Tom's Hardware Wants You: CPU Benchmarks 2011/12

Almost a year ago, I started a thread just like this one to elicit your thoughts on the way we test CPUs. Based on your responses, we made a number of changes to our test suite. Of course, we weren't able to accommodate every request. But each suggestion that made sense received careful consideration.

Well, it's that time again. We want to know how you want to see processors evaluated as we alter the benchmark suite.

As before, I think we have a reasonable mix of gaming-, productivity-, media-, and synthetic-based tests in our automated CPU software package. With that said, our ears are always open.

Now, keep in mind that we’d prefer to use free/trial/open source software, making reproducibility an option for as many of our readers as possible.

Here’s our current list of metrics. We aren't doing much with OpenCL or DirectCompute right now, preferring to save those tests for the graphics suite. With that said, feature like Quick Sync and AMD's Fusion initiative merit a reconsideration about folding in benchmarks with unique support for architecture-specific optimizations. This is something we want to be very careful with, though, as to not bias the weight of our evaluation toward any one company. Keep that in mind as we open up the dialogue to accelerated apps.

  • Apple iTunes
  • Lame MP3
  • HandBrake CLI
  • MainConcept Reference v.2
  • WinRAR
  • WinZip
  • 7-Zip
  • Autodesk 3ds Max
  • Adobe Premiere Pro CS5
  • Adobe After Effects CS5
  • Adobe Photoshop CS5
  • Cinebench
  • Blender
  • ABBYY FineReader 10
  • PCMark 7
  • 3DMark 7
  • SiSoftware Sandra 2011
  • Metro 2033
  • F1 2010
  • Aliens Vs. Predator
  • Just Cause 2

Occasionally you’ll see us swap out a game, exclude a more workstation-oriented app in a mainstream, or add something like Fritz, but that’s the lineup I’m using as of…well, now. So, give us some suggestions on what you’d like to see over the next year. Bear in mind that the apps we pick from the comments you leave will likely be the ones most applicable to the largest number of people, though there is room for a handful of more fringe tests, too.

Thanks again for the input. I’ll be stopping by the comments section here, of course. And if you want to reach me directly, you're always welcome to hit me up on Twitter.

Chris Angelini
Worldwide Editor-in-Chief, Tom’s Hardware

Chris Angelini
Chris Angelini is an Editor Emeritus at Tom's Hardware US. He edits hardware reviews and covers high-profile CPU and GPU launches.
  • kingkstur
    StarCraft 2! I used to see it in some benchmarks, but it slowly got faded out. =
    Reply
  • wintermint
    Can you guys add more programs that benefit from hyper-threading? It will be easier to refer a CPU to new system builders if there's benchmarks in front of our face :)
    Reply
  • cuecuemore
    kingksturStarCraft 2! I used to see it in some benchmarks, but it slowly got faded out. =Yep, this is a great test for single-core strength along with the original Crysis. (and Warhead??) AvP is a terrible CPU benchmark; on my X4 965 it locks all 4 cores at 40%, regardless of whether I'm running one or two GTX 460s. For more proof, let's look at what Tom's had to say :) :
    http://www.tomshardware.com/reviews/best-gaming-cpu-core-i3-2100-phenom-ii-x6-1075t,2859-5.html
    In my opinion, BF:BC2 is the gold standard for CPU scaling at the moment, and will probably be supplanted by BF3.

    +1 to SiSoftware Sandra 2011, this is the only synthetic I use, but it's awesome.
    Reply
  • echdskech
    I've been using Tom's CPU charts since it came out to spec out not only desktops but also entry level servers built on enthusiast parts. Most of them run server software on Linux though so I'd like to cast a vote for the Phoronix test suite.
    Reply
  • genghiskron
    The cpu benchmark would be a joke without battlefield 3.
    i dont know if this applies to the cpu benchmark suite itself, but i think it would be interesting to come up with an appropriate "background program suite" for the gaming benchmarks.
    basically, there was a huge debate over whether the i3-2100 would outperform the phenom ii quads in a real-world environment because a real user runs a number of programs while gaming. id like to see this tested in some way. i think this will also be an interesting benchmark in the context of the upcoming bulldozer release, should that happen before the end of the world.
    Reply
  • dragonsqrrl
    My only request would be to add in some sort of Maya rendering (software or mental ray) or simulation performance bench. I know you already include Cinebench and Blender, but performance and scaling can vary quite significantly between these programs. And amongst pro's and hobbyists alike, Maya is one of the most widely adopted 3D content creation programs out there. If you have a content creation section in your bench suite, Maya just seems like a must.

    And I know you guys are probably already considering this, but adding BF3 (when it's released) to your gaming section would be great as well.
    Reply
  • Matlab, taking advantage of multi-core processors and GPU processing
    Reply
  • Soma42
    JacobiMatlab, taking advantage of multi-core processors and GPU processing
    Agreed I'd love to see benchmarks for MATLAB, Solidworks or more CAD programs
    Reply
  • I suggest to simulate real-life experience simulation rather than synthetic performance benchmark. So we should just use Apps and Games. Should also add video editing/converting tool (e.g. mov to rmvb ..etc). Games should use either BF3 or MW3. Others like Guild Wars 2, latest Need for Speed & NFL/NBA/FIFA series, follow by either Civ5/StarCraft2/Diablo3. Not sure if there's any capable benchmark for Flash/HTML5 games, but it will be good as well to check Civ World (Facebook) and others.
    Reply
  • bystander
    wintermintCan you guys add more programs that benefit from hyper-threading? It will be easier to refer a CPU to new system builders if there's benchmarks in front of our face
    That's a good idea. All previous gaming articles on the subject show no difference, but Resident Evil 5, and Metro 2033 definitely showed improvement with and without on my system. Those can't be the only two.
    Reply