposed to help Vista work faster overall, and I stick in a ReadyBoost 4GB USB drive and use all it, I want something to say "Ok, it's working, and here's the difference in performance with it and without it."

Right now, there is no such tool.

Let's hope someone figures out a way to prove it works, and soon, or it'll just look like Microsoft blowing smoke up our collective asses.

:D
 
Duh.

Being the dumbass that I am on occasion - we all are sometimes, admit it - I didn't post results, so here they are:

SanDisk Cruzer Micro 1GB:

C:\Users\br0adband>winsat disk -read -ran -ransize 4096 -drive h
Windows System Assessment Tool
> Running: Feature Enumeration v1.0.0.0 ''
> Run Time 00:00:00.33
> Running: Storage Performance Assessment via Profiling v1.0.0.0 '-read -ran -ransize 4096 -drive h'
> Run Time 00:00:03.87
> Disk Performance 3.28 MB/s
> Total Run Time 00:00:04.62

C:\Users\br0adband>winsat disk -write -ran -ransize 512000 -drive h
Windows System Assessment Tool
> Running: Feature Enumeration v1.0.0.0 ''
> Run Time 00:00:00.33
> Running: Storage Performance Assessment via Profiling v1.0.0.0 '-write -ran -ransize 512000 -drive h'
> Run Time 00:05:39.49
> Disk Performance 2.56 MB/s
> Total Run Time 00:05:40.24

SanDisk Cruzer Micro 4GB:

C:\Users\br0adband>winsat disk -read -ran -ransize 4096 -drive f
Windows System Assessment Tool
> Running: Feature Enumeration v1.0.0.0 ''
> Run Time 00:00:00.38
> Running: Storage Performance Assessment via Profiling v1.0.0.0 '-read -ran -ransize 4096 -drive f'
> Run Time 00:00:04.12
> Disk Performance 3.17 MB/s
> Total Run Time 00:00:04.91

C:\Users\br0adband>winsat disk -write -ran -ransize 512000 -drive f
Windows System Assessment Tool
> Running: Feature Enumeration v1.0.0.0 ''
> Run Time 00:00:00.33
> Running: Storage Performance Assessment via Profiling v1.0.0.0 '-write -ran -ransize 512000 -drive f'
> Run Time 00:14:06.43
> Disk Performance 2.53 MB/s
> Total Run Time 00:14:07.16


One word of warning: the write test takes a while, as mentioned by the OP but he didn't mention how long it can take. He said "minutes" and that's what I assumed when I tested my 1GB drive but as shown in the results above, the 4GB write test took 14 minutes! So, if anyone has a 4GB or 8GB USB drive for this testing, it's gonna take time to do it. :)

Also, regarding the U3 issue, it's like this: Vista requires a driver to support the U3 features adequately, and Microsoft doesn't write drivers, so guess who needs to write the driver?

:D
 
I've used Readyboost with a USB drive and an SD card. They both work great. My SD card is faster so it works better than the USB drive did.

I can tell you that it seems to have no effect on normal system operation. It *does* seem to have a significant effect on some games (the kind that load and unload textures etc alot)
 
I agree bbz_Ghost, there needs to be a windows level benchmark showing the ReadyBoost improvement. It dosent affect bootup, and its not designed to help something as a fullsystem AV scan. "It just feels faster" is for the birds.

At TechED in boston when MS showed off ReadyBoost they opened 5 Office appications 3 times in a row and showed how it got faster every time with SuperFetch's "Learning" mode.
So am assuming a test suth as opening 1000 1MB photos in new windows would be a way to test...in theory. You just have to do it 3 times in a row w/o readydrive, and then 3 times with. According to spec, it should get faster each time.

Also according to a writer at Bit-tech.com games with long loading times sutch as BF2 should benifit as well.

At least we can benchmark the drives themselves in preperation.
 
Yeah BF2 is helped by this, as are some other games like MMO's that tend to load and unload lots of relatively small files at once. (under a few mb each texture for example)

As far as I can tell it doesnt help anything I do outside of games. I cant see any option in 3dmark that would be affected by this either so its probably not easy to benchmark that way.

If you have a file I/O benchmark that actually uses swap file you could benchmark it that way since ReadyBoost is basically an "addition" to the swapfile that is only used for small transfers. (most benchmarks load files to RAM so you dont get a performance check on the swap file)
 
I've been doing a lot of stuff lately, but still nothing has appeared that I've been able to locate to "prove" the performance benefits of ReadyBoost. When people say "oh yeah it makes <insert app/game name here> much faster" those sorts of "boosts" can be relative to a lot of thing: what's running in the background, etc.

It's not quantifiable data to say "Oh I swear it's like twice as fast;" we have to see some hard data and repeatable results before most of us will accept that says, "yes, it's working to improve performance and here's how much."

Hopefully, that's something we'll be able to do at some point, but I'll keep searching for something that can be used to get some real data on ReadyBoost and how it actually improves system performance.
 
Apacer HT203 "Handy Steno" 4GB:

C:\Windows\system32>winsat disk -read -ran -ransize 4096 -drive f
Windows System Assessment Tool
> Running: Feature Enumeration v1.0.0.0 ''
> Run Time 00:00:00.42
> Running: Storage Performance Assessment via Profiling v1.0.0.0 '-read -ran -ransize 4096 -drive f'
> Run Time 00:00:03.70
> Disk Performance 4.55 MB/s
> Total Run Time 00:00:04.67

C:\Windows\system32>winsat disk -write -ran -ransize 512000 -drive f
Windows System Assessment Tool
> Running: Feature Enumeration v1.0.0.0 ''
> Run Time 00:00:00.45
> Running: Storage Performance Assessment via Profiling v1.0.0.0 '-write -ran -ransize 512000 -drive f'
> Run Time 00:04:42.94
> Disk Performance 9.74 MB/s
> Total Run Time 00:04:43.88

For giggles, I also ran the tests on a 128MB CompUSA-branded USB drive. Got 3.28MB/sec 4K reads, and 2.33 MB/sec 512K writes. I wasn't expecting much here, but it's still fast enough to meet the speed requirements for ReadyBoost. Too small, though...
 
I have a sandisk cruzer micro 1gb and it cam preloaded with the U3 software. Should I delete it before I use the usb drive for ready boost?
 
skriefal:

Damn, I'm jealous. :)

I actually did one test on my 4GB and for some reason it came up with write speeds around 8.5MB/s, so being the idiot that I am (or dumbass, depending on your POV), I redid the test and it dropped to the results posted above.

Makes me wonder if I should retest.

I think, if everyone might agree, that a "standard" format for testing would be appropriate so the results are somewhat relevant. But then again, I'm a dumbass - that time I said it. :)

I wiped both my drives clean and fresh formatted both with FAT32 before the testing, and I also unplugged all my USB devices, including my wireless mouse. I'm on a laptop with 4 USB 2.0 ports, so I wanted to make sure it was only the Cruzers that had access to that bandwidth when I used them.

I might retest later on after a clean reboot just to see if it changes. But that one time it showed 8.5MB/s and I was like, "Wow..." and now I can't repeat that result. Just a fluke, I suppose.
 
I actually ran the 512KB write test three times. The other two runs yielded similar results (9.73MB/s and 10.01MB/s), so I just posted the times from the first run (9.74MB/s). The tests were run on a Dell E1505 Core Duo laptop. The USB drive was NOT reformatted prior to running these tests, so that might have lowered the results a bit.
 
gilga said:
I have a sandisk cruzer micro 1gb and it cam preloaded with the U3 software. Should I delete it before I use the usb drive for ready boost?
Click to expand...

You cannot totally delete it all, and you do not have to, readyboost will use what free space is avialable.
 
warmace said:
You cannot totally delete it all, and you do not have to, readyboost will use what free space is avialable.
Click to expand...

Actually, you can get rid of it all:

http://www.u3.com/uninstall/

If you seriously use that stuff then obviously you wouldn't remove it, but personally, after I got both the 1GB and 4GB Cruzer drives, and I kept having two "drives" show up in Explorer (and on the iMac I owned a few weeks ago) when I inserted them, I got rid of that U3 garbage with all possible haste.

It shouldn't make any difference in benchmarks, however.

I think I'll load up the 1GB and 4GB drives with their original content and then re-run the tests. Who knows, maybe it'll actually speed up that way.

:D
 
Corsair Flash Voyager 4 Gb, properties = Device does not meet the requirements for ready boost :(
 
I was preparing to buy the OCZ Rally2 drive last night. I figured if i want to get the very most out of readyboost, I couldent go wrong with the fastest flash drive on the market.


I guess i was wrong. It failed in a Extremetech Review using Vista RC1. It must of been its poor speed at tiny I/O, where this drive is unbeatable in ISO sized files.

Back to the drawing board.


Nasty_Savage said:
Corsair Flash Voyager 4 Gb, properties = Device does not meet the requirements for ready boost :(
Click to expand...
I would be interested to find out why, I thought that was a quality product. :confused: Hopefully you can post your speed tests. (PM me if you need any help)
 
warmace said:
I would be interested to find out why, I thought that was a quality product. :confused: Hopefully you can post your speed tests. (PM me if you need any help)
Click to expand...

Well, just cause it doesn't work with ReadyBoost doesn't mean it's not a quality product. :)
 
Blitzrommel said:
Well, just cause it doesn't work with ReadyBoost doesn't mean it's not a quality product. :)
Click to expand...

True.

I found a couple sites that have have user submitted compatibility charts, but neither of them state which version of vista they were tested on, or the I/O speeds they achieved.

http://www.grantgibson.co.uk/misc/readyboost/

http://www.neowin.net/forum/index.php?showtopic=514636&st=0

the tests we are using are good because it shows why it may have failed, but the 512000 write test should more accurately be 524288? The whole bits to bytes and fat32 thing still messes with me.

Read test: winsat disk –read –ran –ransize 4096 –drive f (Vista requires > 2.5MB/sec throughput)
Write test: winsat disk –write –ran –ransize 524288 –drive f (Vista requires > 1.75MB/sec throughput)
 
After thinking a little more about how small of a margin my drive failed the test, .03MB/sec, i started toying with the U3 drive to improve performance. After a format at 16kb allocation size the drive passed! :eek: Im now in the process of benchmarking the drive with different allocation sizes to see what changes. Ill post all my findings once finished, 15 min write tests could take all day.
 
I've got a USB 2.0 bay type multi card reader connected directly to my mobo with a 150X OCZ 2GB SD card in it at home. I'll run the tests this evening to see what the reported speeds are and will post them after. It's using the full card for readyboost but like everyone else I can only say certain things "seem" faster....
 
warmace said:
After thinking a little more about how small of a margin my drive failed the test, .03MB/sec, i started toying with the U3 drive to improve performance. After a format at 16kb allocation size the drive passed! :eek: Im now in the process of benchmarking the drive with different allocation sizes to see what changes. Ill post all my findings once finished, 15 min write tests could take all day.
Click to expand...
Can't wait for your findings. I have an old 512mb Sandisk Cruzer I got at BF 2 years ago and have not made much use of it. I decide to pop it in for ReadyBoost and it didn't pass the test so I ran the posted benchmarks. No problems with read speeds but write speeds just fall short. I'm also interested to know if any of you people that have ReadyBoost working see an improvement in your Windows Experience Index score.
 
Lexar JumpDrive 256MB

Microsoft Windows [Version 6.0.6000]
Copyright (c) 2006 Microsoft Corporation. All rights reserved.

C:\Users\Jon>winsat disk -read -ran -ransize 4096 -drive e
Windows System Assessment Tool
> Running: Feature Enumeration v1.0.0.0 ''
> Run Time 00:00:00.28
> Running: Storage Performance Assessment via Profiling v1.0.0.0 '-read -ran -ra
nsize 4096 -drive e'
> Run Time 00:00:03.89
> Disk Performance 5.44 MB/s
> Total Run Time 00:00:04.95

C:\Users\Jon>winsat disk -write -ran -ransize 512000 -drive e
Windows System Assessment Tool
> Running: Feature Enumeration v1.0.0.0 ''
> Run Time 00:00:00.23
> Running: Storage Performance Assessment via Profiling v1.0.0.0 '-write -ran -r
ansize 512000 -drive e'
> Run Time 00:02:04.80
> Disk Performance 7.75 MB/s
> Total Run Time 00:02:05.78
 
rapperwith1p said:
Can't wait for your findings. I have an old 512mb Sandisk Cruzer I got at BF 2 years ago and have not made much use of it. I decide to pop it in for ReadyBoost and it didn't pass the test so I ran the posted benchmarks. No problems with read speeds but write speeds just fall short. I'm also interested to know if any of you people that have ReadyBoost working see an improvement in your Windows Experience Index score.
Click to expand...

Yeah, I've noticed the Windows Experience Index score takes memory performance into consideration as well as capacity. I have 2GB in my laptop, but it was still like a 4.7.
 
rapperwith1p said:
I'm also interested to know if any of you people that have ReadyBoost working see an improvement in your Windows Experience Index score.
Click to expand...

Readyboost made no change to my Windows Experience Index Score.

Blitzrommel said:
Yeah, I've noticed the Windows Experience Index score takes memory performance into consideration as well as capacity. I have 2GB in my laptop, but it was still like a 4.7.
Click to expand...
Keep in mind that your score is out of a total 5.1. So in perspective 4.7/5.1 dosent look so bad. For reference, my notebook (Latitude D820) got a memory score of 4.9 with 2Gig's of DDR2-667 SDRAM.

OK, my testing last night is finished, and are pritty much useless. Except to show the allocation size of a flash drive has no clear effect of flash drive speeds, and PCMark06 is either not ready for vista, or i need better drivers (die to the majority of tests that failed).
I was able to format the flash drive in 3 diffrent allocation sizes, and ran 13 tests on each size. Then for fun i ran PCMark06, 5 times without ready boost, and 3 times with ready boost.
So i don't feel that last nights tests were an entire waste, I'm posting the results for you to interpret as you wish. All tests were run on a Dell Latitude D820 with 2gigs of DDR2-667 and a 1Gig flash drive, and no other devices attached USB or other.

resultsza0.jpg

EDIT: Oops, I accidentally colored the lowest score in red on the Write test average.

Lessons learned? Yes!

If ReadyBoost is advertised to move superfetch data to a faster storage space hence "allows applications and files to load much faster", then probibly a real world test is in order. My next test will be either BF2142 map load, or HalfLife2 Lost Coast load.

Another thing i learned is just enabling readyboost does not instantly speed up a App. You need to wait for data on it to be populated, and apps don't show signs of speeding up until they have been run 4 to 5 times. This is especially important on a fresh vista install.
Q: Isn't this just putting the paging file onto a flash disk?
A: Not really - the file is still backed on disk. This is a cache - if the data is not found in the ReadyBoost cache, we fall back to the HDD.

Q: Aren't Hard Disks faster than flash? My HDD has 80MB/sec throughput.
A: Hard drives are great for large sequential I/O. For those situations, ReadyBoost gets out of the way. We concentrate on improving the performance of small, random I/Os, like paging to and from disk.
Click to expand...
Edit: If ReadyBoost concentrates on small I/O's that leads me to believe our 4k drive benchmark is more important than the large one.

Another thing i learned is Readyboost is especially tricky to test for, because theres nothing to say that app is even preloaded into readyboost. Expecially on a PC with a snappy harddrive and 2Gig's of ram, using a 1gig flash drive. On a older PC with 256MB ram and a 5400rpm HDD, this may be easier to document with a 1Gig stick. I should go buy a 1:1 flash drive for improved results according to Jim Allchin.
 
Wow, that's impressive. I was tempted to spend an entire day doing testing like that myself, over and over again and compile the results, but I do so many various things I just can't dedicate my single laptop to that task.

But again, wow, that's impressive. :)

Now, as you said, the real test would be <hint, hint> real-world testing, and currently there doesn't seem to be anything available to prove ReadyBoost is doing much of anything at all. I will admit myself that yes things do seem to be a bit snappier when I have my 4GB SanDisk Cruzer Micro plugged in and assign all of it's space (3.8GB roughly) to ReadyBoost duties, but I can't prove anything is actually faster.

Sooo... until we can figure out how it's actually affecting performance, I'd say use it if you've got it and just hope for the best.

You'd think Microsoft would know people like us hardcore testers and performance freaks are gonna want PROOF this shit works or it's just the proverbial smoke I mentioned earlier.

:D
 
I just realized we may already have the tool we need to do some testing, or at least proper monitoring:

Resource Monitor, accessed from Task Manager's Performance tab. I fired it up, expanded it to full screen on my external monitor (1680x1050) and then inserted the 4GB Cruzer and turned on ReadyBoost for it.

The first minute or so was just Vista loading up the Cruzer; that much was obvious under the Disk activity section. There was consistent data activity to the drive (strictly writes) hovering around 360MB/s so it was pumping something into the ReadyBoost.sfcache file.

I've been doing some steady websurfing, loading very image heavy pages, jumping tabs, closing the browser, reopening the same pages, etc. One perfect use for this kind of thing would obviously be reassigning the Temporary IE cache to the Cruzer - I've done such things in the past when I used RAMDisk software from SuperSpeed under XP/2K3. Putting the entire IE cache inside RAM makes the browsing experience SCREAM with performance; going Back and Forward - most notably with Opera - is like "blink, blink" because the pages literally blink onscreen most of the time, there's barely any noticeable refresh delay.

I'm all for performance, and right now I think Resource Monitor might be the app of choice to keep tabs on this ReadyBoost thing. It's the only tool so far that actually sees what's going on on at least some level with respect to ReadyBoost and data activity.

Hope this helps... and check Resource Monitor, I firmly believe it's one of the best new things about Vista overall.
 
Heh, i like that kinda like "Boost em' if you got em" kind of attitude.

Digging deeper to find more info, info to fgive me a little background on the topic i found out that Windows Help is dated, MS wrote all of this to work with a driver according to ReadyBoost having a dependency on the "ReadyBoost Caching Driver" further confirming this is a prefetch of data on slower parts of the HDD.

A funny occurrence is the description for the ReadyBoost service uses itself in the description. ReadyBoost Description - "Provides support for improving system performance using ReadyBoost."

 
Good old Paint... hehehe

Here's a tip not many people know about just yet:

Most of us do know the Alt+PrintScreen tip to get just the current window to the Clipboard for a screenshot, and when you open Paint (if that's what you use) to paste it, you get an image like the one above with all that extra white space.

Now comes Paint in Vista with one new option that older versions didn't have: Crop.

Yep, Crop makes an appearance in Vista's Paint. Paste the Clipboard into Paint, choose Image - Crop and wham, no more extra whitespace!!! Save it out and do what you will with it.

</tip_off>

:D
 
I was just lazy and forgot to trim the white, usually i grab the handles on the edge of the pic.

Notice whenever you use the sniping tool, there is a annoying red border, like my benchmark photo.


Im currently writing a application which, in theory, should be able to let me benchmark ReadyBoost itself, and not just the drive directally. I hope it works.
 
One question:

You keep referencing PCMark06 and all I see at Futuremark's website is PCMark05. Typo or... ? I've never heard of any other PCMark so I'm just curious.

Thanks...
 
Kingston DataTraveler 512MB - Fail
Read Disk Performance 3.34 MB/s
Write Disk Performance 1.23 MB/s

OCZ Rally 512MB - Pass
Read Disk Performance 5.30 MB/s
Write Disk Performance 7.10 MB/s
 
OOps, i ment to say PCMark05

Cool! Its good to see someone getting a pass with a OCZ Rally, there is still hope! :eek: Those are some scary fast numbers! :eek: Right up there with the jumpdrive.
 
I work under this motto: If you dont notice a real performance difference yourself, and have to rely on benchmarks to tell you theres a performance gain, does it really matter?
 
I've tested an OCZ Rally 2 (2GB model) in the past (without the Winsat) and it failed the RC2 test. Seems like they're not as fast as the old ones.

@tutelary
Some people are noticing actually differences. That's not the issue. The problem is that it's turning out to be very difficult to quantify the improvement using benchmarks (the opposite of the usual).
 
I could have sworn I read something during the beta that the slower the machine is, the more the benefit ReadyBoost would make. (to a certain point I'm sure) Anyone else read/hear about this? I never tried it in the betas, thought it was honestly a bit of a gimmick and honestly still think it is to some degree still a gimmick. Yeah yeah, why put it in the OS if it's a gimmick, right? I think that if the machine is already fast enough and well past the recommended specs for Vista, then is this ReadyBoot really needed? Or is it indeed really targeted for those Vista machines whose specs are towards the minimum side of things? :confused:
 
ir0nw0lf:

You could have something there, because from the materials I've read, they (meaning the Microsoft media machine) do make it a point to specifically mention the performance improvements for older hardware.

Let's face it: a lot of people have older machines, even ones running off regular PC133 SDRAM that are going to try Vista. It's not like Microsoft can stop any of them, and since their machines are hard limited to the speed of the RAM they've got and limited to how much they can have, this "neat new technology" with ReadyBoost and cheap USB drives is an option.

If it actually works, that's a fantastic and relatively low-priced solution to a lot of people - but we (meaning us HardFolks) know if you're trying to run Vista on a machine over a year old, it's tooth and nail all the way, even if you dropped an 8800GT in it. :)

One funny on-topic thing though:

I recently bought an OCZ "Ultra-High Speed 150x" 2GB SD card from Fry's for my Canon Powershot S3 IS digital camera. Figured, what the hell, toss it into the 4n1 media reader on Ultimate 64 and see what it can do.

The winsat test won't even run on it. :p Is that hilarious or what. So much for "high speed" advertising.

And a 1GB PNY SD card I have, low end stuff, well... that is ReadyBoost capable. Go figure.

The PNY card does 2.58MB/s reads, 3.6MB/s writes, for those that might be curious.
 
Sandisk MiniSD 512mb (connected to dell e1505 via internal reader). Passes.

Read: 3.62 MB/s
Write: 2.50 MB/s
 
I know the OCZ Rally 2GB failed supposeiduly in the readyboost flash drive test.. but i did my own testing and low and behold it supports readyboost! i tried it 3 different times and i got this average...
Read: 2.82
Write: 4.84
 
Its too bad i found out the OCZ rally passes readyboost test After the sale ended. :( Im sure itll go back on sale again.

I'm still working on writing an app to test if ReadyBoost is working, and the more I test the more i learn about SuperCache.

ReadyBoost is tricky to test for because SuperCache is working so well. I wrote an application that times how long (in ms) it takes to open 1000 14kb text documents, in an attempt to simulate "small random i/o's". So far all i have found is that any test must be run almost a dozen times for SuperCache to think my 14kb files are more important to cache than say my outlook .ost files.

Every time i run the test its seemingly faster as windows automatically prioritizes the applications files (the 1000 txt documents) into faster portions of the disk.

I still have yet to figure out a fool proof way to force windows to place those files onto the ReadyBoost drive, because windows might be giving that space to something more important like drivers, or core system exe's.

Anyway, heres what i got so far, its a work in progress.
Notice the first 4 times running it there is a great improvement, then it seems to level off. I honestly believe this is SuperCache working fantastically.

*actual results in blue, and projected direction in white.
testingcachelh1.jpg


I had another hair brained theory which i will test sometime later, which is:
If a hard drive has faster seek times than a flash drive, then ReadyBoost may decide the HDD is a better place to put my files than the slow thumb drive.
But that theory was laid to rest when my 7200RPM c: drive was tested with the read test and only recieved a 0.45MB/s. Maby, i missed something that a old 5400rpm hdd will show me.

Another thing i learned is that ReadyBoost is suppose to supplement a lack of ram. So i tried starting a empty virtual machine and allocated it as much ram as i could give it. This tied up almost all of my ram and pegged the ram in use needle at max, then paused the virtual machine to free up the CPU but still hold the ram hostage.

But sadly, i still saw no change in my tests with and without ReadyBoost.

Its not over yet!!! :D
 
Lexar 512 Jump Drive no issues works flawlessly on RTM Vista Ultimate and Premium
 
You must log in or register to reply here.
Back
Top