Sign in to follow this  
blue_heart71

Low burst rate with Hitachi 7200rpm notebook drive

Recommended Posts

Hi all,

I just purchased Hitachi 100gb 7200rpm for my Fujitsu N6420. The burst rate using HDtune benchmatk is very disappointing (around 50mb) while it has to be at least 100mb. any clue what could be the reason?

The Fujitsu 200gb 4200 drive that came with the laptop has higher burst rate (around 74mb/s) but average tranfer rate and access time is lower than the Hitachi.

I have tried SATA controller driver from Intel and Fujitsu's website and got the same result.

The notebook has 2 drive bay and drive both drives in different bays and the results are the same.

Run DFT and no errors.

Run Feature tools from Hitachi which supports SATA but it is reading both hard drives as ATA and when clicking on change SATA option it says the drive is not SATA drive!!! Is that Hitachi's software problem or there is something wrong in the notebook itself?

Frankly speaking, I did not notice any speed increase between both drives.

Is my Hitachi drive faulty?

Share this post


Link to post
Share on other sites

Is the drive's (write) cache enabled via device manager? Are you able to show us the current low level drive configuration, via "hdparm -i" or other similar utility? Are you seeing any other issues aside from burst rate?

Thank you much,

Frank Russo

Share this post


Link to post
Share on other sites

Thanks for your feedback

Write cache is enabled as per hdtune and feature tools by Hitachi.

I am not sure that I got you properly about your question regarding drive configuration. Is there a program called hdparm -i that I can doanload and install to get that info? Do you have a link for that program?

There is no other issue beside the low burst rate. The average transfer rate is 38mb/s, and access time is 15ms.

I forgot to add that I am using Windows Vista home premium 32 bit

Share this post


Link to post
Share on other sites

I just searched for the hdparm, it is Linux program and wont work on Vista :(

However, I am attaching screen shoots from both HDtune and HDtach for both drives (Hitachi 7200 and Fujitsu 4200 rpm)

In the benchmarks below, notics the low vurst rate of the Hitachi

This is HDtune's score for both drives

78110272.G1PKaZO8.HDtunescore.jpg

This is HDtach's score for both drives

78110299.IsME5qJz.HDtachscore.jpg

Share this post


Link to post
Share on other sites

Yes the burst rate is a little low as reported by HDtune, but surprisingly it is just fine (about 100MB/sec) as reported by HDtach.

Share this post


Link to post
Share on other sites

I just got an email from Hitachi as an aswer for my question for them that the feature tool is reading both SATA drives as UDMA 5, and this is what Hitachi answered me:

"If Feature Tool is reporting both drives running in UDMA mode 5, then your BIOS is set to a compatibility mode so the machine thinks the drives are really IDE drives. Because the BIOS is reporting the drives as IDE, Feature Tool will not allow you to change SATA settings"

I think I need to contact Fujitsu to verify what Hitachi is saying

Will keep you updated

Share this post


Link to post
Share on other sites
"If Feature Tool is reporting both drives running in UDMA mode 5, then your BIOS is set to a compatibility mode so the machine thinks the drives are really IDE drives. Because the BIOS is reporting the drives as IDE, Feature Tool will not allow you to change SATA settings"

Why would UDMA 5 cause a bottleneck of any kind on a single drive that's mechanically only capable of about 55MB/sec tops. I don't see anything to be worrying about there.

Like I said before, the hdtach reported burst rate of 99MB/sec just fine. Even if you were able to increase that from 99 MB/sec to 990 MB/sec then I'll bet it wouldn't make one bit of difference in real world performance.

Share this post


Link to post
Share on other sites

As far as I know, the burst speed will affect the speed of transferring the data from harddrive's cash to the bus, so according to HDtune, instead of moving the data between the interface of the drive and the bus is 50mb/s instead of the normal 100mb and this will affect the performance to some extent although I am not sure how much.

However, I am not sure how reliable is HDtune, as I only found it being used by notebookforum.com users. Any idea about this program?

Share this post


Link to post
Share on other sites

I would recommend you grab a third bmark program and validate the results toy have. I think Sandra has a burst rate benchmark. If 2 out of 3 mbenches say the burst rate is 90MB/sec, then you should be all set. Unfortunately, I am unable to see the images above.

Frank

Share this post


Link to post
Share on other sites

Graphs and burst rate look fine for both drives. A burst rate benchmark usually works by creating a small file and reading it >1 time. It uses the average of runs 2,3,and 4 to calculate the burst rate. If the file is too small, you get an inaccurate result. If it is larger than the read cache available on the HDD, you get a low result.

So, we might be able to assume that the file that hdtune writes does not fit into the read cache of the drive, or that another windows process (like system restore) read/wrote to the drive during the benchmark, bumping the benchmark file from cache.

All in all, I wouldn't worry about it. HDTach indicates that the burst speed is fine.

Thank you for your time,

Frank Russo

Share this post


Link to post
Share on other sites

Thanks everyone,

I have run Sisoft and Atto and got wierd result. As you can see from the attached images, the Hitachi has higher reading score than the Fujitsu while Fujitsu has higher writing degree and its write score is higher than its reading score, so I am confused. However, the way I am analyzing HDtune's benchmark (burst speed), is that it measures both the reading and writing speed and posts the average, and it could be true for the HDtach as well, but this is just an assumption.

Can some sheds some light on Atto's beanchmark and how to read the score?

The images will reveal what I am analyzing, I am attaching the new benchmark in addition to the old benchmarks and a separate link with every image just in case the image is not showing in the message

So what do you think?

http://www.pbase.com/shg2/image/78212705

78212705.f8Faz8r6.SisoftHitachifujitsu.jpg

http://www.pbase.com/shg2/image/78211457

78211457.XBjD8qGv.attohitachifujitsu.jpg

http://www.pbase.com/shg2/image/78110299

78110299.IsME5qJz.HDtachscore.jpg

http://www.pbase.com/shg2/image/78110272

78110272.G1PKaZO8.HDtunescore.jpg

Share this post


Link to post
Share on other sites

That's interesting blue_heart. It really seems as if the drive is simply not using (or only minimally using) the on-drive cache for writes. Perhaps this is a deliberate decision by the authors of the firmware. Is it possible that they profiled the performance and decided it was more efficient to only use the on-drive cache for reads. I don't know so hopefully someone with more knowledge could answer this question.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this