AES-NI: Much Faster Encryption & Bitlocker Performance

Westmere (and thus Clarkdale) adds some new instructions to x86, although the big expansion comes with AVX and Sandy Bridge next year. Westmere gets six new encryption/decryption instructions. The group of instructions accelerate AES (Advanced Encryption Standard) and are thus referred to as AES-NI.

Many businesses require that all corporate PCs have the contents of their hard drives encrypted through the use of software like Bitlocker that comes with Microsoft Windows 7. These sorts of applications can be accelerated by AES-NI and to show the potential benefit I dug up a test I did while preparing for one of our SSD articles several months ago.

If you turn on Bitlocker in Windows 7 there's an immediate and measurable impact to performance. Disk performance generally drops by a noticeable amount and actual application usage performance drops by a smaller amount. Every write to the disk has to be encrypted first so there's some obvious CPU overhead. Clarkdale should reduce that overhead significantly as the common encryption operations are now hardware accelerated.

To test this I ran three tests. I first ran PCMark Vantage's HDD suite on my Windows 7 testbed SSD (an OCZ Summit) on a Core i5 661, then encrypted the drive using Bitlocker and ran the same test on the same processor. For the last test I swapped out the i5 661 for a Lynnfield based Core i5 750 (no AES-NI) and re-ran the HDD test. The results below were quite promising:

Processor PCMark Vantage HDD % of Unencrypted Performance
Clarkdale - Unencrypted 16713  
Clarkdale - Bitlocker Encryption 13785 82.5%
Lynnfield - Bitlocker Encryption 11744 70.3%

 

There's a definite benefit to Clarkdale's AES-NI instructions. There's still a performance hit from enabling Bitlocker, but it's not nearly as great as on Lynnfield and other architectures that don't have AES-NI support. With a smaller decrease in I/O performance from enabling full disk encryption, there's also a smaller hit to application performance as well. This is huge for corporate desktops/notebooks.

Most of those machines aren't quad-core encoding monsters; they use dual-core processors. The upgrade from Core 2 seems like it'd be worth it, or at least AES-NI will probably keep AMD out of the running for consideration.

Intel HD Graphics: A Lot Better ASUS Saves the Day: Simulated Core i3s & The Test
POST A COMMENT

93 Comments

View All Comments

  • Taft12 - Monday, January 4, 2010 - link

    The parent's office PC's aren't bottlenecked by the OS - they're not bottlenecked PERIOD. They run modern productivity apps just fine and would gain little to no benefit from Core i3 (or Windows 7 for that matter). Reply
  • Paulman - Monday, January 4, 2010 - link

    Those office PC's you mentioned aren't bottlenecked by the 2GB of RAM. But I wouldn't say that they aren't bottlenecked, "period". What they ARE bottlenecked by is disk I/O, I'm sure. Throw in a good SSD and you would notice quite a bit of speed improvement, and probably a noticeable difference between the 1.6GHz and 2.4GHz machines.

    The most annoying thing to me whenever I'm using my PC is seeing and hearing my laptop HDD thrash around when launching an app or what not, because everything is held up as a result. Yes, I know it's a laptop HDD, but desktop drives are pretty slow, too.
    Reply
  • FlyTexas - Monday, January 4, 2010 - link

    SSDs are indeed fast, and make the whole computer feel "snappier"...

    However, these office machines never shut down (they hibernate overnight). IE8, Word, Excel, and Acrobat are always open and always stay open. Once loaded in memory, the hard drive is hardly used.

    I've looked at upgrading them to 3GB of RAM, but they aren't using what they have, so why bother? Most of them use right around 1GB of RAM most of the time.

    Could we put 40GB SSDs in? Sure, they are about $130 at Newegg right now... Not the end of the world, until you multiply that times 24 machines. Not a minor expense.
    Reply
  • FlyTexas - Monday, January 4, 2010 - link

    That is so true. This the first time in a long time that the computers have been "fast enough" for everything we use them for.

    There was a time in 1993/1994 that we were in this position, running DOS 6 and Windows 3.1 on 486DX2/66 machines, where the move to the DX4/100 or Pentium saw no benefit until Windows 95 came out. I worked in small shop back then, and we demoed a Pentium 66 machine, and saw zero benefit over the 486DX2/66 machines, other than it cost twice as much.

    Perhaps in 2002, the Athlon XP machines were "fast enough" for Windows XP and Office XP, that was a nice time as well in the business. A Pentium III 550mhz was my last personal Intel chip until 2006, when I got my first Core2Duo machine at home. I had to work with some Pentium 4s at work during that time, Intel really, REALLY dropped the ball with the Pentium 4, IMHO.

    Oh well... I've been doing this a long time, I still remember 5.25" floppy drives, with NO hard drive and those ugly green monitors with Hercules graphics... :)
    Reply
  • lowlight - Monday, January 4, 2010 - link

    But the 45nm package on Westmere doesn't just carry the GPU. They also moved the PCI-E controller and Memory controller there. I guess the "CPU" is still technically 32nm, but compared to Nehalem, half the "CPU" actually resides on a 45nm package on the chip...

    You can see a diagram in this Clarkdale review: http://www.hardcoreware.net/intel-clarkdale-core-i...">http://www.hardcoreware.net/intel-clarkdale-core-i...
    Reply
  • lowlight - Monday, January 4, 2010 - link

    Guess I should have read the whole review... You guys picked it up too! Not many others did though ;) Reply
  • ilnot1 - Monday, January 4, 2010 - link

    I swear I've scoured the pages but I don't see your Test System Setup Chart: how much RAM, which graphics card? If it is there and I missed it I wish you could delete posts. Reply
  • Spoelie - Monday, January 4, 2010 - link

    It's on page 6

    I'd like to know the setup of each memory benchmark on page 2. What memory speeds and settings were used for the latency and bandwidth numbers?
    Reply
  • Anand Lal Shimpi - Monday, January 4, 2010 - link

    All of the CPUs used DDR3-1333 running at 7-7-7-20 timings for that test. I used Everest 1909 (I believe, I'm about 2300 miles away from my testbed right now :-P) and CPU-Z's latency tool to grab the data.

    Take care,
    Anand
    Reply
  • toyota - Monday, January 4, 2010 - link

    I was looking for it too and its not there. Reply

Log in

Don't have an account? Sign up now