Benchmarking Performance: CPU System Tests

Our first set of tests is our general system tests. These set of tests are meant to emulate more about what people usually do on a system, like opening large files or processing small stacks of data. This is a bit different to our office testing, which uses more industry standard benchmarks, and a few of the benchmarks here are relatively new and different.

PDF Opening

First up is a self-penned test using a monstrous PDF we once received in advance of attending an event. While the PDF was only a single page, it had so many high-quality layers embedded it was taking north of 15 seconds to open and to gain control on the mid-range notebook I was using at the time. This put it as a great candidate for our 'let's open an obnoxious PDF' test. Here we use Adobe Reader DC, and disable all the update functionality within. The benchmark sets the screen to 1080p, opens the PDF to in fit-to-screen mode, and measures the time from sending the command to open the PDF until it is fully displayed and the user can take control of the software again. The test is repeated ten times, and the average time taken. Results are in milliseconds.

System: PDF Opening with Adobe Reader DC

Opening PDFs using Adobe is significantly single threaded, hence why Intel takes the win here for the most part - the 7400 is slightly lower on frequency, enough for the 4.0 GHz Ryzen parts to push ahead.

FCAT Processing

One of the more interesting workloads that has crossed our desks in recent quarters is FCAT - the tool we use to measure stuttering in gaming due to dropped or runt frames. The FCAT process requires enabling a color-based overlay onto a game, recording the gameplay, and then parsing the video file through the analysis software. The software is mostly single-threaded, however because the video is basically in a raw format, the file size is large and requires moving a lot of data around. For our test, we take a 90-second clip of the Rise of the Tomb Raider benchmark running on a GTX 980 Ti at 1440p, which comes in around 21 GB, and measure the time it takes to process through the visual analysis tool. 

System: FCAT Processing ROTR 1440p GTX1080 Data

Similarly, FCAT is single threaded as it analyzes frame-by-frame. The extra frequency helps Intel here.

3D Particle Movement v2.1 

This is the latest version of the self-penned 3DPM benchmark. The goal of 3DPM is to simulate semi-optimized scientific algorithms taken directly from my doctorate thesis. Version 2.1 improves over 2.0 by passing the main particle structs by reference rather than by value, and decreasing the amount of double->float->double recasts the compiler was adding in. It affords a ~25% speed-up over v2.0, which means new data. 

System: 3D Particle Movement v2.1

DigiCortex 1.20

Despite being a couple of years old, the DigiCortex software is a pet project for the visualization of neuron and synapse activity in the brain. The software comes with a variety of benchmark modes, and we take the small benchmark which runs a 32k neuron/1.8B synapse simulation. The results on the output are given as a fraction of whether the system can simulate in real-time, so anything above a value of one is suitable for real-time work. The benchmark offers a 'no firing synapse' mode, which in essence detects DRAM and bus speed, however we take the firing mode which adds CPU work with every firing.

System: DigiCortex 1.20 (32k Neuron, 1.8B Synapse)

Agisoft Photoscan 1.0

Photoscan stays in our benchmark suite from the previous version, however now we are running on Windows 10 so features such as Speed Shift on the latest processors come into play. The concept of Photoscan is translating many 2D images into a 3D model - so the more detailed the images, and the more you have, the better the model. The algorithm has four stages, some single threaded and some multi-threaded, along with some cache/memory dependency in there as well. For some of the more variable threaded workload, features such as Speed Shift and XFR will be able to take advantage of CPU stalls or downtime, giving sizeable speedups on newer microarchitectures.

System: Agisoft Photoscan 1.0 Total Time

Because parts of Photoscan are very multithreaded, such as the first stage, there's plenty of scope for the Ryzen CPUs to pull ahead here.

 

 

 

 

Benchmarking Suite 2017: CPU and GPU Benchmarking Performance: CPU Rendering Tests
Comments Locked

254 Comments

View All Comments

  • SkipPerk - Wednesday, May 3, 2017 - link

    These are low-end CPU's. People use those for gaming and web-surfing. I have a proper Xeon machine at work like a normal person. Not to mention, you reference video software. What tiny percentage of computer users ever own or use video software? That is a tiny industry. It reminds me of the silly youtube reviews where the reviewer assumes everyone is editing videos, when less than one percent of us will ever do so.

    Most people buying non-Xeon CPU's really will be using basic software (MS Office, WinZip,...) or games. The only time I have used non-Xeon CPU's for work was when I had software that loved clock speed. Then I got a bunch of 6-core's and overclocked them (it was funny to watch the guys at Microcenter as I bought ten $1k CPUs and cheesy AIO water coolers). Otherwise one uses the right tool for the job.
  • AndrewJacksonZA - Tuesday, April 11, 2017 - link

    On the last page, "On The Benchmark Results"
    "Looking at the results, it’s hard to notice the effect that 12 threads has on multithreaded CPU tests."
    Don't you mean that it's NOT hard to notice?
  • Drumsticks - Tuesday, April 11, 2017 - link

    I didn't see the 7600k in gaming benchmarks, was that a mistake/not ready, or is it on purpose?

    Thanks for the review guys! This new benchmark suite looks phenomenal!
  • mmegibb - Tuesday, April 11, 2017 - link

    I was disappointed not to see the i5-7600k in the gaming benchmarks. Perhaps it wouldn't be much different than the i5-7600, but I have sometimes seen a difference. For my next build, it's looking like it's between the 1600x and the 7600k.
  • fanofanand - Tuesday, April 11, 2017 - link

    "Platform wise, the Intel side can offer more features on Z270 over AM4"

    Aside from Optane support, what does Z270 offer that AM4 doesn't?
  • MajGenRelativity - Tuesday, April 11, 2017 - link

    Z270 has more PCIe lanes off the chipset for controllers and such that AM4 does not
  • fanofanand - Tuesday, April 11, 2017 - link

    I won't disagree with that, but I'm not sure a few extra pci-e lanes is considered a feature. Features are typically something like M.2 support, built-in wifi, things like that. The extra pci-e lanes allows for MORE connected devices, but is a few extra pci-e lanes really considered a feature anymore? With Optane being worthless for 99.99999% of consumers, I'm just not seeing where Z270 gives more for the extra money.
  • JasonMZW20 - Tuesday, April 11, 2017 - link

    Let's do a rundown:

    Ryzen + X370
    20 (3.0) + 8 (2.0)
    Platform usable total: 28

    Core i7 + Z270
    16 + 14 (all 3.0)
    Platform usable total: 30

    Intel's Z270 spec sheet is a little disingenuous, as yes it does have a maximum of 24 lanes, but 10 are reserved for actual features like SATA and USB 2.0/3.x. 14 can be used by a consumer, giving you a total of 2 NVMe x4 + 1 NVMe x2 leaving x4 for other things like actual PCIe slots. That 3rd NVMe slot may share PCIe lanes with a PCIe add-in slot, if configured that way.

    Ryzen PCIe config (20 lanes): 1x16, 2x8 for graphics and x4 NVMe (or x2 SATA when NVMe is not used)

    Core i7 config (16 lanes): 1x16, 2x8, or 1x8+2x4 for graphics

    They're actually pretty comparable.
  • mat9v - Tuesday, April 11, 2017 - link

    No, not more PCIEx lines, those from chipset are virtual, they all go to CPU through DMI bus that is equivalent to (at best) 4 lines of PCIEx 3.0. All those chips (Intel and AMD) offer 16 lines from CPU for graphic card, but Zen also offers 4 lines for NVMe. Chipsets are connected by DMI (in Intel) and 4 lines of PCIEx 3.0 (in AMD), so that is equal, now Intel from those DMI lines offer virtual 24 lines of PCIEx 3.0 (a laugh and half) while AMD quite correctly offers 8 lines of PCIEx 2.0 (equivalent to 4 lines of PCIEx 3.0).
  • psychobriggsy - Wednesday, April 12, 2017 - link

    Indeed. If a user is going to need more than that, they're more likely going to be plumping for a HEDT system anyway. AMD's solution is coming in a bit, but that should be able to ramp up the IO significantly.

Log in

Don't have an account? Sign up now