Hades Canyon NUC Review (NUC8i7HVK)

Performance and Benchmarks

I installed Windows 10 Enterprise version 1803 on the NUC without much effort and made sure it was up-to-date. The graphics driver version 23.20.792.2048 and BIOS version 0037 were used for all the following benchmarks. The system was equipped with a 2×8 GB DDR4-3200 RAM modules (HyperX HX432S20IBK2/16) and an AData XPG SX7000 M.2 NVMe drive.

Benchmarks




3DMark

3DMark is a popular benchmarking suite that benchmarks video and gaming performance of the computer. See below for the results and a comparison to the Skull Canyon as well as the Kaby Lake NUCs.

Basically the Hades Canyon NUC wipes the floor with the other contestants here. There’s almost no point in comparing the Hades Canyon to the previous NUCs.

Cinebench R15

Cinebench runs 3 separate benchmarks and gives us figures that are comparable. First a simple 3D car chase that measures mainly the GPU (OpenGL) performance. The result is in frames per second. Second there’s a rendering of 3D model with all cores. This stresses purely the CPU. Finally there’s another rendering of the same model, this time using just a single CPU core.

In the OpenGL test the Hades Canyon beats all the previous NUCs comfortably. In the pure CPU tests we can see that the i7 CPU in the Skull Canyon isn’t that much slower than the one here.




Geekbench 3

The Geekbench score is mainly dependant on the CPU and here we can see that the older Skull Canyon does not do much worse than the new Hades Canyon.

PassMark CPU test

The score for PassMark CPU performance test was 11359.

SteamVR Performance Test

Since Intel advertises the NUC8i7HVK as a VR capable NUC I gave the SteamVR Performance Test a shot. According to the test the device is on the higher end of “Capable” edging towards the “Ready” territory. The software does recommend an upgrade for the GPU though (which of course is not going to happen on a NUC).

Gaming

Ok, so we know the thing performs in the benchmarks, but can I actually play some recent games with it? I’m not a big gamer, so it took a while to decide which games to try. In the end I tried to look for some recent games to give an idea how well the thing behaves. Resolution was set to 1920×1080 in each game.

Need for Speed Payback

I started by playing a bit of Need for Speed Payback, a racing game released end of last year. While the game itself is pretty disappointing for someone who remembers the original Need for Speed from the 1990s I was happy to find out that the NUC itself had no problems running it at full HD resolution and high detail levels. When measuring the framerate with FRAPS I got a min/avg/max of 50/59.8/71.

Far Cry 5

The latest incarnation of Far Cry is amazingly beautiful and in addition to supporting HDR it also has a nice benchmark option that makes it easy to provide comparable performance data. Even when I yanked all the details to high, the frame rate min/avg/max was 44/51/65. The game generates a nice report of the benchmark, have a look at it by clicking the image above.

Battlefield 4

BF4 is not one of the latest games, but a very popular one and common in benchmarking on various web sites. Thus I decided to take it in so you can have an idea. With ultra settings I was able to reach a min/avg/max of 49/64.85/83 on the scene I tried.

RAM Speed

As I did not get my HyperX DDR4-3200 RAM (HX432S20IBK2) before I got the unit, I initially ran some benchmarks with HyperX DDR4-2400 RAM (HX424S14IBK2). Note that all benchmark results above are ran with the 3200 MHz RAM.

Benchmark DDR4-3200 Result DDR4-2400 Result Difference
3DMark Cloud Gate 25988 26207 -0.8%
3DMark Sky Diver 23374 23351 0.0%
3DMark Fire Strike 8521 8519 0.0%
CineBench CPU 854/182 854/182 0.0%
CineBench OpenGL 135.05 132.34 2.0%
GeekBench 4 Multi-Core 16795 16350 2.7%

The results reveal quite well that there’s not much, if any, advantage to the faster RAM at least when it comes to the benchmarks that I did try. Unlike other NUCs the GPU does have its own HBM2 RAM in this case, so faster RAM does not mean faster GPU RAM in this case.

Thoughts on Performance

Running benchmarks on the Hades Canyon is a bit of a surreal experience after benchmarking countless of slower NUCs. The 3DMark Fire Strike benchmark is actually not a slide show, who knew? But yes, the performance numbers are out of the NUC world. It actually performs better in all benchmarks than my couple-years-old large desktop PC with i5-4690k and Geforce GTX 960.

It was really refreshing booting up a game, selecting 1920×1080 resolution and setting the details all the way to the max and still ending up with a smooth gameplay! There’s really no NUC that I’ve tested before that comes close. To be honest, as I’ve never really been much of a gamer myself, this is probably the fastest computer I’ve ever owned.

It does also support AMDs FreeSync technology even if I was unable to test it.

26 Responses

  1. Peter says:

    Hello!

    Did you test it with a 4K/HDR/10bit video with an usually used local video player like DS Player or MPCHC+MadVR or PowerDVD?

    I’ve read elsewhere, that it has troubles with HDR video passthrough (matter of the IGP maybe).

  2. Any update on the release of the new NUC8 with i3?

  3. Jacob says:

    Can you please measure noise and power consumption during gaming?

    • Olli says:

      It’s a bit difficult to take any generic gaming measurements as different games stress the GPU and the CPU in so different ways and even different scenes in games cause different load. Thus the fan speeds up and down during the gameplay. Expect something between the “Prime95 only” and “Prime95 + 3DMark” during gaming.

  4. Mark Ernst says:

    It seems like your benchmarks are a bit off comparing to mine. I’ve not done the wattage readings for Prime but I’ve measured x264 and x265 playback on FHD in VLC, Netflix and Kodi, and my readings are off (to the lower side) by at least 3 watts, even on idle without powering on to S3 or S4 state I’m reading about 1 to 2 watts instead of 5. Did you upgrade your bios before benching?

    • Olli says:

      It’s very much possible that my meter is off. It’s by no means industrial grade – it’s a consumer grade power meter sold with a price tag of approx. €20. I was using BIOS version 0037.

  5. Waldek says:

    Thanks for the review !
    Has anybody seen a good comparison between the NUC8i7HVK and NUC8i7HNK, i.e. 100W vs 65W models, esp. regarding noise ? My use-case is homelab servers (Vmware/Hyper-V) for education purposes, so not so much interested in the added CUs of the more powerful model, but thermals and noise very important…

  6. Huang says:

    I only care about bean Canyon

  7. Cristian says:

    The nuc’s capabilities with 4K/HDR/10bit video through MPCHC+MadVR, like Peter suggested, is interesting and greatly appreciated to know the results from.

  8. hello what type of memory card have you used on your nuc

  9. Keith Walker says:

    A big +1 for Skull Canyon is that my Linux install was seemless (due in part to lack of GPU), disappointing but not surprised that Hades Canyon video drivers are not ready. This is the achilles heel of Linux- awful video driver support

  10. Andrew S says:

    Beautiful NUC but appears almost impossible to run ESXi on it.

    • ZophiasDad says:

      This seems to be addressed in the latest BIOS update HN0051.bio I have this NUC and have considered downgrading my BIOS to v HN0034 because I think in v 37 they removed Legacy Boot from the BIOS. I was having some issues with my USB 3 ports and support had me flash the BIOS to I think it was v50 and none of my pen drives with Linux installers (brand of pen drive did not matter and they all worked in my Dell Precision M6800 laptop) were even recognized as being plugged in but they were before updating my BIOS.
      info from HN0051 BIOS https://downloadmirror.intel.com/28287/eng/HN_0051_ReleaseNotes2.pdf

      New Fixes/Features:
      •Fixed the issue where an error would occur when installing VMware* ESXi versions 6.5 and 6.7.
      •Fixed the issue that the Intel NUC NUC8i7HNK wouldn’t allow some Linux distributions to install with Secure
      Boot turned off.
      •Fixed the issue where Windows* Task Manager wouldn’t show memory speed correctly

  11. Ben Sanford says:

    I’m trying to see how this model compares with the somewhat similar NUC8i7BEH

    • Mark Ernst says:

      Back with some more crappy news. The LEDs are working fine but they have frozen in mode. No breathing, no config is being load, nothin’. Already picked it up with Intel Support but as of yet I’ve yet to find a solid solution. Performance is solid, nothing wrong there.

      So, for anyone willing to buy this product partially because of the LEDs, don’t bother.

    • nucblognet says:

      NUC8i7BEH is much cheaper, different form factor, about equally performing CPU and much worse GPU. Read my review of NUC8i7BEH from yesterday for benchmarks comparing the two ..

  12. Tigerman82 says:

    Out of curiosity and because I’ve ordered this PC (HVK), why do you recommend the HyperX DDR4-3200 RAM if the results seem to indicate that the difference between it and DDR4-2400 is almost non-existent? I’m only asking because (looking at best EU Amazon deals) the best price on regular Crucial 2x8GB DDR4-2400 (seems to be a popular and cheap choice for this PC in EU) is around €50 cheaper than the best price of HyperX DDR4-3200. So there is a significant price difference. Don’t get me wrong, I do want to get everything out of this PC (productivity+light 1440p gaming), but all the online tests and advice seem to indicate that the performance difference between value RAM running at motherboard’s stock speed and performance RAM running at highest supported speed is minimal.

    • nucblognet says:

      Hi there! In my testing the DDR4-3200 was -0.7% to 2.7% faster than DDR4-2400. I’d say it doesn’t make a difference – if you’ve got a good offer for one or the other, go for that.

      • Tigerman82 says:

        Thanks for the reply. Decided HyperX isn’t worth it (price difference to value RAM too big). However, G.Skill RAM is more competitively priced so might go for G.Skill DDR4-3000 instead of Crucial value RAM.

  13. JohnOrion says:

    Recently got this PC and have come to love it. I’m impressed with how quiet it can be as it is virtually silent after I disabled Turbo Boost and set Fan Control Mode to ‘Quiet’. However, while I mostly do productivity stuff that doesn’t push the Hades Canyon that much, I do occasionally game a little bit. Is it safe to use ‘Quiet’ mode (set at 75 degrees, I believe) instead of the default ‘Balanced’ mode (55 degrees)? I’m thinking long-term here and do not want to cause damage to the hardware.

    I’ve checked HWMonitor a few times right after a gaming session and the highest temp I’ve seen the Hades Canyon hit on ‘Quiet’ is 86 degrees. When gaming, I can faintly hear the blowers but it’s still a lot quieter than, say, a PS4. I’d certainly like to continue using ‘Quiet’ but am unsure if it’s safe in the long run. Thoughts?

  14. Ted says:

    I really wish I could find power consumption stats on the NUC8i7HNK version somewhere, but unfortunately every single reviewer went for the high-end unit. 🙁

    • Ted says:

      I’m in a power-constrained situation, off-grid on solar, so a sane NUC power consumption is critical. Despite the NUC8i7HVK’s watt requirements being so ridiculous for a supposedly 100W TDP CPU/GPU package, I decided to give the 65W TDP NUC8i7HNK a chance.

      Fortunately I can report that its power consumption, with a 15.6″ 4K monitor, is right in line with the 4K XPS 15 and x360 15 2-in-1 laptops with the same chipset. e.g. — 80W-90W under gaming load (Skyrim 1080p, High). I can live with that. 🙂

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.