CPU Office Tests

The office programs we use for benchmarking aren't specific programs per-se, but industry standard tests that hold weight with professionals. The goal of these tests is to use an array of software and techniques that a typical office user might encounter, such as video conferencing, document editing, architectural modelling, and so on and so forth.

All of our benchmark results can also be found in our benchmark engine, Bench.

Chromium Compile (v56)

Our new compilation test uses Windows 10 Pro, VS Community 2015.3 with the Win10 SDK to combile a nightly build of Chromium. We've fixed the test for a build in late March 2017, and we run a fresh full compile in our test. Compilation is the typical example given of a variable threaded workload - some of the compile and linking is linear, whereas other parts are multithreaded.

Office: Chromium Compile (v56)

Having redone our compile testing, we can see that the new Ryzen-2000 series parts do provide a good uplift over the first generation, likely due to the decreased cache latencies and better precision boost. Performance per dollar between the 8700K and the 2700X would seem to be about equal as well.

PCMark8: link

Despite originally coming out in 2008/2009, Futuremark has maintained PCMark8 to remain relevant in 2017. On the scale of complicated tasks, PCMark focuses more on the low-to-mid range of professional workloads, making it a good indicator for what people consider 'office' work. We run the benchmark from the commandline in 'conventional' mode, meaning C++ over OpenCL, to remove the graphics card from the equation and focus purely on the CPU. PCMark8 offers Home, Work and Creative workloads, with some software tests shared and others unique to each benchmark set.

Office: PCMark8 Creative (non-OpenCL)Office: PCMark8 Home (non-OpenCL)Office: PCMark8 Work (non-OpenCL)

PCMark 10

Office: PCMark10 Extended Score (Overall)

GeekBench4

Office: Geekbench 4 - Single Threaded Score (Overall)

Office: Geekbench 4 - MultiThreaded Score (Overall)

If you live and breathe GeekBench 4, then the single threaded results put Intel firmly in first place. For the multi-threaded tests, the top Intel and AMD mainstream parts are going at it almost neck-and-neck, however it is clear that the previous generation quad-cores are falling behind.

Benchmarking Performance: CPU Encoding Tests Benchmarking Performance: CPU Legacy Tests
Comments Locked

545 Comments

View All Comments

  • jor5 - Thursday, April 26, 2018 - link

    Pull this shambles and repost when you've corrected it fully.
  • mapesdhs - Monday, May 14, 2018 - link

    Not an argument. It is just as interesting to learn about how and why this issue occured, to understand the nature of benchmarking. Life isn't just about being spoonfed end nuggets of things, the process itself is relevant. Or would you rather we don't learn from history?
  • peevee - Thursday, April 26, 2018 - link

    When 65W i7 8700 is 15% faster in Octane 2.0 than 105W Rizen 7 2700x, it is just sad.

    Of course, the horrible x64 practically demands than compilers must optimize for a very specific CPU implementation (choosing and sorting instructions in the code accordingly), AMD could have at least realized the fact and optimize their own implementation for the same Intel-optimized code generators...
  • GreenReaper - Thursday, April 26, 2018 - link

    Intel compilers and libraries tend not to use the ideal instructions unless they detect a GenuineIntel signature via CPUID - it'll likely use the default lowest-common-denominator pathway instead.

    TDP is more of a guideline - it doesn't determine actual power usage (we've seen Coffee Lake use way more than the TDP), let alone the power used in a particular operation. Having said that, I wouldn't be surprised if Intel were more efficient in this particular test. But it'd be interesting to know how much impact Meltdown patches have in that area; they might well increase the amount of time the CPU spends idle (but not idle enough to go into a sleep mode) as it waited to fetch instructions.
  • SaturnusDK - Thursday, April 26, 2018 - link

    Compare power consumption to blender score. Ryzen is about 9% more power efficient.

    TDP is literally Thermal Design Power. It has nothing to do with power consumption.
  • peevee - Thursday, April 26, 2018 - link

    "TDP is literally Thermal Design Power. It has nothing to do with power consumption."

    Unless you have invented a way to overcome energy conservation law, power consumed = power dissipated.
  • SaturnusDK - Friday, April 27, 2018 - link

    It's a guideline for cooling solutions. Look at the power consumption numbers in this test for example.

    Ryzen 2700X power consumption under full load 110W.
    Intel i7 8700K power consumption under full load 120W.

    Both are at stock speeds with the Ryzen having 8 cores versus 6 cores, and scoring 2700X 24% higher Cinebench scores. Ryzen is rated at 105W TDP so actual power consumption at stock speed is pretty close. The 8700K uses 120W so it's pretty far from the 95W TDP it is rated at.
  • ijdat - Saturday, April 28, 2018 - link

    The 8700 also uses 120W so it's even further from the 65W TDP it's rated at. In comparison Ryzen 2700 uses 45W when it has the same rated 65W TDP. I know which one I'd prefer to put into a quiet low-power system...
  • mapesdhs - Monday, May 14, 2018 - link

    Perhaps this is AMD's biggest win this time round, potent HTPC setups.
  • peevee - Thursday, April 26, 2018 - link

    "Intel compilers "

    What Intel compilers have to do with it?

Log in

Don't have an account? Sign up now