Office Performance

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

Agisoft Photoscan – 2D to 3D Image Manipulation: link

Agisoft Photoscan creates 3D models from 2D images, a process which is very computationally expensive. The algorithm is split into four distinct phases, and different phases of the model reconstruction require either fast memory, fast IPC, more cores, or even OpenCL compute devices to hand. Agisoft supplied us with a special version of the software to script the process, where we take 50 images of a stately home and convert it into a medium quality model. This benchmark typically takes around 15-20 minutes on a high end PC on the CPU alone, with GPUs reducing the time.

Agisoft PhotoScan Benchmark - Total Time

The different stages of Photoscan have different preferences for frequency and threads, but all the results are in Bench. In this case having threads matters, with the old 8-thread FX CPU barely getting ahead of the Core i3 parts. The i5 takes the lead, showing that having physical cores helps with cache management. This is further reinforced by our results staircase, which put the i3-6320 and i3-6300 at roughly equal timings but the i3-6100 almost 4% behind.

Cinebench R15

Cinebench is a benchmark based around Cinema 4D, and is fairly well known among enthusiasts for stressing the CPU for a provided workload. Results are given as a score, where higher is better.

Cinebench R15 - Single Threaded

Cinebench R15 - Multi-Threaded

Cinebench eats threads, high IPC and high frequency for breakfast. In the single thread test, the high Intel IPC shines through, and our i3 parts sit in unison with AMD behind by up to 40%. In multithreaded land, the 6/8 thread FX processors go ahead of the i3s as expected, and our staircase slightly deviates for the i3-6100 showing that L3 cache creep is slowly coming in.

HandBrake v0.9.9: link

For HandBrake, we take two videos (a 2h20 640x266 DVD rip and a 10min double UHD 3840x4320 animation short) and convert them to x264 format in an MP4 container.  Results are given in terms of the frames per second processed, and HandBrake uses as many threads as possible.

HandBrake v0.9.9 LQ Film

HandBrake v0.9.9 2x4K

The high IPC of the Skylake parts makes a difference for the smaller frame conversion, while threads come into play for the larger resolution frames. In both cases, the regular staircase shows a lack of issues with the L3 cache differences, but it is interesting to see the X4 845 hot on the heels for the high resolution frames despite its cache arrangement. The FX-6350 sits on par with the i3-6100, showcasing the difference between a six-thread much older processor and a four thread latest process part.

Hybrid x265

Hybrid is a new benchmark, where we take a 4K 1500 frame video and convert it into an x265 format without audio. Results are given in frames per second.

Hybrid x265, 4K Video

Hybrid is similar to HandBrake, and we again have a regular staircase.

Performance Comparison: Real World Performance Comparison: Linux
Comments Locked

94 Comments

View All Comments

  • Samus - Tuesday, August 9, 2016 - link

    ^This.

    I love WinRAR, but even a blind test will demonstrate 7zip to be faster, especially at compression. Decompression is often storage limited unless you have a good RAID or PCIe SSD.
  • plopke - Monday, August 8, 2016 - link

    I bought a core i3 6300 for my parents but only paid the price of a i3 6100 around 109 euro, the price gap is so big for such a small performance increase some vendors decided just to dump them out of their inventory.
  • ShieTar - Monday, August 8, 2016 - link

    Nice review. Any chance that the Pentiums and maybe even Celerons will be added to the comparison?

    Everybody reviews the i5&i7, but it's really hard to get any feel on how fast (or slow) the very cheapest Intel CPUs really are.
  • fanofanand - Tuesday, August 9, 2016 - link

    I second this, I see a lot of inexpensive devices running low-end Intel parts and it sure would be nice to see how they stack up.
  • jaydee - Monday, August 8, 2016 - link

    Any attempts at trying Quicksync with the Skylake i3?
  • AndrewJacksonZA - Wednesday, August 10, 2016 - link

    +1 on the Quicksync please.
  • elbert - Monday, August 8, 2016 - link

    The i3's should have been removed from the desktop line up a long time ago. Quad cores come out over 10 years ago and Intel is still trying to sale dual cores for the desktop. Its really a testament to how tight a hold Intel's monopoly is on the PC business. Currently the hold that is killing the PC business.
  • DanNeely - Monday, August 8, 2016 - link

    The persistence of dual core hardware says more about the failure of most software to benefit from larger numbers of cores. Unless you've got at least 3 big threads running in parallel the extra physical cores of the i5/7 only serve to drive up manufacturing costs and power consumption.
  • FunBunny2 - Monday, August 8, 2016 - link

    -- The persistence of dual core hardware says more about the failure of most software to benefit from larger numbers of cores.

    the number of embarrassingly parallel user space applications continues to be stuck at a couple.
  • Ratman6161 - Monday, August 8, 2016 - link

    But, for all those that are just web surfing, reading email, and typing Word documents, how much parallelism do they need? A while back I refurbished a couple of Core 2 duo laptops from about 2009. Upgraded them from 2GB to 4BB of cheap ram and slapped an el-cheapo SSD in place of the old 5400 RPM disk drives and they felt like new machines. For the people doing the sorts of tasks I mentioned above, they are more than good enough - even an i3 is overkill for the task.

Log in

Don't have an account? Sign up now