Author Topic: Viking Modular SATADIMM Review - Benchmarks  (Read 1684 times)

Offline javajolt

  • Administrator
  • Hero Member
  • *****
  • Posts: 36146
  • Gender: Male
  • I Do Windows
    • windows10newsinfo.com
Viking Modular SATADIMM Review - Benchmarks
« on: November 17, 2010, 09:55:15 PM »
◄ To Review

Synthetic Benchmarks

We start out with IOMeter's 2MB sequential transfer test; this test is helpful in seeing how well the drives can hit their advertised transfer speeds.




In our sequential transfer tests, the Viking SATADIMM hangs with the SandForce-based OWC Mercury Extreme Pro 120GB and cruises past the Intel drive. The form factor may be different, but it looks like the SandForce performance is still there.




Switching up to a random transfer test still tells the same story. Viking's SATADIMM hangs tight with OWC's drive.






Once again the SATADIMM nips at OWC's drive's heels. Form factor notwithstanding, the drives are indistinguishable.






When we get to write latency we see a problem that has only materialized on a couple of odd SandForce drives: the average remains consistent between the two, but Viking's drive's peak latency is nearly ten times that of the OWC. It remains to be seen whether or not this will affect the SATADIMM in more real world use, but it's a red flag nonetheless. Neither is as bad as Intel's write latency, but this seems to be one case where SandForce implementations can vary fairly wildly.

Next we make a pit stop with CrystalDiskMark to test for peak performance.


CrystalDiskMark isn't as kind to the SandForce drives as IOMeter can be, but the SATADIMM and OWC drive nonetheless hang close together again, with the SATADIMM even posting slightly faster numbers.

Shifting back to IOMeter, let's see how the SATADIMM performs in simulated server workloads to see if that write latency rears its ugly head.
















Once again, the SATADIMM curls up right next to the OWC Mercury Extreme Pro. The two trade blows in certain disciplines, but essentially the drives are comparable.

Real-World Benchmarks

The synthetic tests have shown that regardless of where you stick a SandForce controller, as long as it's connected to the same SATA port it's going to perform pretty consistently. Still, it's worth seeing if that does translate to real-world usage scenarios.

For all of our real-world benchmarks we use custom StorageMark 2010 traces and try to work the drives the way a regular user might in varying scenarios.

The first trace is the “HTPC” test, which includes playing one 720p HD movie in Media Player Classic, playing one 480p SD movie in VLC, downloading three movies simultaneously through iTunes, and recording one 1080i HDTV stream through Windows Media Center over a fifteen minute period. Pay special attention to this test, because this is one of the situations where a DIMM-mounted SSD could be useful to the hobbyist.


Once again, the SATADIMM cuddles up to the OWC drive, even slightly beating it. This test stresses writes fairly well, and the SATADIMM's measured peak latency doesn't manifest here.

The second trace is the “Productivity” test, which consists of: a three hour period operating with 32-bit Vista running Outlook 2007 and connected to an Exchange server, browsing the internet with both Google Chrome and Internet Explorer 8, editing files in Office 2007, viewing PDFs in Adobe Reader, an hour of local music playback, and then two additional hours of streaming online music from Pandora. Again, this is another situation that the SATADIMM could see itself in, populating a RAM slot in a small form factor, low-power terminal.


In this test, the OWC drive takes an ever so slight lead on the SATADIMM, but the numbers are still close enough for the difference to be largely academic.

The third and final trace is the “Gaming” test; this test is almost entirely read-based, and consists of Windows 7 Ultimate 64-bit configured with Steam, and traces performance loading and running Grand Theft Auto 4, Left 4 Dead 2, and Mass Effect 2.


This is a situation where any potential weaknesses in write latency are easy to ignore; the gaming test is bound by reads, and here the SATADIMM produces a respectable lead over the other SandForce drives.

source:storagereview
« Last Edit: November 17, 2010, 10:06:23 PM by javajolt »