Original Link: https://www.anandtech.com/show/11229/ecs-liva-z-duallan-apollo-lake-ucff-pc-review
ECS LIVA Z Fanless Dual-LAN Apollo Lake UCFF PC Review
by Ganesh T S on March 30, 2017 8:00 AM ESTThe market for UCFF (ultra-compact form factor) PCs has seen tremendous growth since the introduction of the Intel NUCs in early 2013. Various motherboard and system vendors have their own offerings in this space. ECS started selling their own UCFF PCs under the LIVA brand in 2014. The initial focus was on the low end, with units based on the Intel Bay Trail and Braswell platforms. The lineup evolved to accommodate Core M-based units as well as systems in the mini-STX form factor. Today, we are taking a look at the ECS LIVA Z. This fanless Apollo Lake system targets the low-end market segment, and aims to differentiate itself by supporting two LAN ports.
Introduction and Product Impressions
Traditional UCFF PCs (such as the BRIX from GIGABYTE and the Beebox from ASRock) have stayed close to the Intel NUC form factor. ECS had experimented with a smaller form factor (Mini-Lake) in the first-generation LIVA, as well as the LIVA X, and LIVA X2. The LIVA Z, however, adopts a board form-factor similar to that of the Intel NUCs.
The feature set and pricing of the LIVA units make it target the developing and cost-sensitive markets. The LIVA Z is no different. Intel's Apollo Lake SoCs improve upon Bay Trail and Braswell by adopting a newer microarchitecture (Goldmont) for the CPU cores and also getting fabricated in a more power-efficient / mature 14nm process. In the consumer market, the Goldmont cores are exclusive to the Apollo Lake family. The SoCs target the netbook and nettop markets, and it is no surprise that ECS has adopted the netbook variants for the fanless LIVA Z models.
The LIVA Z comes in three variants,corresponding to the three members of the Apollo Lake mobile SoC family (6W TDP) - the Pentium N4200, or the Celeron N3450, or the Celeron N3350. Our review sample, the LIVA-ZN33 is based on the Intel Celeron N3350. The unit ships with 32GB of eMMC on the board, but, it was insufficient to enable the installation of all our benchmark programs. Therefore, we evaluated the PC in two different configurations - the base one with Windows 10 installed on the eMMC and used for typical workloads such as media playback, and a configuration with a M.2 SATA SSD for office workloads. The full specifications of our review configurations are summarized in the table below.
ECS LIVA-ZN33 Specifications | |
Processor | Intel Celeron N3350 Apollo Lake (Goldmont), 2C/2T, 1.1 GHz (Turbo to 2.4 GHz), 14nm, 2 MB L2, 6W TDP |
Memory | Kingston CBD16D3LS1KBG/4G DDR3 4GB 11-11-11-28 @ 1600 MHz |
Graphics | Intel HD Graphics 500 |
Disk Drive(s) | SanDisk eMMC DF4032 (32 GB; eMMC v5.0-compatible) ADATA Premier SP600 SP600NS34 (128 GB; M.2 Type 2242 SATA III; MLC) |
Networking | Intel Dual Band Wireless-AC 3165 (1x1 802.11ac - 433 Mbps) 2x Realtek RTL8168/8111 Gigabit LAN |
Audio | 3.5mm Audio Jack, 1x Digital Microphone Capable of 5.1/7.1 digital output with HD audio bitstreaming (HDMI) |
Display | 1x HDMI 1.4a 1x mini-DisplayPort 1.2 |
Miscellaneous I/O Ports | 3x USB 3.0 Type-A 1x USB 3.0 Type-C |
Operating System | Retail units come barebones, or with OS depending on SKU. Our unit was the non-OS version, but we installed Windows 10 Pro x64 |
Pricing (As configured) | $180 (eMMC / No OS) $220 (eMMC / Win 10 Home x64) $250 (with M.2 SATA SSD, as configured / No OS) |
Full Specifications | ECS LIVA Z Specifications |
The ECS LIVA-ZN33 package comes with a 65W (19V @ 3.42A) AC adapter and a VESA mount / screws in addition to the main unit. Windows 10 is the only officially supported OS, and all the drivers were downloaded on our review unit from the ECS website.
One of the interesting hardware components in the LIVA-ZN33 is the integrated digital microphone. This allows the end user to configure it as an always-listening machine (if needed), without the need to connect an external microphone. The other selling point is the availability of two GbE RJ-45 ports. Intel Ethernet controllers would have sweetened the deal, but, ECS has taken the cost factor into consideration and opted for Realtek controllers.
The industrial design is attractive and the unit feels solid. The chassis is made of plastic (not uncommon at the targeted price point). The internal heat sink which faces the bottom lid, however, is solid metal, and possesses good characteristics for effective heat transfer from the SoC. It would have been nice if the design allowed for the generated heat to dissipate via convection. In any case, we will take a look at the effectiveness of the thermal design in a later section.
Platform Analysis
The specifications of the Intel Celeron N3350 indicate that the SoC can support up to 6 PCIe 2.0 lanes, 2 SATA ports, and 8 USB ports. The break-up of the high-speed I/O lanes is interesting in the context of the availability of four USB 3.0 ports (3x Type-A + 1x Type-C) as well as two GbE LAN ports in the LIVA Z
Intel Celeron N3350 HSIO Block Diagram [ Courtesy : Intel Pentium and Celeron Processor N- and J- Series Datasheet - Volume 1 of 3 (PDF) ]
The distribution of PCIe lanes in the LIVA Z is as below.
- PCI-E 2.0 x1 port #3 In Use @ x1 (Realtek RTL8168/8111 PCI-E Gigabit Ethernet Adapter)
- PCI-E 2.0 x1 port #4 In Use @ x1 (Intel Dual Band Wireless-AC 3165 AC HMC WiFi Adapter)
- PCI-E 2.0 x1 port #5 In Use @ x1 (Realtek RTL8168/8111 PCI-E Gigabit Ethernet Adapter)
Since the used PCIe lanes are muxed with the USB 3.0 lanes, and, there are four USB 3.0 ports in the system, it stands to reason that at least one USB 3.0 port has its bandwidth shared with the PCIe lanes / networking interfaces. At the price point targeted by the LIVA Z, this is hardly an issue.
In the table below, we have an overview of the various systems that we are comparing the ECS LIVA-ZN33 against. Note that they may not belong to the same market segment. The relevant configuration details of the machines are provided so that readers have an understanding of why some benchmark numbers are skewed for or against the ECS LIVA-ZN33 when we come to those sections.
Comparative PC Configurations | ||
Aspect | ECS LIVA-ZN33 | |
CPU | Intel Celeron N3350 | Intel Celeron N3350 |
GPU | Intel HD Graphics 500 | Intel HD Graphics 500 |
RAM | Kingston CBD16D3LS1KBG/4G DDR3 4GB 11-11-11-28 @ 1600 MHz |
Kingston CBD16D3LS1KBG/4G DDR3 4GB 11-11-11-28 @ 1600 MHz |
Storage | SanDisk eMMC DF4032 (32 GB; eMMC v5.0-compatible) ADATA Premier SP600 SP600NS34 (128 GB; M.2 Type 2242 SATA III; MLC) |
SanDisk eMMC DF4032 (32 GB; eMMC v5.0-compatible) ADATA Premier SP600 SP600NS34 (128 GB; M.2 Type 2242 SATA III; MLC) |
Wi-Fi | Intel Dual Band Wireless-AC 3165 (1x1 802.11ac - 433 Mbps) |
Intel Dual Band Wireless-AC 3165 (1x1 802.11ac - 433 Mbps) |
Price (in USD, when built) | $180 (eMMC / No OS) $220 (eMMC / Win 10 Home x64) $250 (with M.2 SATA SSD, as configured / No OS) |
$180 (eMMC / No OS) $220 (eMMC / Win 10 Home x64) $250 (with M.2 SATA SSD, as configured / No OS) |
Performance Metrics - I
The ECS LIVA-ZN33 was evaluated using our standard test suite for low power desktops / industrial PCs. Not all benchmarks were processed on all the machines due to updates in our testing procedures. Therefore, the list of PCs in each graph might not be the same. In the first section, we will be looking at SYSmark 2014 SE, as well as some of the Futuremark benchmarks.
BAPCo SYSmark 2014 SE
BAPCo's SYSmark 2014 SE is an application-based benchmark that uses real-world applications to replay usage patterns of business users in the areas of office productivity, media creation and data/financial analysis. In addition, it also addresses the responsiveness aspect which deals with user experience as related to application and file launches, multi-tasking etc. Scores are meant to be compared against a reference desktop (the SYSmark 2014 SE calibration system in the graphs below). While the SYSmark 2014 benchmark used a Haswell-based desktop configuration, the SYSmark 2014 SE makes the move to a Lenovo ThinkCenter M800 (Intel Core i3-6100, 4GB RAM and a 256GB SATA SSD). The calibration system scores 1000 in each of the scenarios. A score of, say, 2000, would imply that the system under test is twice as fast as the reference system.
SYSmark 2014 SE also adds energy measurement to the mix. Unfortunately, we were unable to get the power logging working with either the ECS LIVA Z or the ASRock Beebox N3000-NUC.
The scores are what one would expect from an Atom-class CPU. On the performance per dollar front, systems like the ECS LIVA Z may turn out to be the right choice (particulary if the workloads are not time-critical). In any case, we do see a definite improvement in the Apollo Lake Celeron N3350-based ECS LIVA Z over the Braswell Celeron N3000-based ASRock Beebox N3000-NUC. Note, however, that the Intel Celeron N3000 has a 4W TDP, while the N3350 has a 6W TDP.
Futuremark PCMark 8
PCMark 8 provides various usage scenarios (home, creative and work) and offers ways to benchmark both baseline (CPU-only) as well as OpenCL accelerated (CPU + GPU) performance. We benchmarked select PCs for the OpenCL accelerated performance in all three usage scenarios. These scores are heavily influenced by the CPU in the system. Amongst the considered fanless PCs, the ECS LIVA Z is bettered only by the CompuLab fitlet-XA10-LAN which is configured with a much higher TDP (thanks to its efficient thermal design).
Miscellaneous Futuremark Benchmarks
3D Rendering - CINEBENCH R15
We have moved on from R11.5 to R15 for 3D rendering evaluation. CINEBENCH R15 provides three benchmark modes - OpenGL, single threaded and multi-threaded. Evaluation of select PCs in all three modes provided us the following results.
The Celeron N3350 in the ECS LIVA Z does win the single-threaded case. However, its dual-core configuration loses out to the quad-core Celeron N2930-based Zotac CI320 nano in the multi-threaded workload. In the OpenGL case, the higher base frequency (320 MHz vs. 200 MHz) helps the Celeron N3050 in the LIVA x2 get a slight edge over the Celeron N3350 in the ECS LIVA Z.
Performance Metrics - II
In this section, we mainly look at benchmark modes in programs used on a day-to-day basis, i.e, application performance and not synthetic workloads.
x264 Benchmark
First off, we have some video encoding benchmarks courtesy of x264 HD Benchmark v5.0. This is simply a test of CPU performance. As expected, the Intel Celeron N3050 turns out to be one of the leaders, with the second pass result going in favor of the AMD A10 Micro-6700T with a much higher TDP.
7-Zip
7-Zip is a very effective and efficient compression program, often beating out OpenCL accelerated commercial programs in benchmarks even while using just the CPU power. 7-Zip has a benchmarking program that provides tons of details regarding the underlying CPU's efficiency. In this subsection, we are interested in the compression and decompression MIPS ratings when utilizing all the available threads. The quad-core CPU-based CompuLab fitlet-XA10-LAN and the Zotac ZBOX CI320 nano score much better than the ECS LIVA Z. The decompression numbers for the dual-core CPU-based systems are all around the same ballpark.
TrueCrypt
As businesses (and even home consumers) become more security conscious, the importance of encryption can't be overstated. CPUs supporting the AES-NI instruction can accelerate the encryption and decryption processes. The Apollo Lake Intel Celeron N3350 does support AES-NI. TrueCrypt, a popular open-source disk encryption program can take advantage of the AES-NI capabilities. The TrueCrypt internal benchmark provides some interesting cryptography-related numbers to ponder. In the graph below, we can get an idea of how fast a TrueCrypt volume would behave in the ECS LIVA-ZN33 and how it would compare with other select PCs. This is a purely CPU feature / clock speed based test.
Agisoft Photoscan
Agisoft PhotoScan is a commercial program that converts 2D images into 3D point maps, meshes and textures. The program designers sent us a command line version in order to evaluate the efficiency of various systems that go under our review scanner. The command line version has two benchmark modes, one using the CPU and the other using both the CPU and GPU (via OpenCL). The benchmark takes around 50 photographs and does four stages of computation:
- Stage 1: Align Photographs
- Stage 2: Build Point Cloud (capable of OpenCL acceleration)
- Stage 3: Build Mesh
- Stage 4: Build Textures
We record the time taken for each stage. Since various elements of the software are single threaded, others multithreaded, and some use GPUs, it is interesting to record the effects of CPU generations, speeds, number of cores, DRAM parameters and the GPU using this software.
Dolphin Emulator
Wrapping up our application benchmark numbers is the Dolphin Emulator benchmark mode results. This is again a test of the CPU capabilities - in particular, the single-threaded performance plays a major role, and we see the ECS LIVA Z come out on top.
Networking 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 ECS LIVA-ZN33. 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. Note that these benchmarks were run with the SSD as the primary drive andthe eMMC mounted as a secondary drive. PCMark 8 allows specification of the logical drive for benchmarking. The storage workload was run twice (once for the SSD and once for the eMMC partitions). Both results are presented in the graphs below.
As expected, eMMC performance can't obviously match up to the SSD. However, as the storage subsystem score shows, the difference is not acute for light home / office use.
We also ran CrystalDiskMark 5.1.2 x64 in the eMMC-only configuration to get some best-case numbers for different artificial workloads.
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.
All the considered PCs use a 1x1 WLAN configuration. The Intel AC3165 performs up to its potential and the antenna placement / plastic chassis ensures that the bandwidth numbers for the ECS LIVA Z are quite good compared to the other systems in the comparison list.
HTPC Credentials
Given the fanless nature of the ECS LIVA Z and the presence of Intel HD Graphics, we expect many purchasers to use it as a media playback machine / HTPC. Given the specifications, it is quite clear that we are not looking at a madVR capable machine, but one targeted at the entry-level / average HTPC user or someone looking for a HTPC to put in a second or third room (non-primary HTPC). There are two HTPC aspects that we will explore in this section, one related to network streaming (OTT services), and the other related to local file playback. Prior to that, we have a small sub-section dealing with refresh rate accuracy.
Refresh Rate Accuracy
Starting with Haswell, Intel, AMD and NVIDIA have been on par with respect to display refresh rate accuracy. The most important refresh rate for videophiles is obviously 23.976 Hz (the 23 Hz setting). As expected, the ECS LIVA-ZN33 has no trouble with refreshing the display appropriately in this setting.
The gallery below presents some of the other refresh rates that we tested out. The first statistic in madVR's OSD indicates the display refresh rate.
Network Streaming Efficiency
Evaluation of OTT playback efficiency was done by playing back our standard YouTube test stream and five minutes from our standard Netflix test title. Using HTML5, the YouTube stream plays back a 1080p encoding. Since YouTube now defaults to HTML5 for video playback, we have stopped evaluating Adobe Flash acceleration. Note that only NVIDIA exposes GPU and VPU loads separately. Both Intel and AMD bundle the decoder load along with the GPU load. The following two graphs show the power consumption at the wall for playback of the HTML5 stream in Mozilla Firefox (v 52.0.1).
GPU load was around 49.81% for the YouTube HTML5 stream and 0.0294% for the steady state 6 Mbps Netflix streaming case.
Netflix streaming evaluation was done using the Windows 10 Netflix app. Manual stream selection is available (Ctrl-Alt-Shift-S) and debug information / statistics can also be viewed (Ctrl-Alt-Shift-D). Statistics collected for the YouTube streaming experiment were also collected here.
Note that these power consumption numbers were obtained for the eMMC-only configuration. As such, the OTT streaming workloads operate in a more power-efficient manner compared to PCs with SSDs such as the ASRock Beebox N3000-NUC, but, it is not as power-efficient as the previous-generation LIVA units based on the Atom-class SoCs.
Decoding and Rendering Benchmarks
Prior to testing out the decoding and rendering performance, we take a look at the DXVA decoding support available via Intel's Apollo Lake GPU drivers. Note that we are using driver version v21.20.16.4551 for the GPU.
DXVAChecker confirms that the Apollo Lake GPU is able to decode all commonly-used codecs (including VP9 and HEVC Main10) up to 4K resolution.
In order to evaluate local file playback, we concentrate on EVR-CP and Kodi. We already know that EVR works quite well even with the Intel IGP for our test streams. In our earlier reviews, we focused on presenting the GPU loading and power consumption at the wall in a table (with problematic streams in bold). Starting with the Broadwell NUC review, we decided to represent the GPU load and power consumption in a graph with dual Y-axes. Elevent different test streams of 90 seconds each were played back with a gap of 30 seconds between each of them. The characteristics of each stream are annotated at the bottom of the graph. Note that the GPU usage is graphed in red and needs to be considered against the left axis, while the at-wall power consumption is graphed in blue and needs to be considered against the right axis.
Frame drops are evident whenever the GPU load consistently stays above the 85 - 90% mark. In our test cases, this happened for 60fps interlaced content as well as 60 fps 4K content. Given that EVR-CP involves a bit mor shader work compared to lean renderers like EVR, and the fact that the Intel HD Graphics in the Intel Celeron N3350 is not very powerful, the results are entirely plausible.
Kodi, on the other hand, has absolutely no trouble in playing back all our test streams (including the HEVC Main10 videos).
Power Consumption and Thermal Performance
The power consumption at the wall was measured with a 1080p display being driven through the HDMI port. In the graphs below, we compare the idle and load power of the ECS LIVA-ZN33 with other low power PCs evaluated before. For load power consumption, we ran the AIDA64 System Stability Test with various stress components, as well as our custom stress test,and noted the maximum sustained power consumption at the wall.
The numbers were recorded for the eMMC-only configuration. Since the load power consumption peak happened in our custom stress test (which doesn't stress the SSD / storage), we didn't feel the need to repeat the custom stress test in the SSD configuration. We, however, noticed that the baseline number (eMMC configuration) shifted up by around 800mW to 1W in the scenarios where the SSD remained relatively idle.
Our thermal stress routine starts with the system at idle, followed by four stages of different system loading profiles using the AIDA64 System Stability Test (each of 30 minutes duration). In the first stage, we stress the CPU, caches and RAM. In the second stage, we add the GPU to the above list. In the third stage, we stress the GPU standalone. In the final stage, we stress all the system components (including the disks). Beyond this, we leave the unit idle in order to determine how quickly the various temperatures in the system can come back to normal idling range. The various clocks, temperatures and power consumption numbers for the system during the above routine are presented in the graphs below. The first set of three graphs correspond to the eMMC-only configuration, while the second set correspond to the SSD + eMMC configuration.
We repeated the same observations with our legacy stress test using the latest versions of Prime95 and Furmark - Prime95 v28.10 for 30 minutes (after launching with the max. stress option), followed by Furmark v1.18.20 for 30 minutes. The Prime95 load is then removed, allowing just the GPU alone to be stressed for 30 minutes. The system is then left idle.
According to the official specifications, the junction temperature of the Intel Celeron N3350 is 90C. The thermal design does manage to keep the core temperatures well below that. We see that the system is essentially limited by the 6W package TDP. Our testing shows that the thermal design can easily handle a sustained 6W power consumption by the SoC for at least 2 hours. The idling temperatures are around 40 - 45C.
Another important aspect to keep note of while evaluating fanless PCs is the chassis temperature. Using the Android version of the FLIR One thermal imager, we observed the chassis temperature after the CPU package temperature reached the maximum steady state value in the above graph.
We have additional thermal images in the gallery below. The FLIR One thermal imager app reported a maximum chassis temperature of around 53C in our measurements.
Given that the heat sink faces downwards. the positioning of the chassis and its all-plastic non-metallic nature ensure that the externally accessible parts remain at a relatively low temperature.
Final Words
The ECS LIVA Z provided us with the opportunity to take a first look at the performance of a passively-cooled
mini-PC based on the Intel Apollo Lake platform. Compared to the previous LIVA units, the Z family (they also have a Kaby Lake-based LIVA Z Plus in the pipeline) has moved on from providing systems with extremely small footprints. The boards are slightly larger than the traditional NUCs (115 x 111mm, as compared to 101.6 x 101.6mm). Fortunately, the larger footprint has enabled ECS to put in dual LAN ports in all the models. Dual LAN ports usually lend themselves to network appliances (x86 routers, firewalls, and the like). However, both ports in the LIVA Z are backed by Realtek controllers. Intel controllers deliver better performance and are more widely supported in the x86-based networking appliance OS market. Therefore, we wouldn't advise the usage of the LIVA Z solely as a network appliance (where performance might be important). That said, the availability of three high-performance network interfaces (2x GbE LAN and 1x 1x1 802.11ac WLAN) can lend itself to some specific use-cases.
The fanless nature makes the system suitable for certain HTPC workloads. The LIVA Z fulfills all the basic necessities - 4K output (with 4Kp60 supported on the mini-DP port) and full hardware decode for 4Kp60 content in a variety of codecs. That said, it is not a 4K OTT box because it lacks the necessary DRM capabilities as well as a HDCP 2.2-capable HDMI 2.0 port necessary for Netflix 4K (currently available only on select Kaby Lake systems).
On the price front, the ECS LIVA Z is a winner. $180 delivers a system ready for OS installation, while $220 fetches a ready-to-go system with Windows 10 Home x64 pre-installed. The overall performance benefits are a step up from similarly priced Bay Trail and Braswell systems.
The I/O and industrial design are pleasing to the eye. On the motherboard front, we would like to have support for the installation of a M.2 2280 SSD (currently, the choices for M.2 2242 SSDs supported by the LIVA Z are limited). Based on our market research, the ADATA SP600 presents the best balance of cost (128GB at $70) and flash quality (2D MLC). The JMicron controller is not going to create benchmark records, but, the performance is a good match for the typical LIVA Z use-cases.
A 6W TDP SoC doesn't stress the thermal design aspect too much. Therefore, the usage of an all-plastic chassis is excusable in this fanless PC. However, we believe the chassis design could be altered for better airflow and convective cooling. In terms of storage, I have always been critical of 32GB drives as primary OS drives. The unit also ships with only a single DDR3L SODIMM slot occupied. It would be an interesting exercise to determine if operation in dual-channel mode delivers any tangible performance benefits. Thankfully, ECS has provided the option to use a SSD for the boot drive. Users can also install a second SODIMM - however, it is recommended that sticks from a dual SODIMM kit be used for this purpose.
One of the issues I encountered while performing the OS installation was that the UEFI BIOS would not allow for booting from a USB installation drive with a MBR partition. Setting the BIOS to legacy mode allowed MBR keys to boot, but, wouldn't allow installation to the eMMC partition. We resolved the issue by using a Windows installation USB key formatted with the GPT option.
Coming to the business end of the review, we can say that, taken standalone, the ECS LIVA ZN33 more than delivers $180 worth of computing performance. It also manages to keep thermals under limit for consumer workloads. Features such as the Type-C USB 3.0 port and 802.11ac Wi-Fi are welcome changes compared to the first-generation Bay Trail-based LIVA units. The idle and load power consumption profiles of the LIVA Z are excellent and continue the tradition of the LIVA PCs being one of the most power-efficient lineups in the market. The LIVA Z has no show-stopping drawbacks. Based on the workloads, it is definitely an option to consider.