Jump to content


Photo

Samsung SSD 830 RAID Review Discussion


  • You cannot start a new topic
  • Please log in to reply
17 replies to this topic

#1 Kevin OBrien

Kevin OBrien

    StorageReview Editor

  • Admin
  • 1,439 posts

Posted 05 January 2012 - 11:51 AM

Today we look at the Samsung SSD 830 in RAID0, where it gets over 900MB/s in sequential transfer speed in IOMeter.

Samsung SSD 830 RAID Review

#2 mike2h

mike2h

    Member

  • Member
  • 258 posts

Posted 05 January 2012 - 04:50 PM

thx. like the layout & the drives u picked for comparison.
can u please tell me why u picked the 128 sector size? stil learning a lot of this stuff and sincve im going to be seeting up a raid 0 setup on my iwfes comp with 2 wd re4's a litle info would be helpful.

#3 Kevin OBrien

Kevin OBrien

    StorageReview Editor

  • Admin
  • 1,439 posts

Posted 05 January 2012 - 05:13 PM

128K stripe worked out best when it came to giving the best ground in our traces. It really depends on your workload though. Let me try to dig out an old forum thread where we covered stripe size versus real-world scenarios.

#4 mike2h

mike2h

    Member

  • Member
  • 258 posts

Posted 05 January 2012 - 06:43 PM

128K stripe worked out best when it came to giving the best ground in our traces. It really depends on your workload though. Let me try to dig out an old forum thread where we covered stripe size versus real-world scenarios.

thx! would appreciate that very much

#5 Shadmaster

Shadmaster

    Member

  • Member
  • 1 posts

Posted 06 January 2012 - 11:26 AM

While I find the data interesting and of worth to see what a RAID0 looks like on this product. I cannot see why or how you can justify comparing RAID'd drives with Single drives and call it a valid comparison. Shouldn't the comparison data be RAID'd drives as well to add any value or validity to the data you gathered. Otherwise the data on comparison is just muddying the review

#6 Kevin OBrien

Kevin OBrien

    StorageReview Editor

  • Admin
  • 1,439 posts

Posted 06 January 2012 - 01:15 PM

While I find the data interesting and of worth to see what a RAID0 looks like on this product. I cannot see why or how you can justify comparing RAID'd drives with Single drives and call it a valid comparison. Shouldn't the comparison data be RAID'd drives as well to add any value or validity to the data you gathered. Otherwise the data on comparison is just muddying the review


Generally speaking not many manufacturers send more than one drive in for review, making full comparisons between RAID options sparse in the same performance category. For that reason we will generally plot RAID results in with the same batch of drives we originally used when comparing that product, primarily to show how much performance changes compared to the original drive.

#7 mike2h

mike2h

    Member

  • Member
  • 258 posts

Posted 07 January 2012 - 02:58 PM

128K stripe worked out best when it came to giving the best ground in our traces. It really depends on your workload though. Let me try to dig out an old forum thread where we covered stripe size versus real-world scenarios.

i looked around a little bit & couldnt find it. if u give me general clue i will locate it. if it isnt in your archives that is...:)

#8 Kevin OBrien

Kevin OBrien

    StorageReview Editor

  • Admin
  • 1,439 posts

Posted 07 January 2012 - 11:02 PM

Just found it here

http://forums.storag...__1#entry264933

#9 mike2h

mike2h

    Member

  • Member
  • 258 posts

Posted 08 January 2012 - 02:11 AM

thx!

#10 undersea

undersea

    Member

  • Member
  • 1 posts

Posted 08 January 2012 - 08:51 PM

Does trim function in raid?

#11 h4lf

h4lf

    Member

  • Member
  • 24 posts

Posted 10 January 2012 - 05:11 AM

Would love to see steady state results for Samsung 830 single-drive and RAID. Would be very interesting as I heard garbage collection is quite slow in this model.

#12 Kevin OBrien

Kevin OBrien

    StorageReview Editor

  • Admin
  • 1,439 posts

Posted 10 January 2012 - 11:54 AM

Going to be gathering those results later this week when we are back in the office after CES. Around that time I will probably knock out whatever other controller/NAND combo that is also laying around.

#13 mike2h

mike2h

    Member

  • Member
  • 258 posts

Posted 11 January 2012 - 12:56 AM

what impact does slow garbage collection have? curious as this is one of my 3 ssd candidates.

#14 h4lf

h4lf

    Member

  • Member
  • 24 posts

Posted 11 January 2012 - 04:24 AM

"Slow" garbage collection really means that garbage collection is "delayed" until the drive is idle. Garbage collection is used to clean up the drive in order to maintain it at peak performance. By delaying garbage collection you can increase performance *right now* but if you leave it for too long then you get a build up of garbage and then the performance turns to rubbish, excuse the pun. The SR steady state tests illustrate this concept. (note: for a client/consumer drive this doesn't really matter unless you use your SSD for long sustained periods).

#15 mike2h

mike2h

    Member

  • Member
  • 258 posts

Posted 11 January 2012 - 02:33 PM

"Slow" garbage collection really means that garbage collection is "delayed" until the drive is idle. Garbage collection is used to clean up the drive in order to maintain it at peak performance. By delaying garbage collection you can increase performance *right now* but if you leave it for too long then you get a build up of garbage and then the performance turns to rubbish, excuse the pun. The SR steady state tests illustrate this concept. (note: for a client/consumer drive this doesn't really matter unless you use your SSD for long sustained periods).

thx!!
for me(average use) it wont be a 'problem'

#16 qubes

qubes

    Member

  • Member
  • 1 posts

Posted 20 January 2012 - 11:06 PM

While I read for only 2 pages into the threat on prior research on stripe size, I wonder if you can include some tests in the future on RAID0 SSDs. From my understanding of RAID0, the stripe size is usually kept larger on mechanical drives as it help keep the costs of seeks low when you have 2 spindles to precess incoming IO requests. When SSDs have no real penalty for random access patterns, wouldn't a stripe size on the order of the SATA interface be faster? ie, use a 4k stride.

I currently don't have a second SSD to test this against, so please take this into consideration the next time you have 2 identical drives to test against!

Cheers,

Q

#17 Kevin OBrien

Kevin OBrien

    StorageReview Editor

  • Admin
  • 1,439 posts

Posted 30 January 2012 - 01:10 AM

Well the way most RAID systems are designed, you only have a set number of stripe options (usually 64KB is smallest) to choose from. We have Crucial m4 RAID0 review upcoming that we can play around with the stripe sizes once the review is complete to see how things change in different areas.

#18 Bohemus

Bohemus

    Member

  • Member
  • 1 posts

Posted 01 February 2013 - 02:16 PM

Well the way most RAID systems are designed, you only have a set number of stripe options (usually 64KB is smallest) to choose from. We have Crucial m4 RAID0 review upcoming that we can play around with the stripe sizes once the review is complete to see how things change in different areas.

Interested to find out how you handled the lack of TRIM support when using these drives in a RAID configuration. Over time most reviews I have seen shows serious performance degradation without TRIM.



0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users