qasdfdsaq

Seek times / Burst rates with different IDE controllers

15 posts in this topic

Does anyone know which is the best IDE/IDE-RAID controller availible? I've run several different tests with several different controllers and consistently my RAID controller (even without running a RAID array) is scoring better access times and burst rates with the same drives than standard onboard controllers...

Share this post


Link to post
Share on other sites

From the reviews I've read, it seems 3ware is the best in terms of driver performance (better/smoother performance in concurrent disk access) and Highpoint is the best in raw RAID-0 performance (best read/write throughput). 3ware is also better than Promise in RAID-5 performance.

Tech-report has a pretty good review on IDE RAID controller last month you might want to check out.

Share this post


Link to post
Share on other sites

Promise RAID controllers (as well as just the IDE controllers) are ATA133, so they'll have higher burst rates than the ATA100 controllers on-board. Haven't heard anything about the better seek times when using RAID controller though.

Cheers,

Jan

Share this post


Link to post
Share on other sites
Promise RAID controllers (as well as just the IDE controllers) are ATA133, so they'll have higher burst rates than the ATA100 controllers on-board.

This really only makes a difference if you're putting two drives on a channel, which is a great way to suboptimize the array's performance.

Share this post


Link to post
Share on other sites
Promise RAID controllers (as well as just the IDE controllers) are ATA133, so they'll have higher burst rates than the ATA100 controllers on-board.

This really only makes a difference if you're putting two drives on a channel, which is a great way to suboptimize the array's performance.

I think even with a single drive, when you have a cache hit, the transfer rate should be close to interface speed.

Share this post


Link to post
Share on other sites

Sechs: I was talking about burst speed (aka cached transfer), just as docinthebox (and also the topic :wink:) mentioned. SATA should have 150MB/s burst, if the drive is native SATA.

Jan

Share this post


Link to post
Share on other sites

Not even close. I haven't yet seen an single ATA drive that could fill the entire bandwidth of a ATA/66 connection with a burst, let alone sustained.

Share this post


Link to post
Share on other sites

By the way, if you're using a VIA motherboard with a VIA southbridge and you're looking for good burst speeds, be sure to apply the RAID performance patch:

http://www.viaarena.com/?PageID=66

Read this article for more details:

http://www.tecchannel.de/hardware/817/index.html

Apparently, there's a bug in VIA's PCI implementation such that the max throughput of the PCI bus is far less than the theoretical bandwidth of 132MB/s.

Share this post


Link to post
Share on other sites
Not even close.  I haven't yet seen an single ATA drive that could fill the entire bandwidth of a ATA/66 connection with a burst, let alone sustained.

What about this thread then? (sorry bit old :wink:) Chefanoz is reporting 73.2MB/s on read burst speed.

Jan

Share this post


Link to post
Share on other sites
Not even close.  I haven't yet seen an single ATA drive that could fill the entire bandwidth of a ATA/66 connection with a burst, let alone sustained.

What about this thread then? (sorry bit old :wink:) Chefanoz is reporting 73.2MB/s on read burst speed.

Jan

or this one:

http://www.xbitlabs.com/storage/ibm-120gxp/

Scroll down to the graph labelled "Read Burst Speed, KB/sec".

At UDMA/33, the IBM Deskstar 120GXP, Maxtor D740X-6L, WD1000BB, WD800BB and Seagate Barracuda ATA IV have read burst speeds of 86613, 86103, 85706, 85450 and 68817 KB/sec respectively.

Share this post


Link to post
Share on other sites
At UDMA/33, the IBM Deskstar 120GXP, Maxtor D740X-6L, WD1000BB, WD800BB and Seagate Barracuda ATA IV have read burst speeds of 86613, 86103, 85706, 85450 and 68817 KB/sec respectively.

Sorry I meant UDMA/100.

Share this post


Link to post
Share on other sites

You can really see ATA33 cap there... ATA100 specs say that one must sustain the 100 MB/s for 3 ms... so the amount of data transferred during the burst is not much (0,3 MB) so it will fit in the cache. Thanks for the link, docinthebox.

Jan

Share this post


Link to post
Share on other sites

Hmm, actually the time is much less than 3ms. More likely 30 µs... I can't find the correct data to verify. Anyone else?

Jan

Share this post


Link to post
Share on other sites
Not even close. I haven't yet seen an single ATA drive that could fill the entire bandwidth of a ATA/66 connection with a burst, let alone sustained.

hmm....

image004.gif

my two 180gxp's can sustain 58MB/sec pretty well... i'd say that IS close

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