HighPoint 2720SGL RocketRAID Controller Review – Amazing 3GB/s Recorded With 8 Crucial C400 SSDs

Our recent RAID review of three Intel 520 SSDs on the Highpoint 2720SGL 6Gbps RocketRAID Controller opened our eyes as we didn’t think performance of 1.5GB/s could be reached with three SSDs.

Those results were just a taste of what was yet to come as we simply had to throw eight Crucial M4 256GB SATA 3 SSDs onto the RocketRAID for a true run. How does 3GB/s performance grab you for a start?  Unbelievable?  That’s what we thought as well but we did it!

INTRODUCTION

Putting aside the 3GB/s performance for just a second, lets head back a bit to our review of the very same HighPoint 2720SGL controller with the same SSDs and older firmware. The performance of the controller was much better than expected, however, it was something else that got our attention, this something being the popularity of the review with our readers.  This was evident in not only the amount of comments on the article as well as forum pm’s received, but also in the high number of e-mails received with respect to the article.

As a result, we made sure we had that ‘face to face’ with HighPoint at CES and were asked to test new drivers that they had for the RocketRAID, drivers that would bring about an incredible 50% performance increase!   An offer like this was simply to good to pass up.

M4-and-HighPoint

SIMPLICITY, PRICE AND PERFORMANCE

The long list of reasons why our first review garnered so much attention just doesn’t seem to end for HighPoint.  To say that such an amazing performance was achievable at such a low price was an understatement at the RocketRAID coming in at a mere $160, a price that saw many an enthusiast looking very closely at this card.  And then there was RAID 5 and RAID 6 configurable straight out of the box, this being an absolutely fantastic feature for many home users with large HDD arrays, specifically Windows Home Servers and FTP Servers.

Add to this the ability to simply plug this card into systems with older chipsets and users now have a simple upgrade to SATA 3 functionality and the RocketRAID is now a product suited for mainstream users as well as enthusiasts.  Having Crucial/Micron as a supporter has enabled us to throw in eight Crucial M4/C400 SATA 3 SSDs in order to highlight the tremendous performance potential of this card.

blank

HIGHPOINT 2720SGL CHARACTERISTICS

The HighPoint 2720SGL RAID card used for this report utilizes the manufacturers newest firmware revision, which drastically improves end performance.  A key consideration with this controller is that its slim form factor makes installation very simple and its small footprint and low heat signature make this a very unobtrusive device when installed into the host computer. A key benefit we see with such a low-heat solution is that it doesn’t require any complicated cooling measures and doesn’t result in unbearable heat within the system chassis.

Highpoint-FrontHighpoint-Back

The HighPoint controller uses a low powered ROC to boost the device through boot-up, to which the driver and the host system then perform the brunt of the computation. The card does not utilize any cache, other than the cache that is contained on the SSDs themselves.

The RocketRAID configuration sets a 64K stripe and which is not user configurable. Our configuration entailed write cache and NCQ being enabled and all 8 of the SSDs set to RAID 0.

With the promise of 50% higher speeds with Highpoints new RocketRAID drivers, there was bound to be some impressive performance.

8 comments

  1. blank

    Can anyone tell me the limits of SATA I aka 150mb/s
    I only need sequential read/write and randrom write 4kb

  2. blank

    This is pretty awesome. I’m having a chat with macperformanceguide.net, about whether this route would be better than just getting a PCI-E SSD equivalent. Like the RevoDrive 3.
    Are those just as likely to fail in Raid 0? as they are just onboard raided Sandforce SSD’s no?

    Any thoughts would be appreciated.

  3. blank

    Where can I download the new driver ???

  4. blank
    Fernando Martinez

    Guys, can you make the same, but using par drives on tests?

    Example: with 2 SSDs, then 4 (as RAID 0) ? Of course, no need 8 since it was already tested here.

    Thanks!

  5. blank

    That’s great and all, but HighPoint is the kind of company that exist in eastern europe. You know what I mean? You pay them for the product, and you assume all responsibility for support of that product, it going bad, not performing as advertised, and so on. There is no customer support with HighPoint. Actually, I just tried to contact them today and their online support (which is the only way to get in touch with them) script is not posting issues, it just redirects you to a blank post your issue screen. Just absolutely lousy company.

    So, if you ever consider silly highpoint, know that you won’t be getting support so you better be so good that you never need ask questions.

  6. blank
    John Joseph Mc Dermott

    This was an amazing card to begin with. I want one for my gaming rig. GTX 690 first though.

  7. blank

    What version of the raid controller firmware did you use for the benchmark ?

    I’m using RR272x v1.5 bios/fw, testing with 4 x 256 GB ADATA XPG SX900
    I get close to 500 Mbyte/s (read) per drive, which is great !

    But there seems to be something wrong, one out of 4 drives only reach 250 MB/s,
    even when moving them around, using the second port, in different combination,
    one is not going as fast as the others, but it is not the same SSD drive !

    Same if I put all 4 SSD on the same SAS port or 2 + 2 drives/SAS port

    (I’m testing with cables from 2 vendors, no diff)

    So it is not one faulty drive.

    After many reboots I sometimes can get all to perform equal,
    but after next reboot one is dropped down to half speed.

    I have verified that I’m using a slot @ PCIe 2.0 x8 speed (5GB/s).

    SSD fw is 5.02a (latest right now)

    Any advice ?

Leave a Reply

Your email address will not be published. Required fields are marked *