The Test

To keep the review length manageable we're presenting a subset of our results here. For all benchmark results and even more comparisons be sure to use our performance comparison tool: Bench.

Motherboard: ASUS P8Z68-V Pro (Intel Z68)
ASUS Crosshair V Formula (AMD 990FX)
Intel DX79SI (Intel X79)
Hard Disk: Intel X25-M SSD (80GB)
Crucial RealSSD C300
Memory: 4 x 4GB G.Skill Ripjaws X DDR3-1600 9-9-9-20
Video Card: ATI Radeon HD 5870 (Windows 7)
Video Drivers: AMD Catalyst 11.10 Beta (Windows 7)
Desktop Resolution: 1920 x 1200
OS: Windows 7 x64

Cache and Memory Bandwidth Performance

The biggest changes from the original Sandy Bridge are the increased L3 cache size and the quad-channel memory interface. We'll first look at the impact a 15MB L3 has on latency:

Cache/Memory Latency Comparison
  L1 L2 L3 Main Memory
AMD FX-8150 (3.6GHz) 4 21 65 195
AMD Phenom II X4 975 BE (3.6GHz) 3 15 59 182
AMD Phenom II X6 1100T (3.3GHz) 3 14 55 157
Intel Core i5 2500K (3.3GHz) 4 11 25 148
Intel Core i7 3960X (3.3GHz) 4 11 30 167

Cachemem shows us a 5 cycle increase in latency. Hits in L3 can take 20% longer to get to the core that requested the data, if this is correct. For small, lightly threaded applications, you may see a slight regression in performance compared to Sandy Bridge. More likely than not however, the ~2 - 2.5x increase in L3 cache size will more than make up for the added latency. Also note that despite the large cache and thanks to its ring bus, Sandy Bridge E's L3 is still lower latency than Gulftown's.

Memory Bandwidth Comparison - Sandra 2012.01.18.10
  Intel Core i7 3960X (Quad Channel, DDR3-1600) Intel Core i7 2600K (Dual Channel, DDR3-1600) Intel Core i7 990X (Triple Channel, DDR3-1333)
Aggregate Memory Bandwidth 37.0 GB/s 21.2 GB/s 19.9 GB/s

Memory bandwidth is also up significantly. Populating all four channels with DDR3-1600 memory, Sandy Bridge E delivered 37GB/s of bandwidth in Sandra's memory bandwidth test. Given the 51GB/s theoretical max of this configuration and a fairly standard 20% overhead, 37GB/s is just about what we want to see here.

Overclocking Windows 7 Application Performance
Comments Locked

163 Comments

View All Comments

  • yankeeDDL - Monday, November 14, 2011 - link

    I'm with you xpclient.
    I will never understand Microsoft's fanboys. Why do they expect the OS to make such huge impact on the benchmarks?
    It's like for motherboards: you can differentiate for ease of use, stability, features, supported hardware ... but the benchmarks, will substantially be the same.
    XP trails Windows7 on multi core because it was never designed to support so many cores, and MS has no interest in updating it.
  • Per Hansson - Monday, November 14, 2011 - link

    Hi, what resolution and in game graphical settings where used for the World of Warcraft gaming test?
    It always amazes me how well that game scales with super high-end CPU's, but if it's at just a silly low resolution it does not really matter, so, which is it? (This isn't mentioned in "bench" either....)
  • Per Hansson - Monday, November 21, 2011 - link

    Here is Anand's reply to this question incase anyone cares ;)
    ---
    1680 x 1050, 4X AA, all detail settings maxed (except for weather) :)

    Take care,
    Anand
  • Mightytonka - Monday, November 14, 2011 - link

    Typo on the second page. Should read RST (Rapid Storage Technology), right?
  • Ocire - Monday, November 14, 2011 - link

    Nice review! :-)

    If you want to test PCIe bandwidth, you could use the bandwidth-test that comes with the CUDA SDK. It's easy to setup, and you can also test configurations with multiple GPUs. You should get quite reliable results for pure PCIe Gen.2 performance with that.
    It would be really interesting to get some performance numbers for PCIe as that is the bottleneck in quite a few GPU-computing scenarios.

    Cheers!
  • LancerVI - Monday, November 14, 2011 - link

    Sounds like a great, enthusiast proc, married to a mainstream chipset at enthusiast prices.

    That means 'no joy' for me.

    I guess I'll be hanging on to my little i7 920 that could for a bit longer. Going on 4 years now. That's unheard of for me!!!
  • hechacker1 - Monday, November 14, 2011 - link

    I'm also going to hang onto the i7 920. Overclocked the x58 platform can still compete with the best of of Sandy Bridge for almost any workload. Sure we're missing some IPC and power enhancements, but nothing worth spending serious cash on.

    What I'm looking at now is the Gulftown prices. I'm hoping they come down from the $1000 Extreme part, and perhaps we get an affordable 6-core chip for the X58 platform.

    I'd be happy to stick with Gulftown until we see affordable 8 core parts, or major IPC improvements.
  • Makaveli - Monday, November 14, 2011 - link

    the 980 non X version is already going for $550

    This is your next upgrade, as is mine i'm also on a 920.

    I doubt that price will drop any lower.
  • hechacker1 - Monday, November 14, 2011 - link

    Yeah I'm going to be watching the prices closely for a while.

    At $550 for the lowest end Gulftown on newegg, it's still not affordable for me.

    It's still cheaper than the intro prices of Sandy Bridge-E including a new motherboard though.

    I'll have to watch forums closely for people wanting to sell their chips, I imagine you can snag one used for a good deal.
  • davideden - Monday, November 14, 2011 - link

    I currently have a Core i7 2600K LGA 1155 processor. I am assuming that I won't be able to use this with the new LGA 2011 socket on the new Sandy Bridge E motherboards. Will there be any cheaper processors in the near future that are at the price point of the Core i7 2600K that are compatible? I was disappointed with not being able to utilize triple channel with my memory or being able to use all my sticks of ram with the current LGA 1155 motherboards. The quad channel ram along with the 8 slots have me most excited for the new platform as I do video editing/motion graphics/3D work. Thanks!

Log in

Don't have an account? Sign up now