Gemini Lake SFF PC Showdown: Intel's June Canyon (NUC7PJYH) and ECS's LIVA Z2 Reviewed
by Ganesh T S on December 20, 2018 8:00 AM ESTNetworking and Storage Performance
Networking and storage are two major aspects which influence our experience with any computing system. This section presents results from our evaluation of these aspects in the Intel NUC7PJYH. On the storage side, one option would be repetition of our strenuous SSD review tests on the drive(s) in the PC. Fortunately, to avoid that overkill, PCMark 8 has a storage bench where certain common workloads such as loading games and document processing are replayed on the target drive. Results are presented in two forms, one being a benchmark number and the other, a bandwidth figure. We ran the PCMark 8 storage bench on selected PCs and the results are presented below.
There is really not much to choose from between the three systems - our choice of storage drives (Crucial BX300) was the same for both the Gemini Lake units, and that is reflected in the above graphs.
On the networking side, we restricted ourselves to the evaluation of the WLAN component. Our standard test router is the Netgear R7000 Nighthawk configured with both 2.4 GHz and 5 GHz networks. The router is placed approximately 20 ft. away, separated by a drywall (as in a typical US building). A wired client is connected to the R7000 and serves as one endpoint for iperf evaluation. The PC under test is made to connect to either the 5 GHz (preferred) or 2.4 GHz SSID and iperf tests are conducted for both TCP and UDP transfers. It is ensured that the PC under test is the only wireless client for the Netgear R7000. We evaluate total throughput for up to 32 simultaneous TCP connections using iperf and present the highest number in the graph below.
In the UDP case, we try to transfer data at the highest rate possible for which we get less than 1% packet loss.
Despite all the WLAN modules being 1x1, the June Canyon NUC has much better numbers. In part, this is due to the availability of antenna diversity in the AC 9462 (which allows the better of two antennae to be utilized for communication).
Facebook
Twitter
RSS
59 Comments
View All Comments
mode_13h - Friday, December 21, 2018 - link
I think the main reason for Intel doubling last-level cache vs. Apollo Lake is all the cheapo systems using this in single-channel mode.IntelUser2000 - Thursday, December 20, 2018 - link
"Doubling the internal cache has led to significant performance increase in many real-life workloads."Come on. It's not due to cache. Goldmont Plus cores in Gemini Lake has substantial architectural improvements. Doubled caches are usually responsible for maybe 5% increase in performance.
shelbystripes - Friday, December 21, 2018 - link
Ummm... that’s not really a valid assumption. Sure, if the system already has enough cache, adding more cache will not substantially increase performance.But the cache size is actually small enough to restrain performance (which can happen with these smaller, lower-cost parts). The “doubling” here is going from 2MB to 4MB L2 cache, which for the quad-core designs compared here, means effectively from 0.5MB per core to 1MB of L2 cache per core.
That sounds like a lot of L2 cache, until you realize there’s no L3 cache. That’s it, 0.5-1MB per core of last-level cache, and then you’re going to system RAM.
Is there even an Intel Core CPU made today with only 0.5MB of last level cache? Those tend to have only 256KB of L2, but then at least 1MB of L3 per core. That’s enough cache that adding more cache won’t help you much. Given the smaller, simpler design of Atom, I’m not surprised going up to 1MB of L2 cache per core would yield substantial performance benefits.
Brunnis - Friday, December 21, 2018 - link
Goldmont Plus has substantial architectual enhancements that are much more likely to account for the lion’s share of the performance increase. The article makes it seem Goldmont Plus is mainly about larger L2, which is a bit misleading. See this link:https://en.wikichip.org/wiki/intel/microarchitectu...
Brunnis - Friday, December 21, 2018 - link
Even smaller compute heavy benchmarks perform 20-30% faster, which is usually not the case for a mere L2 size increase (I’ve never seen that, at least).mode_13h - Friday, December 21, 2018 - link
The performance impact of cache is highly workload-dependent. However, it does sound like there are some significant improvements:https://en.wikichip.org/wiki/goldmont_plus#Key_cha...
Smell This - Thursday, December 20, 2018 - link
So ...How many tens of billions of dollars has Chipzilla spent subsidizing the 'Next Units' and Atom 'Fails'?
The 'new' NUCs are not, really, all that. An AMD Ryzen V1000 SoC mini-ITX FP5 BGA at 12/14nm would 'Temash' the Atom at 10-12w.
ZOLTAC ... make it so.
Death666Angel - Thursday, December 20, 2018 - link
I'd take some more AM4 mSTX motherboards. There are quite a few Intel ones, but the beefier iGPU for AMD would make for a more well rounded system.Alien88 - Saturday, December 22, 2018 - link
Check out the Udoo Bolt...LMonty - Friday, December 21, 2018 - link
Hello Ganesh, could you pls. confirm whether the NUC operates in dual channel mode when using 32GB of RAM? I saw one review on Amazon complaining that his J5005 NUC was running in single channel mode, when using 2x8GB sticks (16GB total).