The market demand for small form-factor (SFF) PCs was kickstarted by the Intel NUC in the early 2010s. Since then, many vendors have come out with their own take on the Intel NUC using both Intel and AMD processors. In recent years, we have also seen various Asian companies such as Beelink, Chuwi, GEEKOM, GMKtec, MinisForum, etc. emerging with a focus solely on these types of computing systems. We had looked at GEEKOM's Jasper Lake offering - the MiniAir 11 - last year, and came away satisfied with the build quality of the device.

Earlier this year, GEEKOM announced a tie-up with ASUS to market specific configurations of the ASUS ExpertCenter PN53 under their own brand as the GEEKOM AS 6. Based on AMD's Rembrandt line of notebook processors, the GEEKOM AS 6 comes with a choice of the Ryzen 9 6900HX, Ryzen 7 6800H, or the Ryzen 7 7735H.

The CPU choices all belong to the Rembrandt family of SoCs utilizing CPU cores with the Zen 3+ microarchitecture. While the Ryzen 9 6800HX and the Ryzen 7 6800H were released in early 2022, the Ryzen 7 7735H was launched as a China-only SKU earlier this year along with a few other 'Rembrandt Refresh' SoCs. We have already taken a detailed look at the ASRock Industrial 4X4 BOX-7735U based on one such refresh SKU - the Ryzen 7 7735U.

The initial wave of systems based on Rembrandt SoCs impressed with their performance, but the connectivity aspects did not measure up to the possibilities enabled by Thunderbolt in the Intel ecosystem. Over the last few quarters, that has perceptibly changed with the USB4 capabilities of the SoCs (including PCIe tunneling support) getting reflected in the newer set of Rembrandt systems. Most mid-range and high-end systems based on Rembrandt now come with full-featured USB4 ports, and the GEEKOM AS 6 is no exception. The company sent across their flagship configuration - the AS 6 equipped with a Ryzen 9 6900HX, 32GB of DDR5 RAM, and a 1TB Gen4 x4 NVMe SSD - to put through our evaluation routine for small form-factor computing systems. This review explores the performance profile and value proposition of the GEEKOM AS 6.

Introduction and Product Impressions

Small form-factor (SFF) systems have been replacing bulky desktops for many use-cases in recent years. The rapid growth in this segment has created a number of companies (mainly based out of Asia) focusing on these systems. GEEKOM is a private label brand of Shenzhen Jiteng Network Technology Co., Ltd. - an OEM / ODM for small form-factor computing systems. The company manufactures both Intel and AMD-based systems. For the latter, the company has a tie-up with ASUS. The GEEKOM AS 5 (using AMD Cezanne SoCs) introduced last year was a rebranded version of the ASUS ExpertCenter PN52. The GEEKOM AS 6 we are looking at in this review is a rebrand of the ASUS ExpertCenter PN53, and uses high-end Rembrandt SoCs. The Ryzen 9 6900HX processor in the AS 6 review system has a 8C / 16T (Zen 3+) configuration fabricated in TSMC's 6nm process. In addition to the minor improvements in CPU performance over the Zen 3 / Cezanne family, the integrated GPU has also migrated from the Vega-based ones in previous generations to the RDNA2-based Radeon 680M.

GEEKOM's package components for the AS 6 differs from the ASUS PN53's. We have seen in previous ASUS mini-PC reviews that a wired keyboard and mouse are always bundled in the retail kit. GEEKOM opts to keep things simple and compact by shipping the main system, a VESA mount, a few screws for the SATA and M.2 drives, and a 150W power adapter (20V @ 7.5A) along with a quick-start guide.

The company also keeps the ASUS branding in the front panel, and the system supports all of the software value additions available for the ASUS PN53. On the hardware front, the system clocks in at 120mm x 130mm x 58mm (compared to a full-height mainstream NUC at 117mm x 112mm x 54mm). The larger dimensions allow for a wider variety of ports in the system. Unlike the mainstream Intel NUCs that adopt a single PCB layout, the GEEKOM AS 6 adopts a motherboard / daughterboard strategy with a flexible printed circuit (FPC) connector, as shown below.

The daughterboard includes a Gen4 x4 M.2 SSD slot as well as the SATA connector and the non Type-C display outputs. The DDR5 SODIMM slots and one of the Gen4 x4 M.2 SSD slots on the main board are pre-populated in the GEEKOM configurations. The WLAN / BT module is under the primary NVMe SSD slot (similar to the mainstream Intel NUCs). The metal frame, coupled with the thermal pads for both Gen4 M.2 slots point to some thought being put in to keep the SSD temperatures under check. There is also a 2.5" SATA drive slot in the system. However, with no direct active airflow on that side of the board, it may not be ideal to install one in addition to the M.2 SSDs. The external I/O includes five display outputs (including the two USB4 Type-C ports), but only four of them can be active at a time.

The GEEKOM AS 6 review configuration was equipped with 2x Crucial CT16G48C40S5.M8A1 DDR5-4800 SODIMMs and a Kingston NV2 PCIe Gen4 x4 M.2 2280 NVMe SSD. Windows 11 Professional came pre-installed.

The full specifications of the review sample (as tested) are summarized in the table below. The Ryzen 9 6900HX has a configurable TDP (cTDP) ranging from 35W to 54W. As we shall see in our detailed investigation into the thermal characteristics in a later section, the form-factor of the system and the notebook-style thermal solution restricts the operation of the SoC in a 35W cTDP mode.

GEEKOM AS 6 (ASUS PN53) Specifications (as tested)
Processor	AMD Ryzen 9 6900HX Zen 3+ (Rembrandt) 8C/16T, 3.3 - 4.9 GHz TSMC 6nm, 16MB L3, 45W Max / Normal / Target TDP : 54W / 45W / 35W
Memory	Crucial CT16G48C40S5.M8A1 DDR5-4800 SODIMM 40-39-39-77 @ 4800 MHz 2x16 GB
Graphics	AMD Radeon 680M (Rembrandt) - Integrated (12 CUs @ 2.4 GHz)
Disk Drive(s)	Kingston NV2 SNV2S1000G (1 TB; M.2 2280 PCIe 4.0 x4 NVMe;) (Kioxia BiCS5 112L 3D TLC; Silicon Motion SM2267XT Controller)
Networking	1x 2.5 GbE RJ-45 (Realtek RTL8125) Mediatek MT7922 (RZ616) Wi-Fi 6E (2x2 802.11ax - 1.9 Gbps)
Audio	Realtek ALC256 (3.5mm Audio Jack in Front) Digital Audio with bitstreaming support over HDMI, Display Port, and Type-C
Video	2x HDMI 2.0 (4Kp60) 1x Display Port 1.4a (5Kp60) 2x Display Port 1.4 over USB4 Type-C (5Kp30)
Miscellaneous I/O Ports	2x USB 3.2 Gen 1 (Front) 1x USB4 Type-C (Front) 1x USB4 Type-C (Rear) 3x USB 3.2 Gen 1 Type-A (Rear)
Operating System	Windows 11 Enterprise (22000.2176)
Pricing	(Street Pricing on July 20th, 2023) US $709 (as configured, with OS, using coupon code as640a)
Full Specifications	GEEKOM AS 6 (ASUS ExpertCenter PN53) Specifications

The next section details the the various BIOS options and follows it up with a detailed platform analysis.

Setup Notes and Platform Analysis

Our review sample of the GEEKOM AS 6 came with all necessary components pre-installed - including the OS. Prior to setting up the OS on first boot, we took some time to look into the BIOS interface. ASUS primarily targets business installations with their ExpertCenter PN series mini-PCs. Unlike their enthusiast-focused motherboards, the BIOS is not a fancy one navigable using both mouse and keyboard. Despite the keyboard-only navigation, there are plenty of configurable options. The video below presents an overview of the BIOS interface.

The system is equipped with both firmware and discrete TPMs, with the firmware TPM enabled by default. Various TPM aspects can be configured in the 'Trusted Computing' sub-section. The VRAM allocation can also be specified - by default, it is set to 'Auto', but available options range from 128 MB to 8 GB. It is also possible to disable specific USB ports in the BIOS - extremely useful for business installations where it may be necessary for IT administrators to lock down systems and prevent users from connecting potentially harmful USB drives to it. Enabling the UEFI network stack (off by default) allows for network boot. The LAN, WLAN, and Bluetooth can also be disabled via the BIOS. HDMI CEC control is available, but only for the HDMI port on the left.

The BIOS can also be used to configure the two M.2 SSDs installed in the two Gen4 x4 NVMe slots in RAID mode. The 'APM Settings' allows configuration of power management-related aspects such as CEC / ErP / system behavior upon coming out of a power failure / Wake-on-LAN etc. Though the system has only one SATA port for a 2.5" drive, the BIOS also includes options related to M.2 SATA ports. The M.2 ports in the AS 6 support only NVMe SSDs. Virtualization support (enabled by default) can also be controlled via the 'Platform Configuration' sub-section.

The 'Monitor' section presents a real-time view of the temperatures of the processor package and the motherboard, along with the core voltage and fan speed. It also allows users to finely control the fan speed setting using the ASUS QFan feature. The 'Fan Control' option is set to 'Performance Mode' by default, with a 'Quiet Mode' also available. The 'Boot' section includes various boot configuration options such as boot logo display control, fast boot (enabled by default), etc. Boot priorities can be set, and a one-time override can also be chosen. Secure boot is enabled by default, and can be changed if needed. Options such as keys can also be reset / reprogrammed. The BIOS also includes an option to upgrade the BIOS from within its own interface. The EzFlash option triggers a reboot and presents the ASUS EzFlash Utility screen. The BIOS CAP file can be chosen from any file system present in the drives attached to the system and an update can be processed. We used this feature to update the BIOS version from 2.03.00 to 2.06.00 (downloaded from ASUS's support site for the PN53), and the process completed without any issues. After the update, the GEEKOM boot logo was replaced by ASUS's 'In Search of Incredible' graphics, but there was no change in the functioning of the system otherwise.

The AMD Rembrandt SoC Platform

AMD's Rembrandt line of low-power processors targets mobile systems. Similar to previous generation products in that line (such as Renoir and Cezanne), Rembrandt also opts for a tightly integrated single die SoC package. This is unlike Intel's mobile offerings where the PCH is in a separate die, and the link between it and the CPU / iGPU processor die can act as a bandwidth bottleneck. AMD's Rembrandt block diagram gives a quick overview of the features and connectivity options.

The topology indicates the presence of a Zen 3+ CCD along with an RDNA2 iGPU with up to 12 CUs. The 35W+ Rembrandt SoCs contain either 6 or 8 CPU cores, while the iGPU is configured with either 6 or 12 CUs. The SKUs otherwise differ in the base and boost clocks for the CPU cores and the clock rate of the iGPU. The Ryzen 9 6900HX in the GEEKOM AS 6 is a top-end config (8C/16T with a 12CU RDNA2 iGPU), with the CPU cores operating at 3.3 GHz. The cores can boos up to 4.9 GHz. The iGPU runs at 2.4 GHz.

The overall high-speed I/O distribution in the GEEKOM AS 6 board design is brought out in the bus layout diagram below.

Despite the built-in connectivity options, the board design includes a couple of Genesys Logic USB hub chips to enable some of the front and rear USB ports. The bottom USB 3.2 Gen 1 ports in the rear are enabled by an ASMedia ASM3042 xHCI controller, and the SATA port by an ASMedia ASM1061 bridge chip, each of which connect to the SoC via PCIe x1 links. Other than those, the WLAN and LAN components also utilize one PCIe lane each. There are two USB4 controllers built into the SoC. When a Thunderbolt (read, PCIe tunneling-capable) device is connected, it enumerates as a PCIe bus under the USB4 bridge entry. On the other hand, when a legacy USB device is connected, a bunch of USB controllers are enumerated under an internal PCIe GPP bridge to bus.

In today's review, we compare the GEEKOM AS 6 and a host of other systems based on processors with TDPs ranging from 15W to 55W. The systems do not target the same market segments, but a few key aspects lie in common, making the comparisons relevant.

Comparative PC Configurations
Aspect	GEEKOM AS 6 (ASUS PN53)	GEEKOM AS 6 (ASUS PN53)ASRock NUC BOX-1360P-D5 (Performance)Intel NUC13ANKi7 (Arena Canyon)ASRock 4X4 BOX-5800U (Performance)Intel NUC12WSKi7 (Wall Street Canyon)ASRock 4X4 BOX-7735U (Normal)ASRock NUC BOX-1260PASUS PN50ASRock 4X4 BOX-7735U (Performance)
CPU	AMD Ryzen 9 6900HX Zen 3+ (Rembrandt) 8C/16T, 3.3 - 4.9 GHz TSMC 6nm, 16MB L3, 45W Max / Normal / Target TDP : 54W / 45W / 35W	AMD Ryzen 9 6900HX Zen 3+ (Rembrandt) 8C/16T, 3.3 - 4.9 GHz TSMC 6nm, 16MB L3, 45W Max / Normal / Target TDP : 54W / 45W / 35W
GPU	AMD Radeon 680M (Rembrandt) - Integrated (12 CUs @ 2.4 GHz)	AMD Radeon 680M (Rembrandt) - Integrated (12 CUs @ 2.4 GHz)
RAM	Crucial CT16G48C40S5.M8A1 DDR5-4800 SODIMM 40-39-39-77 @ 4800 MHz 2x16 GB	Crucial CT16G48C40S5.M8A1 DDR5-4800 SODIMM 40-39-39-77 @ 4800 MHz 2x16 GB
Storage	Kingston NV2 SNV2S1000G (1 TB; M.2 2280 PCIe 4.0 x4 NVMe;) (Kioxia BiCS5 112L 3D TLC; Silicon Motion SM2267XT Controller)	Kingston NV2 SNV2S1000G (1 TB; M.2 2280 PCIe 4.0 x4 NVMe;) (Kioxia BiCS5 112L 3D TLC; Silicon Motion SM2267XT Controller)
Wi-Fi	1x 2.5 GbE RJ-45 (Realtek RTL8125) Mediatek MT7922 (RZ616) Wi-Fi 6E (2x2 802.11ax - 1.9 Gbps)	1x 2.5 GbE RJ-45 (Realtek RTL8125) Mediatek MT7922 (RZ616) Wi-Fi 6E (2x2 802.11ax - 1.9 Gbps)
Price (in USD, when built)	(Street Pricing on July 20th, 2023) US $709 (as configured, with OS)	(Street Pricing on July 20th, 2023) US $709 (as configured, with OS)

Benchmarks were processed afresh on all of the above systems with the latest BIOS for each. The next few sections will deal with comparative benchmarks for the above systems.

Our 2022 Q4 update to the test suite for Windows 11-based systems carries over some of the standard benchmarks we have been using over the last several years. While UL's PCMark makes the list, we have opted to temporarily suspend reporting of BAPCo's SYSmark scores (pending fixture of the energy consumption aspect). Instead, BAPCO's CrossMark multi-platform benchmarking tool has been added to the set along with UL's Procyon suite. While CrossMark employs idle time compression and processes all workloads in an opaque manner, UL's Procyon processes real-world workloads with user interactions (like BAPCo's SYSmark). We have augmented the UL Procyon suite benchmark with our own custom energy measurement setup.

UL PCMark 10

UL's PCMark 10 evaluates computing systems for various usage scenarios (generic / essential tasks such as web browsing and starting up applications, productivity tasks such as editing spreadsheets and documents, gaming, and digital content creation). We benchmarked select PCs with the PCMark 10 Extended profile and recorded the scores for various scenarios. These scores are heavily influenced by the CPU and GPU in the system, though the RAM and storage device also play a part. The power plan was set to Balanced for all the PCs while processing the PCMark 10 benchmark. The scores for each contributing component / use-case environment are also graphed below.

UL PCMark 10 - Performance Scores

The systems being compared against each other are all configured in a slightly different manner. The Rembrandt SKUs are all 8C/16T configurations with the Ryzen 9 6900HX enjoying a slight clock speed advantage over the Ryzen 7 7735U in the ASRock 4X4 BOX-7735U. The 'Performance' setting for the latter sets the target TDP around 42W, while we will see in a later section that the GEEKOM AS 6 is set for a 35W target TDP. The Arena Canyon NUC and the ASRock NUC BOX-1360P/D5, on the other hand, employ a hybrid processor with a 40W target TDP.

The Raptor Lake-based systems with the 40W TDP configuration enjoy a significant edge in the 'Essentials' segment. However, across all of Productivity, Gaming, and Digital Content Creation, the Rembrandt processors outclass the Raptor Lake ones. Eight high-performance cores seem to be better suited for these workloads compared to a 4P / 8E configuration, and the RDNA2 iGPU with 12 CUs also enjoys a bit of an edge over the Intel Iris Xe Graphics with 96EUs.

UL Procyon v2.1.544

PCMark 10 utilizes open-source software such as Libre Office and GIMP to evaluate system performance. However, many of their professional benchmark customers have been requesting evaluation with commonly-used commercial software such as Microsoft Office and Adobe applications. In order to serve their needs, UL introduced the Procyon benchmark in late 2020. There are five benchmark categories currently - Office Productivity, AI Inference, Battery Life, Photo Editing, and Video Editing. AI Inference benchmarks are available only for Android devices, while the battery life benchmark is applicable to Windows devices such as notebooks and tablets. We presents results from our processing of the other three benchmarks.

UL Procyon - Office Productivity Scores

The results from the UL Procyon Office Productivity Suite are similar to what was seen for the Essentials suite in PCMark. The Raptor Lake-P systems with a 40W TDP outperform the Rembrandt systems across the board.

From an energy consumption viewpoint, the Intel NUC (Arena Canyon) seems to be the most optimized of the lot, getting the job done fast enough to be second in the performance charts without consuming too much power in the process. The GEEKOM AS 6 doesn't seem to be energy efficient for this workload.

Moving on to the evaluation of Adobe Photoshop and Adobe Lightroom, we find the Raptor Lake-P systems having the edge over the Rembrandt ones.

While the ASRock Industrial 4X4 BOX-7735U in a 28W configuration gets the job done with minimum energy consumption, the GEEKOM AS 6 manages to make it to the middle of the pack in this aspect.

UL Procyon evaluates performance for video editing using Adobe Premier Pro. While rendering and playback can make use of the GPU, the main encoding task is still left to the CPU. Here, more high-performance cores with a bigger power budget can potentially help. The ASRock Industrial NUC BOX-1360P/D5 completes the job in the least time, but is beat on energy efficiency only by the 28W Rembrandt configuration (which is in the middle of the pack with respect to time taken).

The GEEKOM AS 6 gets the job done faster than the Arena Canyon NUC and all other contenders (except the NUC BOX-1360P/D5), but energy efficiency is not its forte. Only the Wall Street Canyon NUC fares worse in that aspect.

BAPCo CrossMark 1.0.1.86

BAPCo's CrossMark aims to simplify benchmark processing while still delivering scores that roughly tally with SYSmark. The main advantage is the cross-platform nature of the tool - allowing it to be run on smartphones and tablets as well.

BAPCo CrossMark 1.0.1.86 - Sub-Category Scores

The relative performance seen in the UL Procyon office workloads translate to BAPCo's CrossMark also, but the gulf is a bit wider here. Idle time compression seems to favor the Raptor Lake-P systems more, and the GEEKOM AS 6 is placed in the middle of the pack overall. Since CrossMark attempts to consolidate different workloads together without idle time intervals and play it back in a non-real-time environment, it is not entirely representative of real-world performance like PCMark 10 and UL Procyon. Therefore, tasks requiring frequent user interaction are better represented by those other benchmarks.

System Performance: Application-Specific Workloads

Standardized benchmarks such as UL's PCMark 10 and BAPCo's SYSmark take a holistic view of the system and process a wide range of workloads to arrive at a single score. Some systems are required to excel at specific tasks - so it is often helpful to see how a computer performs in specific scenarios such as rendering, transcoding, JavaScript execution (web browsing), etc. This section presents focused benchmark numbers for specific application scenarios.

3D Rendering - CINEBENCH R23

We use CINEBENCH R23 for 3D rendering evaluation. R23 provides two benchmark modes - single threaded and multi-threaded. Evaluation of different PC configurations in both supported modes provided us the following results.

The Raptor Lake-P systems have a slightly higher processor power budget, but the single-threaded performance (even on a per-clock basis) is better in them compared to the Rembrandt systems. The GEEKOM AS 6 is in the middle of the pack, but comes out with a better performance in the multi-threaded cases. Only the Arena Canyon NUC outperforms the GEEKOM AS 6 in that workload.

Transcoding: Handbrake 1.5.1

Handbrake is one of the most user-friendly open source transcoding front-ends in the market. It allows users to opt for either software-based higher quality processing or hardware-based fast processing in their transcoding jobs. Our new test suite uses the 'Tears of Steel' 4K AVC video as input and transcodes it with a quality setting of 19 to create a 720p AVC stream and a 1080p HEVC stream.

The presence of eight high-performance cores in the Rembrandt systems gives them a very good advantage for this multi-threaded benchmark. Higher power budget translates to higher performance given that they all have the same core counts. The Arena Canyon NUC enjoys a slight advantage over the GEEKOM AS 6, as the eight efficiency cores with a higher power budget seem to give it the edge over four high-performance cores in the latter.

Hardware transcoding is available in the form of the VCE feature in the GPU. Intel systems have QuickSync, but a direct comparison is not advisable given that the encoding quality may differ. So, the comparison in the above graphs is restricted to AMD systems. The FPS is a function of the GPU clock rate and is directly tied to the available power budget. So, it is no surprise that the 4X4 BOX-7735U in Performance mode with a 42W TDP has a slight edge over the GEEKOM AS 6.

Archiving: 7-Zip 21.7

The 7-Zip benchmark is carried over from our previous test suite with an update to the latest version of the open source compression / decompression software.

Raptor Lake-P and Rembrandt are neck-to-neck in terms of compression rates, with the GEEKOM AS 6 at the top (and the Arena Canyon NUC / ASRock Industrial NUC BOX-1360P/D5 following closely behind). However, decompression is a completely different story, with the Rembrandt systems enjoying as much as a 25% advantage in processing rates.

Web Browsing: JetStream, Speedometer, and Principled Technologies WebXPRT4

Web browser-based workloads have emerged as a major component of the typical home and business PC usage scenarios. For headless systems, many applications based on JavaScript are becoming relevant too. In order to evaluate systems for their JavaScript execution efficiency, we are carrying over the browser-focused benchmarks from the WebKit developers used in our notebook reviews. Hosted at BrowserBench, JetStream 2.0 benchmarks JavaScript and WebAssembly performance, while Speedometer measures web application responsiveness.

Raptor Lake-P's single-threaded advantage gives it the edge across all considered browsers in the JetStream 2.0 workload. The GEEKOM AS 6 slots in the middle of the pack. The same behavior is seen in the Speedometer 2.0 benchmark also.

From a real-life workload perspective, we also process WebXPRT4 from Principled Technologies. WebXPRT4 benchmarks the performance of some popular JavaScript libraries that are widely used in websites.

A mixed real-world browser workload in the form of WebXPRT4 also lines up the systems in an order similar to what was seen in Speedometer and JetStream. JavaScript performance seems to be primarily impacted by single-threaded performance.

Application Startup: GIMP 2.10.30

A new addition to our systems test suite is AppTimer - a benchmark that loads up a program and determines how long it takes for it to accept user inputs. We use GIMP 2.10.30 with a 50MB multi-layered xcf file as input. What we test here is the first run as well as the cached run - normally on the first time a user loads the GIMP package from a fresh install, the system has to configure a few dozen files that remain optimized on subsequent opening. For our test we delete those configured optimized files in order to force a fresh load every second time the software is run.

As it turns out, GIMP does optimizations for every CPU thread in the system, which requires that higher thread-count processors take a lot longer to run. So the test runs quick on systems with fewer threads, however fast cores are also needed. Raptor Lake-P edges out Rembrandt in this case across all considered systems, with the GEEKOM AS 6 not faring particularly well in the bottom half of the pack.

Cryptography Benchmarks

Cryptography has become an indispensable part of our interaction with computing systems. Almost all modern systems have some sort of hardware-acceleration for making cryptographic operations faster and more power efficient. In the case of IoT servers, many applications - including web server functionality and VPN - need cryptography acceleration.

BitLocker is a Windows features that encrypts entire disk volumes. While drives that offer encryption capabilities are dealt with using that feature, most legacy systems and external drives have to use the host system implementation. Windows has no direct benchmark for BitLocker. However, we cooked up a BitLocker operation sequence to determine the adeptness of the system at handling BitLocker operations. We start off with a 4.5GB RAM drive in which a 4GB VHD (virtual hard disk) is created. This VHD is then mounted, and BitLocker is enabled on the volume. Once the BitLocker encryption process gets done, BitLocker is disabled. This triggers a decryption process. The times taken to complete the encryption and decryption are recorded. This process is repeated 25 times, and the average of the last 20 iterations is graphed below.

Hardware acceleration is available for the operations in all of the systems. The time taken for processing is directly dependent on the available power budget. The AMD systems have a significant advantage over the Intel ones in both encryption and decryption, possibly due to faster clocks across more cores.

GPU Performance: Synthetic Benchmarks

AMD's Rembrandt SoCs includes an integrated GPU update to enable it to compete against Intel's Xe iGPUs. The new RDNA2 microarchitecture is present in the Ryzen 9 6900HX in the form of the Radeon 680M. With 12 CUs and 768 shader units clocked at 2.4 GHz, AMD claims that the GPU should be capable of playing virtually any modern game at Full HD resolutions. For full-blown desktop systems or mini-PCs targeting the gaming market, we look at gaming workloads. However, for general-purpose mini-PC models like the GEEKOM AS 6, we restrict ourselves to a series of canned benchmarks from Kishonti and Futuremark / UL. Prior to that, a look at the capabilities of the GPU via GPU-Z is warranted.

GPU-Z indicates the presence of hardware ray-tracing - a first for an integrated GPU.

AMD Radeon 680M Features

In the ray tracing department, Rembrandt already scores over Intel's latest iGPU. The remaining subsections below look into the performance aspects.

GFXBench

The DirectX 12-based GFXBench tests from Kishonti are cross-platform, and available all the way down to smartphones. As such, they are not very taxing for discrete GPUs and modern integrated GPUs. We processed the offscreen versions of the 'Aztec Ruins' benchmark.

At 1080p, the souped-up Raptor Lake-P in the ASRock Industrial NUC BOX-1360P/D5 has a healthy boost over the Arena Canyon NUC. The Rembrandt-based systems slot between the two, with the GEEKOM AS 6 at the top of that pack. However, at 1440p, the Rembrandt systems move to the top of the pack.

UL 3DMark

Four different workload sets were processed in 3DMark - Fire Strike, Time Spy, Night Raid, and Wild Life.

3DMark Fire Strike

The Fire Strike benchmark has three workloads. The base version is meant for high-performance gaming PCs. It uses DirectX 11 (feature level 11) to render frames at 1920 x 1080. The Extreme version targets 1440p gaming requirements, while the Ultra version targets 4K gaming system, and renders at 3840 x 2160. The graph below presents the overall score for the Fire Strike Extreme and Fire Strike Ultra benchmark across all the systems that are being compared.

UL 3DMark - Fire Strike Workloads

The GFXBench scores showed Rembrandt pulling away from Raptor Lake-P at 1440p, and that is evident in the Extreme and Ultra scores above also. The GEEKOM AS 6 has the highest iGPU clocks among all the Rembrandt systems, and that is reflected in its appearance at the top of the pack in both settings.

3DMark Time Spy

The Time Spy workload has two levels with different complexities. Both use DirectX 12 (feature level 11). However, the plain version targets high-performance gaming PCs with a 2560 x 1440 render resolution, while the Extreme version renders at 3840 x 2160 resolution. The graphs below present both numbers for all the systems that are being compared in this review.

UL 3DMark - Time Spy Workloads

The 42W Ryzen 7 7735U in the ASRock Industrial 4X4 BOX-7735U manages to perform as well as the GEEKOM AS 6's 35W Ryzen 9 6900HX in both settings, but they are both at the top of the pack.

3DMark Wild Life

The Wild Life workload was initially introduced as a cross-platform GPU benchmark in 2020. It renders at a 2560 x 1440 resolution using Vulkan 1.1 APIs on Windows. It is a relatively short-running test, reflective of mobile GPU usage. In mid-2021, UL released the Wild Life Extreme workload that was a more demanding version that renders at 3840 x 2160 and runs for a much longer duration reflective of typical desktop gaming usage.

UL 3DMark - Wild Life Workloads

This workload sees the Raptor Lake-P's iGPU hold the edge when configured with higher clocks / higher sustained power budget (as in the NUC BOX-1360P/D5). The GEEKOM AS 6 follows closely behind.

3DMark Night Raid

The Night Raid workload is a DirectX 12 benchmark test. It is less demanding than Time Spy, and is optimized for integrated graphics. The graph below presents the overall score in this workload for different system configurations.

The GEEKOM AS 6 and the 4X4 BOX-7735U are comfortably on top, with almost a 20% advantage over the best that Raptor Lake-P has to offer.

3DMark Port Royal

UL introduced the Port Royal ray-tracing benchmark as a DLC for 3DMark in early 2019. The scores serve as an indicator of how the system handles ray-tracing effects in real-time.

The GPU in the Ryzen 9 6900HX is clocked higher than the one in the Ryzen 7 7735U, and that is reflected in the Port Royal performance scores graphed above.

Workstation Performance - SPECworkstation 3.1

SFF PCs traditionally do not lend themselves to workstation duties. However, a recent trend towards miniaturized workstations has been observed. While systems in the GEEKOM AS 6's form-factor are still not capable enough to become workstations, the rapid performance improvements over the years has encouraged us to benchmark some of these UCFF / SFF systems for content creation workloads and professional applications. Towards this, we processed the SPECworkstation 3.1 benchmark from SPEC.

The SPECworkstation 3.1 benchmark measures workstation performance based on a number of professional applications. It includes more than 140 tests based on 30 different workloads that exercise the CPU, graphics, I/O and memory hierarchy. These workloads fall into different categories.

• Media and Entertainment (3D animation, rendering)
• Product Development (CAD/CAM/CAE)
• Life Sciences (medical, molecular)
• Financial Services
• Energy (oil and gas)
• General Operations
• GPU Compute

Individual scores are generated for each test and a composite score for each category is calculated based on a reference machine (HP Z240 tower workstation using an Intel E3-1240 v5 CPU, an AMD Radeon Pro WX3100 GPU, 16GB of DDR4-2133, and a SanDisk 512GB SSD). Official benchmark results generated automatically by the benchmark itself are linked in the table below for the systems being compared.

SPECworkstation 3.1 Official Results (2K)
GEEKOM AS 6 (ASUS PN53)	Run Summary
ASRock NUC BOX-1360P-D5 (Performance)	Run Summary
Intel NUC13ANKi7 (Arena Canyon)	Run Summary
ASRock 4X4 BOX-5800U (Performance)	Run Summary
Intel NUC12WSKi7 (Wall Street Canyon)	Run Summary
ASRock 4X4 BOX-7735U (Normal)	Run Summary
ASRock NUC BOX-1260P	Run Summary
ASUS PN50	Run Summary
ASRock 4X4 BOX-7735U (Performance)	Run Summary

Details of the tests in each category, as well as an overall comparison of the systems on a per-category basis are presented below.

Media and Entertainment

The Media and Entertainment category comprises of workloads from five distinct applications:

• The Blender workload measures system performance for content creation using the open-source Blender application. Tests include rendering of scenes of varying complexity using the OpenGL and ray-tracing renderers.
• The Handbrake workload uses the open-source Handbrake application to transcode a 4K H.264 file into a H.265 file at 4K and 2K resolutions using the CPU capabilities alone.
• The LuxRender workload benchmarks the LuxCore physically based renderer using LuxMark.
• The Maya workload uses the SPECviewperf 13 maya-05 viewset to replay traces generated using the Autodesk Maya 2017 application for 3D animation.
• The 3ds Max workload uses the SPECviewperf 13 3dsmax-06 viewset to replay traces generated by Autodesk's 3ds Max 2016 using the default Nitrous DX11 driver. The workload represents system usage for 3D modeling tasks.

This category sees the Raptor Lake-P and Rembrandt systems getting ordered on the basis of the configured sustained TDP. The 4X4 BOX-7735U enjoys a lead with its 42W TDP, and the 40W RPL-P systems (including the Arena Canyon NUC) follow behind. The GEEKOM AS 6 with its 35W setting is in the middle of the pack, ahead of the previous generation systems.

Product Development

The Product Development category comprises of eight distinct workloads:

• The Rodinia (CFD) workload benchmarks a computational fluid dynamics (CFD) algorithm.
• The WPCcfd workload benchmarks another CFD algorithm involving combustion and turbulence modeling.
• The CalculiX workload uses the Calculix finite-element analysis program to model a jet engine turbine's internal temperature.
• The Catia workload uses the catia-05 viewset from SPECviewperf 13 to replay traces generated by Dassault Systemes' CATIA V6 R2012 3D CAD application.
• The Creo workload uses the creo-02 viewset from SPECviewperf 13 to replay traces generated by PTC's Creo, a 3D CAD application.
• The NX workload uses the snx-03 viewset from SPECviewperf 13 to replay traces generated by the Siemens PLM NX 8.0 CAD/CAM/CAE application.
• The Solidworks workload uses the sw-04 viewset from SPECviewperf 13 to replay traces generated by Dassault Systemes' SolidWorks 2013 SP1 CAD/CAE application.
• The Showcase workload uses the showcase-02 viewset from SPECviewperf 13 to replay traces from Autodesk???s Showcase 2013 3D visualization and presentation application

The trend observed in the Media and Entertainment category repeats here, with the 4X4 BOX-7735U faring better than the GEEKOM AS 6.

Life Sciences

The Life Sciences category comprises of four distinct test sets:

• The LAMMPS set comprises of five tests simulating different molecular properties using the LAMMPS molecular dynamics simulator.
• The NAMD set comprises of three tests simulating different molecular interactions.
• The Rodinia (Life Sciences) set comprises of four tests - the Heartwall medical imaging algorithm, the Lavamd algorithm for calculation of particle potential and relocation in a 3D space due to mutual forces, the Hotspot algorithm to estimate processor temperature with thermal simulations, and the SRAD anisotropic diffusion algorithm for denoising.
• The Medical workload uses the medical-02 viewset from SPECviewperf 13 to determine system performance for the Tuvok rendering core in the ImageVis3D volume visualization program.

The addition of a GPU-centric component in the workload allows the GEEKOM AS 6 to leapfrog the RPL-P-based systems, but the 42W configuration of the Ryzen 7 7735U still remains on top.

Financial Services

The Financial Services workload set benchmarks the system for three popular algorithms used in the financial services industry - the Monte Carlo probability simulation for risk assessment and forecast modeling, the Black-Scholes pricing model, and the Binomial Options pricing model.

While the 42W configuration of the 4X4 BOX-7735U continues to remain on top, some of the components in this workload are also better served with high single-threaded performance. We see the Arena Canyon NUC slotting into the second place, with the GEEKOM AS 6 slipping to third, and the TDP ordering being maintained.

Energy

The Energy category comprises of workloads simulating various algorithms used in the oil and gas industry:

• The FFTW workload computes discrete Fourier transforms of large matrices.
• The Convolution workload computes the convolution of a random 100x100 filter on a 400 megapixel image.
• The SRMP workload processes the Surface-Related Multiples Prediction algorithm used in seismic data processing.
• The Kirchhoff Migration workload processes an algorithm to calculate the back propogation of a seismic wavefield.
• The Poisson workload takes advantage of the OpenMP multi-processing framework to solve the Poisson's equation.
• The Energy workload uses the energy-02 viewset from SPECviewperf 13 to determine system performance for the open-source OPendTec seismic visualization application.

This workload also includes a GPU-centric component, and that helps the GEEKOM AS 6 and the other high TDP Rembrandt system come out in the top two (with the difference being in the realm of run-to-run variations). The high-performance RPL-P system (NUC BOX-1360P/D5) also delivers very similar performance.

General Operations

In the General Options category, the focus is on workloads from widely used applications in the workstation market:

• The 7zip workload represents compression and decompression operations using the open-source 7zip file archiver program.
• The Python workload benchmarks math operations using the numpy and scipy libraries along with other Python features.
• The Octave workload performs math operations using the Octave programming language used in scientific computing.
• The Storage workload evaluates the performance of the underlying storage device using transaction traces from multiple workstation applications.

The components in this workload benefit from single-threaded performance as well as high-performance storage subsystems. The SSD used in the Arena Canyon NUC, NUC BOX-1360P/D5, and 4X4 BOX-7735U use SSDs with DRAM for the FTL, while the GEEKOM AS 6 came pre-configured with a DRAM-less SSD for cost optimization. As a result, there is a significant gulf in the scores of the other systems and the GEEKOM AS 6, with the latter making an entry in the bottom half of the pack.

GPU Compute

In the GPU Compute category, the focus is on workloads taking advantage of the GPU compute capabilities using either OpenCL or CUDA, as applicable:

• The LuxRender benchmark is the same as the one seen in the media and entertainment category.
• The Caffe benchmark measures the performance of the Caffe deep-learning framework.
• The Folding@Home benchmark measures the performance of the system for distributed computing workloads focused on tasks such as protein folding and drug design.

We only process the OpenCL variants of the benchmark, but the Intel iGPU drivers have an issue with the Caffe benchmark. All the Intel-based systems have a default 0.01 scoring for that component, and we see that reflected in their appearance in the bottom half of the graph below.

Within the Rembrandt-based systems, the 42W configuration of the Ryzen 7 7735U has the edge over the 35W Ryzen 9 6900HX in the GEEKOM AS 6, relegating it to the second place.

System Performance: Multi-Tasking

One of the key drivers of advancements in computing systems is multi-tasking. On mobile devices, this is quite lightweight - cases such as background email checks while the user is playing a mobile game are quite common. Towards optimizing user experience in those types of scenarios, mobile SoC manufacturers started integrating heterogenous CPU cores - some with high performance for demanding workloads, while others were frugal in terms of both power consumption / die area and performance. This trend is now slowly making its way into the desktop PC space.

Multi-tasking in typical PC usage is much more demanding compared to phones and tablets. Desktop OSes allow users to launch and utilize a large number of demanding programs simultaneously. Responsiveness is dictated largely by the OS scheduler allowing different tasks to move to the background. The processor is required to work closely with the OS thread scheduler to optimize performance in these cases. Keeping these aspects in mind, the evaluation of multi-tasking performance is an interesting subject to tackle.

We have augmented our systems benchmarking suite to quantitatively analyze the multi-tasking performance of various platforms. The evaluation involves triggering a ffmpeg transcoding task to transform 1716 3840x1714 frames encoded as a 24fps AVC video (Blender Project's 'Tears of Steel' 4K version) into a 1080p HEVC version in a loop. The transcoding rate is monitored continuously. One complete transcoding pass is allowed to complete before starting the first multi-tasking workload - the PCMark 10 Extended bench suite. A comparative view of the PCMark 10 scores for various scenarios is presented in the graphs below. Also available for concurrent viewing are scores in the normal case where the benchmark was processed without any concurrent load, and a graph presenting the loss in performance.

UL PCMark 10 Load Testing - Digital Content Creation Scores

UL PCMark 10 Load Testing - Productivity Scores

UL PCMark 10 Load Testing - Essentials Scores

UL PCMark 10 Load Testing - Gaming Scores

UL PCMark 10 Load Testing - Overall Scores

The gaming section comes to the rescue of the Rembrandt systems, allowing them to retain the top two spots in the overall scores. However, we see some benefits to the hybrid architecture in RPL-P as the introduction of concurrent loading tends to bog down the Rembrandt systems much more in the sub-components. The end result is that the RPL-P systems move to the top and push the Rembrandt systems down in the presence of intensive background tasks.

Following the completion of the PCMark 10 benchmark, a short delay is introduced prior to the processing of Principled Technologies WebXPRT4 on MS Edge. Similar to the PCMark 10 results presentation, the graph below show the scores recorded with the transcoding load active. Available for comparison are the dedicated CPU power scores and a measure of the performance loss.

Principled Technologies WebXPRT4 Load Testing Scores (MS Edge)

The RPL-P systems had come out on top in the browser benchmarks, and the addition of background loading doesn't change the relative ordering of the systems. The GEEKOM AS 6 continues to remain in the middle of the pack.

The final workload tested as part of the multitasking evaluation routine is CINEBENCH R23.

3D Rendering - CINEBENCH R23 Load Testing - Single Thread Score

3D Rendering - CINEBENCH R23 Load Testing - Multiple Thread Score

Intensive background tasks keep the efficiency cores busy, and take up some of the power budget. As a result, while the RPL-P systems came out on top in the absence of the background loads in both ST and MT modes, the Rembrandt systems manages to outwit them in raw scores after the introduction of the load.

After the completion of all the workloads, we let the transcoding routine run to completion. The monitored transcoding rate throughout the above evaluation routine (in terms of frames per second) is graphed below.

The transcoding rate during different segments is also recorded below.

GEEKOM AS 6 (Ryzen 9 6900HX) ffmpeg Transcoding Rate (Multi-Tasking Test)
Task Segment	Transcoding Rate (FPS)
	Minimum	Average	Maximum
Transcode Start Pass	3	13.8	44
PCMark 10	0	12.02	42
WebXPRT 4	3	12.27	22.5
Cinebench R23	2	12.73	39
Transcode End Pass	4.5	13.59	40.5

The comparison is against RPL-P systems such as the NUC BOX-1360P/D5. The GEEKOM AS 6 has a much lower delta in the transcoding rate between different task segments, but that translates to a lower primary workload score as seen in the graphs above. Overall, the RPL-P systems seem to prioritize foreground task better compared to Rembrandt-based systems such as the GEEKOM AS 6.

HTPC Credentials

The 2022 Q4 update to our system reviews brings an updated HTPC evaluation suite for systems. After doing away with the evaluation of display refresh rate stability and Netflix streaming evaluation, the local media playback configurations have also seen a revamp. This section details each of the workloads processed on the GEEKOM AS 6 (ASUS PN53) as part of the HTPC suite.

YouTube Streaming Efficiency

YouTube continues to remain one of the top OTT platforms, primarily due to its free ad-supported tier. Our HTPC test suite update retains YouTube streaming efficiency evaluation as a metric of OTT support in different systems. Mystery Box's Peru 8K HDR 60FPS video is the chosen test sample. On PCs running Windows, it is recommended that HDR streaming videos be viewed using the Microsoft Edge browser after putting the desktop in HDR mode.

YouTube Streaming Statistics

The GPU in GEEKOM AS 6 (ASUS PN53) supports hardware decoding of VP9 Profile 2, and we see the stream encoded with that codec being played back. The streaming is perfect, thanks to the powerful GPU and hardware decoding support - the few dropped frames observed in the statistics below are due to mouse clicks involved in bringing up the overlay and the transition to different encode resolutions in the initial stages.

The streaming efficiency-related aspects such as GPU usage and at-wall power consumption are also graphed below.

 

The NUC BOX-1360P/D5 RPL-P system is the most energy efficient of the tested lot by a huge margin, but the Rembrandt systems (including the GEEKOM AS 6) slot in right behind.

Hardware-Accelerated Encoding and Decoding

The transcoding benchmarks in the systems performance section presented results from evaluating the VCE encoder within Handbrake's framework. The capabilities of the decoder engine are brought out by DXVAChecker.

Video Decoding Hardware Acceleration in GEEKOM AS 6 (ASUS PN53)

On paper, this codec list is quite comprehensive and should cover most home consumer and digital signage requirements

Local Media Playback

Evaluation of local media playback and video processing is done by playing back files encompassing a range of relevant codecs, containers, resolutions, and frame rates. A note of the efficiency is also made by tracking GPU usage and power consumption of the system at the wall. Users have their own preference for the playback software / decoder / renderer, and our aim is to have numbers representative of commonly encountered scenarios. Our Q4 2022 test suite update replaces MPC-HC (in LAV filters / madVR modes) with mpv. In addition to being cross-platform and open-source, the player allows easy control via the command-line to enable different shader-based post-processing algorithms. From a benchmarking perspective, the more attractive aspect is the real-time reporting of dropped frames in an easily parseable manner. The players / configurations considered in this subsection include:

• VLC 3.0.18
• Kodi 20.2
• mpv 0.35.1 (hwdec auto, vo=gpu-next)
• mpv 0.35.1 (hwdec auto, vo=gpu-next, profile=gpu-hq)

Fourteen test streams (each of 90s duration) were played back from the local disk with an interval of 30 seconds in-between. Various metrics including GPU usage, at-wall power consumption, and total energy consumption were recorded during the course of this playback.

All our playback tests were done with the desktop HDR setting turned on. It is possible for certain system configurations to automatically turn on/off the HDR capabilities prior to the playback of a HDR video, but, we didn't take advantage of that in our testing.

 

Lower TDP Rembrandt systems fare better in the energy consumption metric. All codecs play back well, except for the AV1 clip. VLC doesn't take advantage of the GPU hardware decoding capabilities for that codec. As a result, the playback for that clip, as well as the energy numbers are not ideal in the GEEKOM AS 6.

 

The user experience with Kodi is not much different, as AV1 hardware acceleration is again unused. Since the Kodi GUI is kept active at the main screen between the playback of different streams, the GPU remains active and continues to render the UI. This results in the energy numbers creeping up when the full playback period is considered.

 

AV1 decode acceleration is utilized by mpv. Unfortunately, despite the decoder engine running at full tilt, there is some problem with the video output path resulting in approximately half the frames getting dropped. In terms of energy efficiency, the GEEKOM AS 6 is slotted in the middle of the pack.

 

The case of the AV1 playback being spotty despite the hardware acceleration is seen here again. The activation of the GPU shaders for rendering in the GPU-HQ case results in relatively higher energy numbers for all the systems.

Power Consumption and Thermal Characteristics

The power consumption at the wall was measured with a 4K display being driven through the HDMI port of the system. In the graph below, we compare the idle and load power of the GEEKOM AS 6 (ASUS PN53) with other systems evaluated before. For load power consumption, we ran the AIDA64 System Stability Test with various stress components, as well as our custom stress test with Prime95 / Furmark, and noted the peak as well as idling power consumption at the wall.

The peak numbers are consistent with the TDP and suggested PL1 / PL2 values for the processors in the systems, and do not come as any surprise. The idle power consumption numbers are more interesting, as there is more scope for BIOS optimizations as well as board design itself to affect that. In general, Intel's own NUC systems are well-optimized - the RPL-P Arena Canyon NUC idles at 4.84W. The GEEKOM AS 6 is similar to other AMD systems at around 9W. The presence of a number of bridge chips in the main and daughterboards may also be a contributor to this relatively high number.

Stress Testing

Our thermal stress routine is a combination of Prime95, Furmark, and Finalwire's AIDA64 System Stability Test. The following 9-step sequence is followed, starting with the system at idle:

• Start with the Prime95 stress test configured for maximum power consumption
• After 30 minutes, add Furmark GPU stress workload
• After 30 minutes, terminate the Prime95 workload
• After 30 minutes, terminate the Furmark workload and let the system idle
• After 30 minutes of idling, start the AIDA64 System Stress Test (SST) with CPU, caches, and RAM activated
• After 30 minutes, terminate the previous AIDA64 SST and start a new one with the GPU, CPU, caches, and RAM activated
• After 30 minutes, terminate the previous AIDA64 SST and start a new one with only the GPU activated
• After 30 minutes, terminate the previous AIDA64 SST and start a new one with the CPU, GPU, caches, RAM, and SSD activated
• After 30 minutes, terminate the AIDA64 SST and let the system idle for 30 minutes

Traditionally, this test used to record the clock frequencies - however, with the increasing number of cores in modern processors and fine-grained clock control, frequency information makes the graphs cluttered and doesn't contribute much to understanding the thermal performance of the system. The focus is now on the power consumption and temperature profiles to determine if throttling is in play.

The cTDP of the processor is evident in the reported STAPM power numbers. This 'skin-temperature aware power management' number has a slow rise towards the 35W number even as the package power peaks at 65W for around 8s, with another level at 54W for around 4 minutes, before reaching a stable plateau at 35W.

The core temperature is kept south of 90C during the 54W package power duration. Once the STAPM reaches 35W, throttling takes effect and the package power goes down. The core temperature stabilizes around 75C, pointing to a very effective thermal solution. The SSD thermal solution is also very good, with the thermal pad and the metal frame combining to keep it under 50C even under disk stress. The only hot spot under stress seems to be the DIMMs which stabilize around 81C and 78C. Form-factor limitations do not allow a heat sink for the SODIMMs, but this is something that may need addressing with active airflow over the memory modules in the future, or even vertical placement of the SODIMMs with additional air gap between the modules.

Miscellaneous Aspects and Concluding Remarks

Networking and storage are aspects that may be of vital importance in specific PC use-cases. The GEEKOM AS 6 (ASUS PN53) comes with a single 2.5 Gbps RJ-45 port and a 2x2 Wi-Fi 6E WLAN subsystem. This is high-end, considering that many mini-PC manufacturers are still advertising 1 Gbps and Wi-Fi 5 / Wi-Fi 6 in their units. A dual LAN solution would have been a nice-to-have feature, but the system does come with more USB ports than is usual. A USB to LAN adapter, or even better - a USB4 dock with different port varieties (even 10 Gbps LAN) can be connected to expand networking functionality, if needed.

On the storage side, GEEKOM advertises a Gen4 NVMe SSD, but it is Gen4 in name only. The pre-configured system comes with a Kingston NV2 PCIe 4.0 x4 NVMe SSD. It is a DRAM-less offering from Kingston aimed primarily at OEMs. There are many variants of this SSD in the market - some with the Silicon Motion SM2267XT controller, and others with the Phison E21T. Some use TLC, while others use QLC. It is a complete hit or miss (our SSD was a SM2267XT variant with 112L BiCS5 3D TLC), and when a vendor has such multiple variants under one model name, our suggestion is to avoid it as much as possible. If I were to recommend a SSD for a mini-PC, the Kingston NV2 would probably come somewhere in the bottom of the list. From a benchmarking perspective, we provide results from the WPCstorage test of SPECworkstation 3.1. This benchmark replays access traces from various programs used in different verticals and compares the score against the one obtained with a 2017 SanDisk 512GB SATA SSD in the SPECworkstation 3.1 reference system.

SPECworkstation 3.1.0 - WPCstorage SPEC Ratio Scores

The graphs above present results for different verticals, as grouped by SPECworkstation 3.1. The storage workload consists of 60 subtests. Access traces from CFD solvers and programs such as Catia, Creo, and Soidworks come under 'Product Development'. Storage access traces from the NAMD and LAMMPS molecular dynamics simulator are under the 'Life Sciences' category. 'General Operations' includes access traces from 7-Zip and Mozilla programs. The 'Energy' category replays traces from the energy-02 SPECviewperf workload. The 'Media and Entertainment' vertical includes Handbrake, Maya, and 3dsmax.

Given the DRAM-less nature and choice of controller in the Kingston NV2, its Gen4 designation does nothing to prevent it from faring worse than even the Kingston KC2500 (a Gen3 SSD with SM2262EN and 96L 3D TLC) in the Wall Street Canyon NUC. These characteristics also explain the relative numbers for the different workloads in the WPCstorage suite.

Closing Thoughts

The GEEKOM AS 6 (ASUS PN53) provided us with the opportunity to evaluate a high-end Rembrandt-based AMD mini-PC with premium features. We had already reviewed the ASRock Industrial 4X4 BOX-7735U with full USB4 40 Gbps support, and the GEEKOM AS 6 retains the same functionality. An added advantage is that the rear USB4 port can also be used as a DC-In with a 100W+ USB-PD power source. On the other hand, the ASRock Industrial system includes Realtek DASH functionality for remote management. Such a feature would help expand the target market for the ASUS PN53 / GEEKOM AS 6. Changing a couple of the USB 3.2 Gen 1 Type-A ports to USB 3.2 Gen 2 (10 Gbps) would have been welcome, but the presence of two USB4 ports - one in the front, and another in the rear, more than makes up for that. To keep up with the changing times, enabling USB 3.2 Gen 2x2 (20 Gbps) support on the USB4 ports would have been nice, but that is available only on specific Raptor Lake-P USB4 / Thunderbolt 4 ports for now. Hopefully, that is something that can be addressed in AMD's Phoenix-based systems or in later generations.

Thanks to the tie-up with ASUS, GEEKOM is able to deliver a full-featured BIOS. Aspects such as the ability to upgrade the BIOS from within itself, and the ability to override boot devices on a one-time basis are user-friendly features that are usually not present in the vanilla BIOS used by most Asian vendors.

The thermal design of the GEEKOM AS 6 (ASUS PN53) is top-notch given the form-factor constraints, and we believe it can accommodate a cTDP slightly higher than the configured 35W. Unfortunately, there is no direct way in the BIOS to tweak this number. The system can sustain 54W package power for more than four minutes, which should serve it well for consumer workloads that are bursty in nature. In this review, we compared it against a bunch of systems that could sustain between 28W and 42W for long durations, and the GEEKOM AS 6 came out to be a promising performer. The integrated GPU's performance is excellent, and energy efficiency is passable. It emerges as pole performer in many multi-threaded workloads, and goes neck-to-neck against RPL-P systems configured for a slightly higher TDP in others.

The GEEKOM AS 6 (ASUS PN53) and a Mainstream Half-Height Intel NUC for Size Comparison

GEEKOM's tie-up with ASUS is puzzling at first glance. While ASUS is a well-established vendor in the personal computing space, GEEKOM is essentially an upstart in the field. However, one must note that GEEKOM is a private label brand of Shenzhen Jiteng Network Technology Co.. This company also provides design and manufacturing services. It is our guess that ASUS manufactures the PN52 (AS 5) and PN53 (AS 6) mini-PCs at Jiteng under an ODM deal (where Jiteng just does the realization of the design provided by ASUS). ASUS is keeping the mid-range high-volume configuration with the Ryzen 5 6600H ($429 barebones) to be sold under their own branding, while the high-end lower-volume products using the Ryzen 9 6900HX ($709 with 1TB SSD / 32GB RAM, coupon code - as640a) and Ryzen 7 7735H ($689 with 1TB SSD / 32GB RAM) are left for GEEKOM to sell under their own label. The Ryzen 7 6800H is available in both the PN53 ($669 barebones) and GEEKOM flavors ($659 with 1TB SSD / 32GB RAM), but only the ASUS version is available on Amazon. The GEEKOM AS 6 is being sold exclusively via their own storefront, and will not be available on Amazon.

The value proposition of the GEEKOM AS 6 needs to be discussed after taking pricing into account. The GEEKOM AS 6 with the Ryzen 7 6800H, SSD, RAM, and Windows 11 Pro OS is $10 cheaper than the corresponding barebones version of the ASUS ExpertCenter PN53. The Ryzen 9 6900HX configuration will probably deliver similar value for mainstream users. Power users may want to replace the SSD with a more performant one. Traditionally, buying from no-name brands or upstarts like GEEKOM - particularly from their own storefronts - at 'too good to be true' prices is fraught with risk. However, with ASUS backing up on the BIOS and drivers maintenance front, the risk is alleviated to some extent. Phoenix-based mini-PCs have also started to appear in the market from vendors like Beelink and MinisForum. This may put some price pressure on Rembrandt-based systems, but that is good news for consumers. In terms of connectivity and peripherals (USB4 support, in particular), the Rembrandt and Phoenix-based systems are equivalent. While Zen 4 and RDNA3 can deliver better performance numbers, AMD's drivers are still in the nascent stage. Rembrandt systems have been around for a few quarters now, and will probably be a more stable investment for the immediate future.