We’ve moved!

You will be redirected to our new home in five seconds. If not, click here.

GigaOM Network: GigaOM | Earth2Tech | jkOnTheRun | NewTeeVee | OStatic | TheAppleBlog | WebWorkerDaily | Jobs Live Events | About | Contact

May 08, 2007

Windows ReadyBoost: how much does it help a Samsung Q1P UMPC?

CrystalmarkAfter yesterday's ReadyBoost post, I decided to truly put the feature to the test. Hey, what I perceive as a noticeable performance improvement might just be old age setting in, right? A little empirical analysis and a scientific approach might be in order here. To test the impact of ReadyBoost, I fired up the Vista partition on my Samsung Q1P, which has a 1 GHz Pentium M processor and 1 GB of RAM. For my ReadyBoost scenario, I added a 2 GB SanDisk Cruzer USB drive that labeled 'Enhanced for ReadyBoost'. Using the CrystalMark 2004R2 software application, I ran the benchmarks both with and without the ReadyBoost device.

Without ReadyBoost, the Samsung Q1P scored a 14873 overall; quite interesting since the new Q1 Ultra scored 13112 using the same software, which is almost a 14% overall difference, but I digress. ;)

With ReadyBoost, the same system scored 15663 overall, for a relative overall increase of about 5.3%. So where does ReadyBoost really provide the 'boost'? I did some reviewing of the granular level results and here's where CrystalMark sees the biggest boost:

HDD
Overall Benchmark without ReadyBoost: 2112
Overall with ReadyBoost: 2922 (+38%)

Random Read (64K) without ReadyBoost: 3.08 MB/s (123 CrystalMark points)
Random Read (64K) with ReadyBoost: 21.36 MB/s (854 CrystalMark points)

There's little to no significant difference in any of the other benchmarks; it's all in the system reads of data from the 'hard drive'. I put hard drive in quotes because the system is also reading data from the ReadyBoost cache when it can in that scenario; that's essentially what ReadyBoost is: a high speed cache.

Keeping this exercise in perspective, you may have ReadyBoost results that are better or worse than what I found. The experience will vary due to a number of variables: system specifications, speed of USB flash drive, etc... Overall, I think ReadyBoost is worth taking advantage of for the cost. You should be able to find a USB flash drive that's fast enough to use for as little as $20 or $30. Ideally, I subscribe to the thought that adding more RAM is the best investment you can make to speed up a system. However, with UMPCs we're limited to how much RAM can actually be recognized and used; for my Q1P, I'm maxed out at 1 Gigabyte, so ReadyBoost is an inexpensive and easy way to gain just bit more performance.

If you want the full results of my test in an HTML format, I've got them right here:

Enjoy this post? Receive more jkOnTheRun content for FREE by subscribing to the RSS feed!

Comments

My 8GB Corsair GT USB drive delivers 32MB per second reading. But then its $100.

I just won an eBay auction for a 32GB USB drive for $112. I'm curious to see if it actually has 32GB in it and how fast it is. It might just be usable as a SSD and skip the whole ReadyBoost feature.

I just picked up a Sony Vaio UX380N and a Sandisk Extreme III 2GB Memory Stick Duo. I'll compare data before and after ReadyBoost return later with the results. It'll be sometime in the next few days.

great info - ReadyBoost on the q1p would certainly be better if it were supported with the Compact Flash card. Using ReadyBoost as a USB stick is a bit cumbersome and problematic due to having to remember the darn thing and then having something stick out.

http://www.flickr.com/photos/39128821@N00/481963977/

I wonder why I'm getting a higher score. I'm using Vista Business.

who cares about ReadyBOB ?
just set bios to always boot from usb stick, or get IDE socket and SSD drive for 5 bucks.
Clearly this is another feature some 'architect' did not think over.

who cares about ReadyBOB ?
just set bios to always boot from usb stick, or get IDE socket and SSD drive for 5 bucks.
Clearly this is another feature some 'architect' did not think over.

people (like Kevin) keep taking stabs at the new Intel CPU's based on ONE benchmark. but they are forgetting the main reason for these new Intel chips, MUCH smaller package & use less power. this will allow Intel CPU's to compete with Via in the small market. it seems like people who own big UMPC's have the wrong mindset & just dont understand the goal of the new Intel CPU.

besides, i would never compare these benchmarks floating around of highly optimized enthusiasts machines versus 1 lousy default Q1u configuration.

Ready boost is only a high speed cache when it comes to random data access. Current flash is much much slower than a HDD for sequential 'stream' access. Also writes are much slower than reads. My understanding is that it is not really implemented as a general purpose disk cache, but rather for optimizing app startup and pre-fetch etc. I'm still not sure how this translates to real word performance.

On the Q1 I stopped using it before I could get a feel for it though. I find that hanging a USB dongle off the end of the UMPC not only looks odd, but increases my likelihood of me breaking something. Although I could be convinced to plug it back in if someone could show me that it saved battery life. :)

Frank, I'm sure we're not using the same USB flash drive, so that will have an impact: different read/write speeds, for example. I'm also running Vista Ultimate on a smaller drive partition.

a, I'm not sure what 'people who own big UMPCs' means, but I do understand and appreciate what the new Intel processors are meant to and what they compete against. I've drawn no conclusions on the new CPUs based on my one test above; I simply said it was interesting because it is to me. We've had a number of readers ask about the expected performance of Vista or the 600- and 800 MHz A1xx chips, so I pulled in the comparison. An Intel processor that consumes less power is good for the UMPC market and I've never said otherwise.

I've seen at least one other benchmark result on the web that puts the 800 MHz A1xx chips at the performance level of the Celeron processors. One thing is clear, better battery life means you must throttle the processor back.

If I had an SSD drive in my handtop, would ReadyBoost help?

Scotty: There's a 4GB limit on how much memory you can use for ReadyBoost, so that 32GB drive is probably overkill.

hola quiero el tema de nigga te quiero grasias chao alex

Comments are temporarily disabled for site maintenance and will return at 6 PM PDT.

 

RSS and Mobile-Friendly View

Contributors

Kevin C. Tofel

James Kendrick

Kevin's gear   JK's gear

Awards

Microsoft MVP Awardees

CNET100 2004Weblog Awards
2004ReadersChoice 2004_BoardOfExperts
Powered by TypePad
Member since 05/2004

Copyright Notice


  • Copyright 2008 The GigaOM Network. All rights reserved. The content in this RSS feed, as well as the content presented on the web pages of the blog, is provided for your personal non-commercial use only and may not be republished in whole or in part without the express written or verbal consent of the publisher. All rights are reserved.
StatCounter