4

I have a workstation class laptop running 64bit Windows-7 with 8GB-RAM (expandable to 16GB) and Dual SATA2 7200rpm disks.

I know my system will be making heavy use of the pagefile since I will be running Visual Studio 2008/2010 on the host machine while at the same time running up to two Linux Guests (For development) inside the latest VMWare Workstation 7.0.

Both disks are set as Dynamic Disks as follows:

                   320 GB 7200 rpm Disks (Volumes NOT to scale)
      +------------------------------------------------------------------------+
      | VOL_WIN7 |  VOL_VM_LINUX#1   | VOL_SWAP |  VOL_HOME                    |
disk1 | C:\      |  U:\              | Striped  |  W:\                         |
      | 50 GB    |  90 GB            | 12 GB    |  168 GB                      |
      +------------------------------------------------------------------------+
                                     |stripe-set|
      +------------------------------------------------------------------------+
      | VOL_APPS |  VOL_VM_LINUX#2   | VOL_SWAP |  VOL_DATA                    |
disk2 | D:\      |  V:\              | Striped  |  X:\                         |
      | 50 GB    |  90 GB            | 12 GB    |  168 GB                      |
      +------------------------------------------------------------------------+

I was wondering if there would be a performance advantage of dedicating my fixed size pagefile on the VOL_SWAP Software NTFS RAID-0 Stripe-Set and formatted using the 64K Allocation Unit Size.

Or should I avoid the Stripe-Set idea and just create two independent pagefiles.. one on each VOL_SWAP1 and VOL_SWAP2 partitions (Non Striped) and let the OS pick the swapfile it wants to use at any moment according to the internal Windows Kernel algorithms.. which is based on Least Active Drive (however that is determined)

I was going to put the pagefile volumes in the centre of the physical disks since i've read on Microsoft Engineering blogs that when it comes to pagefile performance.. Seek time is much more costly then the average read/write times when the pagefile is not fragmented.

the above is the Scenario desription for backround.. but my major question is:

Does NTFS Software RAID-0 (Stripe Sets) IMPROVE or REDUCE pagefile performance when swaping?

Any takers?

2 Answers 2

2

And, would this be hardware based raid or software?

The design that you have "sketched" above, really would not have any significant improvements with the RAID... If anything your making your configuration more complex then it needs to be. These are only 320 Gb drives...

This article, http://www.hardwaresecrets.com/article/394/1 Goes into benchmarking a similar setup, and basically burst read has a 50%+ increase, but sustained reading has only a marginal difference... Overall, for the sustained activities only a ~5% difference...

So, does RAID0 really increase disk performance? Yes, definitely. In our tests RAID0 doubled the average transfer rate of our hard disk drive.

How this disk performance increase reflects on overall system performance? It will depend on the kind of application you run: programs that make a lot of disk access will be the ones most benefited, of course. On PCMark05 this increase was between 4.44% and 8.82%, while on SYSmark2004 the best result was on Communication batch, where we saw a 15.25% performance increase. On this same program, Office Productivity performance was increased up to 6.52%, while we saw a performance increase up to 8.52% on 2D Creation batch.

2
  • It would "Software NTFS RAID-0 Stripe-Set".. the type of software raid that is available with NTFS Dynamic Disks as part of the disk manager. Commented Jan 19, 2010 at 3:44
  • Like Benjamin is saying, that's a pretty conveluted setup on only two drives. You're probably introducing more overhead than you're going to gain by doing anything with the swap location. If you want a fast Swap, get it it's own drive (physical). If you want it faster, get two phyiscal drvies and build a hardware RAID0. As soon as the swap is on a drive being used for other things, you're impeding it. Commented Jan 19, 2010 at 4:23
1

Based on this article and responses (see link below).. I get the impression that an NTFS Software RAID-0 Strip-Set is NOT necessary, and in fact, it would be better to just ensure I have 2 (or more) fixed pagefiles on seperate spindle volumes and let the OS internally stripe the pagefile writes accross both disks based on its own algorithms and demand.

http://www.pcmech.com/article/striping-your-swap-file-in-xp/

The last forum-answer (in above link) says he did tests to prove it works as claimed..

So in my above diagram I would make my VOL_SWAP1 and VOL_SWAP2 independant volumes and dedicate them to pagefiles.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .