Using a Raspberry Pi for digital signage is nothing new. The low cost and HD video output mean they are right at home powering the sort of digital signage that’s becoming more common around airports and even roadside advertising hoardings. The MPi4 NEC Media from Sharp NEC Display Solutions, spotted by CNX Software, puts the Compute Module 4 right in the digital signage driving seat as part of Sharp/NEC large format 4K displays.
(Image credit: Sharp/NEC)
While Sharp/NEC have been using Raspberry Pi Compute Modules in their commercial products for five years, the new Compute Module 4 board takes its position as the entry-level option in the company’s range of Smart Display Module cards (the upper tiers are the exclusive territory of Intel processors).
The MPi4 NEC MediaPlayer Kit offers all the usual abilities of the Raspberry Pi 4 SOC – which means up to 4K/60fps HEVC decoding – with the addition of 32GB of eMMC flash storage for holding all your picture and video content and a large heatsink. There’s 4GB of LPDDR4 RAM, Gigabit Ethernet, and a pair of USB 2.0 ports, with a micro USB port which we assume is used to flash the eMMC storage of the Compute Module 4.
The whole thing is run by NEC’s MediaPlayer software, and can be slotted straight into a compatible display, making use of internal connections for power, data, and control interfaces. The software can connect multiple boards into a network, and can handle streaming content via UDP, HTTP or RTSP.
Marketed as an affordable solution, the official site only has the rather foreboding ‘price on application’. If you feel like applying, make that application via Sharp/NEC’s website.
No matter how much web browsers improve, it feels like they can’t keep pace with everything we want to do. Open one too many tabs on a few-year-old laptop, and your fan starts spinning, your battery life dips, your system starts to slow. A faster or cleaner PC might fix it, but a startup called Mighty has a different idea: a $30-a-month web browser that lives in the cloud.
Instead of your own physical computer interacting with each website, you stream a remote web browser instead, one that lives on a powerful computer many miles away with its own 1,000Mbps connection to the internet.
Suddenly, your decent internet connection would feel like one of the fastest internet connections in the world, with websites loading nigh-instantly and intensive web apps running smoothly without monopolizing your RAM, CPU, GPU and battery, no matter how many tabs you’ve got open — because the only thing your computer is doing is effectively streaming a video of that remote computer (much like Netflix, YouTube, Google Stadia, etc.) while sending your keyboard and mouse commands to the cloud.
Skeptical? I definitely am, but perhaps not for the reason you’d think — because I tried this exact idea nearly a decade ago, and it absolutely works in practice. In 2012, cloud gaming pioneer OnLive introduced a virtual desktop web browser that would let you load full websites on an iPad in the blink of an eye and stream 4K video from YouTube. (Quite the feat in 2012!) I called it the fastest web browser you’ve ever used, and OnLive’s asking price was just $5 a month.
Cloud desktop providers like Shadow have also offered similar capabilities; when you rent their gaming-PCs-in-the-cloud ($12-15 a month), you can use those virtual PCs’ built-in web browsers to get similar speeds, thanks to the fact they typically live in data centers with very few hops to (and possibly direct peering arrangements with) major content delivery networks.
Mighty argues that by focusing on the browser (rather than recreating a whole Windows PC), it can give more people what they actually want. “Most people want an experience where the underlying OS and the application (the browser) interoperate seamlessly versus having to tame two desktop experiences,” founder Suhail Doshi commented at Hacker News. Mighty claims it’ll eliminate distracting cookies and ads, automatically notify you about Zoom meetings, quick search Google Docs and presumably other integrations to come. Mighty also says it encrypts your data and keystrokes, among other security promises.
But it’s not entirely clear why it costs so much more, or who would be willing to pay $30 a month for such a subscription — you’d think the kinds of people who can afford a monthly browser bill on top of their monthly internet bill would be the same kinds of people who can afford a faster PC and faster internet to begin with. Gigabit fiber is already a reality for some homes, and it’s not like Mighty will turn your iffy 25/3 connection into a gigabit one; while Doshi tells me it’ll technically work with a 20Mbps connection, he says he’s targeting 80+Mbps households right now.
Then again, it’s not like everyone has a real choice of internet service provider, no matter how much money they make. As Jürgen Geuter (aka tante) points out below, this feels more like an indictment than innovation. It’s been a decade, and we still haven’t solved these problems.
“Streaming your browser to you because rendering the HTML is too slow on your machine” is not innovation but a mark of shame on everyone building websites and browsers.
Tech failed as an industry. https://t.co/JJC0WomArb
— tante (@tante) April 28, 2021
I agree with my colleague Tom: I genuinely want to know who’d actually pay for this and why. Would you?
I want to meet whoever is going to spend $30 a month to stream a Chromium browser from the cloud just to avoid RAM hungry Chrome https://t.co/4pl6jL2zUV
TP-Link’s mid-range AX6000 offers Wi-Fi 6, 2.5Gb WAN and lots of ports at a reasonable $270. While the software is simple and 2.4 Ghz performance isn’t great, this is a solid option, if you don’t expect too much.
For
+ Affordable Wi-Fi 6 option
+ Abundant Ethernet ports
+ Ease of setup and interface
+ Supports Link Aggregation
+ Excellent compatibility across all clients tested
Against
– Dual-band only
– No Wi-Fi 6E
– Simplistic software
– Unexceptional 2.4 GHz throughput speeds
TP-Link’s Archer AX6000 Next-Gen Wi-Fi Router sits somewhere between a full-tilt gaming router like the Asus ROG Rapture GT-AXE11000 and basic budget options like TP-Link’s sub-$70 Archer A7 AC1750. The Archer AX6000’s ‘middle of the road’ approach results in a router that’s larger than some, with lots of antennae (8) and gigabit Ethernet ports (8, plus a 2.5GB WAN). You also get Wi-Fi 6, though not the newer 6E that makes use of the less-cluttered 6 GHz band.
But the Archer AX6000 doesn’t feature the red accents and dancing LED’s that often adorn gaming routers. It also does not have a ridiculously high price tag. As of publication, it was selling for about $270–a far cry from flagship gaming options that often sell in the $500 range.
Is this more general-purpose approach to a router setup still perform well enough for serious gamers to take notice? Read on as we delve into the Archer AX6000’s full feature lest and performance testing to find out.
Design of the TP-Link Archer AX6000
(Image credit: Tom’s Hardware)
The AX6000 takes a more subtle approach to design than its flamboyant gaming counterparts (like the flagship Archer AX11000). While it comes with eight permanently attached antennae and a horizontal design, it goes with a black plastic exterior that looks more business than gaming. But there’s some flash here, in the form of an LED on the top center, behind a shiny gold TP-Link badge, that glows blue when all is working (and red when it isn’t). But the light is easily disabled with a dedicated hardware button–important if your router lives near your TV or in a bedroom.
AX6000 is on the larger end of the spectrum at 10.3 × 10.3 × 2.4 inches (261.2 × 261.2 × 60.2 mm), and it weighs 3.5 pounds (1.59 kg). That’s nearly as big as the Asus Rapture GT-AXE1000, at 10.4 x 10.4 x 2.9 inches and 3.94 pounds. But the Archer AX6000’s antennae are shorter than on most routers, and they come permanently attached, so all you have to do when taking the router out of the box is flip them up.
(Image credit: Tom’s Hardware)
Specifications of the TP-Link Archer AX6000
In terms of hardware specs, the AX6000 can definitely go toe-to-toe with higher-end routers. You get a 1.8 GHz quad-core processor with two additional coprocessors, 1 GB of RAM and 128 MB of flash storage.
Image 1 of 2
(Image credit: Tom’s Hardware)
Image 2 of 2
(Image credit: Tom’s Hardware)
Ports are also ample, with the WAN port a speedy 2.5 Gbps, plus eight gigabit LAN ports, basically giving you an integrated switch for plenty of wired connections. Link Aggregation is also supported, should you want to give extra bandwidth to a particular device. There are also a pair of two USB 3.0 ports (one Type-A and one Type-C) for connecting things like external storage to share on your network.
All that aside, the wireless department is where the AX6000 starts to show its mid-range limitations. The router is dual-band, rather than the more robust tri-band options found on higher-end gear. That means the AX6000 has a single 2.4 GHz option, rated at up to 1148 Mbps and a single 5 GHz for up to 4804 Mbps of throughput–hence (with the help of the usual rounding) the 6000 designation. Also, this router supports Wi-Fi 6 (802.11ax), but not the newest Wi-Fi 6E spec. So your devices get to live in the more crowded and mainstream bandwidths, rather than in the newly opened (and therefore much-less crowded) 6 GHz band.
That said, the aforementioned Asus ROG Rapture router is the only model we’ve tested so far that supports Wi-Fi 6E, and there are still very few devices (aside from some new phones) that are out there waiting for a 6E router to connect to.
Setup of the TP-Link Archer AX6000
(Image credit: Tom’s Hardware)
Setup of the TP-Link Archer AX6000 is fairly simple and took less than 10 minutes. After the requisite unboxing, we appreciate that the eight antennas come pre-attached, so just have to be rotated into their vertical position (although a few may be disappointed that they cannot be manually fine-tuned and adjusted to maximize reception).
Software setup can be done via the TP-Link app for the smartphone, or via the web browser, with the latter as our method of choice. We were given opportunities to choose a router password, the Wi-Fi password, and upgrade the firmware. This got us dialed in right away and was a painless and efficient process.
Features of the TP-Link Archer AX6000
We found the software for this router to be a mixed blessing. On the one hand, the interface is easy to use; we doubt more novice users will get confused using it. The converse for this is that compared to actual gaming routers, it’s less robust. You get fewer fine controls and not as much flexibility as some other offerings.
(Image credit: Tom’s Hardware)
The Quality of Service (QoS) offers a reasonable solution to allow for some adjustability. The choices include Gaming (which was used for our testing), Standard, Streaming, Surfing, and Chatting. There is also a Custom setting to allow for finer control of prioritization of traffic when your network needs don’t fit into a preset setting. There is also an option for Device Prioritization, so for example you can make sure your Gaming PC gets a higher priority than other clients on the network.
Security of the TP-Link Archer AX6000
(Image credit: Tom’s Hardware)
The security functionality on the Archer AX6000 also strikes a delicate balance between simple and capable. We do like that the AX6000 has integrated network security, and are glad that the subscription is included, taking it a notch above budget routers. The security is powered by Trend Micro, but the functions are limited to just three: a Malicious Content Filter, an Intrusion Prevention System, and Infected Device Quarantine. There’s also a Parental Controls function to limit screen time for younger family members.
Performance of the TP-Link Archer AX6000
2.4 GHz near
2.4 GHz far
5 GHz near
5 GHz far
132 Mbps
76 Mbps
789 Mbps
301 Mbps
The TP-Link Archer AX6000 proved to be quite stable during our testing. We connected a variety of clients to it, including laptops with Intel chips, an iPhone, and multiple Android smartphones. All devices connected easily and consistently to the AX6000.
Throughput testing with an Intel AX201 Wi-Fi 6 chipset showed decent, although unexceptional Wi-Fi 6 speeds. At the time that this router was introduced the speeds were quite solid, but given newer Wi-Fi 6E gear and the significantly faster speeds offered, the AX6000 feels more mid-range than it did in 2019. While the 5 GHz near speed of 789 Mbps is certainly solid, the throughput on the 2.4 GHz of 132 Mbps when close, and 76 Mbps on the far test offered no significant advantage over even older 802.11ac routers.
Testing Configuration
QoS
FRAPS Avg
Max
8K Dropped Frames
Pingplotter Spikes
Latency
Ethernet
no
123.7
158
n/a
0
235
Ethernet + 10 8k videos
no
96.8
121
34.90%
3
259
Ethernet + 10 8k videos
yes
119.4
146
28.80%
2
259
5 GHz
no
127.3
149
n/a
0
249
5 GHz + 10 8k videos
no
119.3
147
14.10%
0
275
5 GHz + 10 8k videos
yes
120.7
140
49.90%
1
257
2.4 GHz
no
123.6
156
n/a
0
68
2.4 GHz + 10 8k videos
no
45.2
91
16.30%
9
156
2.4 GHz + 10 8k videos
yes
88.8
123
47.20%
11
306
Again, the AX6000 showed its strengths and weaknesses in our testing. We did obtain some definitely fast gaming scores on our test game of Overwatch, such as 123.7 fps when wired, and an even slightly faster frame rate of 127.3 fps when connected via 5 GHz, both with nothing else running.
From a gameplay standpoint, when connected via 5 GHz, the fps were well maintained, both without the QoS at 119.3 FPS, and a slightly faster, and close to wired speed of 120.7 fps with the QoS activated with our ten 8K videos streaming in the background.
We also had some concerns, such as the high latency on all of the wired connection testing situations that went as high 259 milliseconds. The 2.4 GHz gaming score with background video congestion without QoS dragged to a much slower 45.2 fps. While activating the QoS improved the frame rate to 88.8 fps, we were disappointed to see how much the video streaming got sacrificed as the dropped frames went from 16.3% to a stuttering 47.2%.
Pricing of the TP-Link Archer AX6000
A strong point of this AX6000 is the pricing. At a suggested retail price of $299, it’s an affordable way to acquire a robust higher-end Wi-Fi 6 router. And street price seems to consistently hover at $269 at the time of this writing, dropping this router more solidly into the mid-range category in terms of pricing.
Conclusion
Image 1 of 2
(Image credit: Tom’s Hardware)
Image 2 of 2
(Image credit: Tom’s Hardware)
The TP-Link AX6000 is a decent mid-range router, offering good value for the dollar. We like the solid 5 GHz throughput speeds, the high fps gaming scores when connected via Ethernet or 5 GHz, the integrated 8 port switch, and the included security subscription.
But there are some shortcomings, including slower 2.4 GHz throughput (at this point only really an issue with older devices), the high dropped frame rate on video streaming with QoS activated for gaming, the simplified interface, and the lack of tri-band. Overall, for a general-purpose, mid-range router, there is plenty to like about the AX6000, including decent overall gaming performance.
But if maximum gaming speed is your priority and your network is often congested by lots of people and devices vying for bandwidth, you may want to spend more on something that handles this better, with less latency and frame drops. TP-Link’s own Archer AX11000 performed better on that front for about $100 more, while also delivering an extra 5 GHz band to ease congestion. But that router also suffered from a fair amount of dropped frames in our congestion testing.
The M1-powered Mac Mini can now be configured to included 10-gigabit networking, an option that the computer was notably missing compared to its Intel-powered counterpart (via 9to5Mac). The update to the 2020 computer seems to have come during Apple’s store update that came with the bigger product announcements today, including a new iMac and iPad Pros. Adding the faster networking will add $100 to the price of the computer — the same price as if you were to add 10Gbps to the Intel version.
When Apple added the 10Gbps port as an option to the Mac Mini in 2018, it was welcomed by those running home and professional servers, as well as anyone who needed fast network access from a cheaper Mac (like video professionals using network-attached storage). But when Apple introduced a faster version of the computer with its own custom chip, the ability to add networking faster than the standard Gigabit Ethernet was no longer there. If the lack of 10Gb networking took the (as of now) least expensive M1 computer out of the running for you, it’s now an option again.
You can order the Apple Silicon Mac Mini with faster networking today, but it may be a bit of a wait for the configuration — Apple estimates that I could have a stock M1 Mac Mini shipped to me in two days, but adding the 10Gb ethernet turns that into two weeks.
Nowadays there are loads of small form-factor (SFF) systems featuring fairly high performance, there are also fanless PCss that can offer performance of regular desktops. Unfortunately, SFF and fanless worlds rarely intersect and passively cooled compact desktops are extremely rare. Yet, they exist. Recently Atlast! Solutions introduced its Sigao Model B, which packs Intel’s 10-core Comet Lake CPU into a fairly small fanless chassis.
Image 1 of 5
(Image credit: Atlast! Solutions)
Image 2 of 5
(Image credit: Atlast! Solutions)
Image 3 of 5
(Image credit: Atlast! Solutions)
Image 4 of 5
(Image credit: Atlast! Solutions)
Image 5 of 5
(Image credit: Atlast! Solutions)
The Atlast! Sigao Model B is based around Intel’s 10-core Core i9-10900T processor as well as an Asus H470-I Mini-ITX motherboard. The CPU features a 35W TDP and has a base clock of 1.9 GHz as well as a maximum turbo frequency of up to 4.6 GHz, though we would not expect the processor to hit very high clocks in a fanless system powered by a 200W PSU. The motherboard comes with all the essentials, including Wi-Fi 6 and Bluetooth module, two Gigabit Ethernet ports, three display outputs (DisplayPort, HDMI, USB 3.2 Gen 2 Type-C), one USB 3.2 Gen 2 Type-A connector, four USB 3.2 Gen 1 Type-A ports, and 5.1-channel audio.
The Sigao measures 12.6 x 12.6 x 3.4 inches (320 × 322 × 87.5 mm) without feet, so while it is definitely not as compact as Intel’s NUC or Apple’s Mac Mini, it can still be considered a small form-factor PC.
Atlast! builds its fanless systems to order, so it can equip its Sigao Model B with up to 64GB of DDR4-2666 memory, one Samsung 970 Evo Plus M.2 SSD with a PCIe 3.0 x4 interface and up to 2TB capacity, and two 2.5-inch HDDs or SSDs.
The motherboard has a PCIe 3.0 x16 slot and the system can accommodate a single slot wide add-in card using a riser, though finding a decent mini-ITX 75W single slot graphics card with passive (or even active) cooling is close to impossible, so it is unlikely that the system can be equipped with a standalone AIB. Unfortunately, the motherboard also lacks a Thunderbolt 4 port for an external graphics solution, so it looks like the Sigao Model B has to rely on Intel’s built-in UHD Graphics 630 based on the previous-generation architecture. Meanwhile, if the Asus H470-I motherboard gains Rocket Lake-S support, it should be possible to install a more up-to-date CPU with Xe Graphics featuring leading-edge media playback capabilities.
The Atlast! Sigao Model B is not cheap at all. Even the basic model featuring a Core i9-10900T, 16GB of RAM, and a 250GB SSD costs €1,922 ($2304) with taxes and €1,602 without ($1,920), which is quite expensive even by SFF standards. But a desktop PC that brings together compact dimensions and passive cooling is hard to come by, so its price seems to be justified for those who want both features.
Seeed Studio announced the reTerminal, its latest addon for the Raspberry Pi Compute Module 4. This $195 Compute Module 4 machine learning and industrial application carrier board comes enclosed inside a robust case and sports a 5-inch touchscreen and plenty of expansion possibilities.
Image 1 of 5
(Image credit: Seeed)
Image 2 of 5
(Image credit: Seeed)
Image 3 of 5
(Image credit: Seeed)
Image 4 of 5
(Image credit: Seeed)
Image 5 of 5
(Image credit: Seeed)
Seeed’s reTerminal works with all models of the Raspberry Pi Compute Module 4. For the $195 price tag, it comes with a Compute Module 4 with 4GB RAM and 32GB of eMMC flash. You can easily swap out the Compute Module 4 if you need to change the config to accommodate your use case.
The 5-inch capacitive touch LCD has a 1280 x 720 (293 PPI) resolution, much more useful than 800 x 480 resolutions found on other screens. On the front of the unit, we see four user buttons, which can be programmed as required for useful quick access to common tasks.
On reTerminal’s left side, we find two USB 2.0 ports, Gigabit Ethernet, micro HDMI output (4Kp60), and a USB C power input for the 5V 4A required to power the unit. On the other side, we see a breakout for the Raspberry Pi GPIO. This is not directly compatible with add ons/HATs, but we can use one of the many Raspberry Pi add-on boards with a breakout board.
On the rear, we see a finned radiator used to keep the Compute Module 4’s 1.5 GHz quad-core Cortex-A72 CPU cool. We can only assume that this is passive cooling as there is no mention of active cooling. Also present is a MIPI camera interface, which enables using the official Raspberry Pi Camera and the HQ Camera. A useful addition when we consider that reTerminal is aimed at machine learning applications such as Edge Impulse, PyTorch and TensorFlow. An interesting addition is an industrial high-speed interface, which provides a PCIe 1 lane host, an additional USB 2.0 port, 28 GPIOs, and Power over Ethernet, according to the press release.
A series of M4 screw mounts at various points around the device are used to secure the reTerminal to several external modular accessories. We can see an industrial GPIO breakout from the supplied images and a unit that connects to the industrial high-speed breakout connector. We find a standard camera mount at the bottom of the unit, enabling reTerminal to be held in place using many different camera accessories.
A series of sensors are hidden inside of the reTerminal. First, we have a mic array that can be used with machine learning for voice recognition projects, an accelerometer for display rotation, data collection and gesture input, and a light sensor with proximity detection. An onboard real-time clock provides a means to keep your Compute Module 4’s clock in time, useful for projects where connectivity to NTP servers could be problematic.
Seeed’s reTerminal looks to be a nice and tidy means to use a Raspberry Pi Computer Module 4-powered machine for machine learning. Tom’s Hardware will have all of the reTerminal information you need in a future review. Seeed is offering $5 off if you register here.
A potential new Raspberry Pi competitor will soon be announced, according to a post on embedded systems blog CNX-Software. Based on the open-source RISC-V instruction set architecture, the Allwinner D1 development board is a credit-card-sized single board computer with a single core XuanTie C906 64-bit RISC-V CPU running at 1GHz along with 1GB of DDR3 RAM.
(Image credit: CNX-software)
The board has a lot in common with the single-board computers we already know and love, though seems to be aimed at the less powerful end of the spectrum. The Raspberry Pi 4 has this board beat for pure computational power. There’s the familiar 40-pin GPIO which we assume is compatible with layout used on the Raspberry Pi but that remains to be seen. Also present on the board is a single full size HDMI 1.4 port, Wi-Fi 4 and Bluetooth, two USB Type-C ports and a single USB 2.0 Type-A, a Micro SD slot, Gigabit Ethernet, a 3.5mm audio jack, plus camera and display connectors which look to be the same format as the CSI and DSI connectors found on the Raspberry Pi. Interestingly, there’s 256MB of onboard flash storage, and a four-pin UART header for debugging.
Allwinner is positioning the board as a ‘multi-media decoding platform’, and CNX’s figures show it’s good for H.265 up to 1080p60 or 4Kp30, and H.264 up to 1080p60 or 4Kp24. The display header can run a touchscreen at up to 1080p60, and the HDMI port hits the 1.4 standard.
The Allwinner D1 Linux RISC-V has the same dimensions as a Raspberry Pi 4 at 3.3 x 2.2 inches (85 x 56 mm) but the overall layout is different enough to prevent Raspberry Pi cases from being used. Beyond its video engine, the Allwinner D1 Linux RISC-V development board doesn’t have much in the way of graphics processing. A quoted Coremark score of 3.8/Mhz doesn’t compare well with the Raspberry Pi 4’s 15.1/MHz, but we’d expect this board to be considerably cheaper, and it could be interesting for use in smart displays or networked cameras. Allwinner has its own Debian-based Linux distribution, Tina OS, though its Github page was last updated in 2017, so we’re hopeful for a newer kernel to go with this new board.
Right now that is all we know, but we are hopeful that this board will introduce the RISC-V platform to more makers and lead to more powerful machines coming in the near future.
TP-Link has announced a new family of multi-gigabit switches designed primarily for homes and small offices. The 2.5GbE and 10GbE switches are reasonably priced — one 2.5GbE model is only $130 — and are designed for advanced users that need faster wired network speeds as they use multi-gig NAS, Wi-Fi 6 access points, and other bandwidth-hungry devices.
The newly announced family of switches includes the TL-SG105-M2 5-port 2.5GbE desktop switch, the TL-SG108-M2 8-port 2.5GbE desktop switch, and the TL-SG3210XHP-M2 JetStream managed switch that has eight RJ45 2.5GbE ports as well as two SFP+ 10GbE ports.
The switches are backward compatible with 100Mbps and GbE over CAT5 copper cables, yet to hit 10GbE speeds, the highest-end model requires SFP+ cables, which are not common at homes or SMB offices.
(Image credit: TP-Link)
The 5-port TL-SG105-M2 and the 8-port TL-SG108-M2 switches are rather basic fanless devices that can automatically sense link speeds and intelligently fine-tune for compatibility and optimal performance for all devices. The switches come in metal chassis and will fit almost any home design (assuming, of course, that they are not hidden).
(Image credit: TP-Link)
The 5-port switch is immediately available for $130, whereas the 8-port model carries a $200 MSRP.
(Image credit: TP-Link)
Being a managed switch, the TL-SG3210XHP-M2 JetStream is of course inherently more advanced than its cheaper counterparts. The unit has eight 2.5GbE 802.3at/af-compliant PoE+ ports, two 10 Gbps SFP+ slots, and two ports for management (an RJ45 and a mini-USB). TP-Link positions this switch for relatively large networks employing both wired and wireless clients. Since it is supposed to have rather serious switching capacity, it uses up to 240W of power and has active cooling.
The switch supports a host of security capabilities, including IP-MAC-Port binding, ACL, Port Security, DoS defend, Storm control, DHCP snooping, 802.1X and radius authentication. L2/L3/L4 QoS and IGMP snooping for voice and video applications. In addition, the switch is Omada SDN compatible and features Zero-Touch Provisioning and intelligent monitoring.
The TL-SG3210XHP-M2 JetStream is of course not exactly cheap, but for $350, it is certainly not expensive.
After almost a decade of total market dominance, Intel has spent the past few years on the defensive. AMD’s Ryzen processors continue to show improvement year over year, with the most recent Ryzen 5000 series taking the crown of best gaming processor: Intel’s last bastion of superiority.
Now, with a booming hardware market, Intel is preparing to make up some of that lost ground with the new 11th Gen Intel Core Processors. Intel is claiming these new 11th Gen CPUs offer double-digit IPC improvements despite remaining on a 14 nm process. The top-end 8-core Intel Core i9-11900K may not be able to compete against its Ryzen 9 5900X AMD rival in heavily multi-threaded scenarios, but the higher clock speeds and alleged IPC improvements could be enough to take back the gaming crown. Along with the new CPUs, there is a new chipset to match, the Intel Z590. Last year’s Z490 chipset motherboards are also compatible with the new 11th Gen Intel Core Processors, but Z590 introduces some key advantages.
First, Z590 offers native PCIe 4.0 support from the CPU, which means the PCIe and M.2 slots powered off the CPU will offer PCIe 4.0 connectivity when an 11th Gen CPU is installed. The PCIe and M.2 slots controlled by the Z590 chipset are still PCI 3.0. While many high-end Z490 motherboards advertised this capability, it was not a standard feature for the platform. In addition to PCIe 4.0 support, Z590 offers USB 3.2 Gen 2×2 from the chipset. The USB 3.2 Gen 2×2 standard offers speeds of up to 20 Gb/s. Finally, Z590 boasts native support for 3200 MHz DDR4 memory. With these upgrades, Intel’s Z series platform has feature parity with AMD’s B550. On paper, Intel is catching up to AMD, but only testing will tell if these new Z590 motherboards are up to the challenge.
The ASRock Z590 Phantom Gaming Velocita is a recent addition to ASRock’s arsenal. The Phantom Gaming Velocita targets the gamer market with Killer Networking for both wired and wireless connectivity and even an option to route the network traffic straight from the Killer LAN controller to the CPU. The ASRock Z590 Phantom Gaming Velocita features a dependable 14-phase VRM that takes advantage of 50 A power stages from Vishay. The ASRock Z590 Phantom Gaming Velocita has all the core features for a great gaming motherboard. All that is left is to see how the ASRock Z590 Phantom Gaming Velocita stacks up against the competition!
1x Killer E3100G 2.5 Gb/s LAN 1x Intel I219V Gigabit LAN 1x Killer AX1675x WiFi 6E
Rear Ports:
2x Antenna Ports 1x HDMI Port 1x DisplayPort 1.4 1x Optical SPDIF Out Port 1x USB 3.2 Gen2 Type-A Port 1x USB 3.2 Gen2 Type-C Port 6x USB 3.2 Gen1 Type-A Ports 2x USB 2.0 Ports 2x RJ-45 LAN Ports 5x 3.5 mm HD Audio Jacks
NUC 11 Extreme Compute Element (Image credit: Chiphell)
The Intel NUC 11 Extreme Compute Element (codename Driver Bay) might be right around the corner. A user from the Chiphell forums has shared a screenshot of the alleged specifications for the device, which appears to leverage Intel’s forthcoming 11th Generation Tiger Lake-H 45W chips.
If the information is legit, the NUC 11 Extreme Compute Element will be available with three processor options that may come in the shape of the Core i9-11980HK, Core i7-11800H or Core i5-11400H. The Core i9 and Core i7 Tiger Lake-H 45W chips will arrive with eight Willow Cove cores, while the Core i5 will stick to six cores. All three have Hyper-Threading technology, of course.
The NUC 11 Extreme Compute Element can be outfitted with up to 64GB of DDR4-3200 dual-channel memory. It also comes equipped with three PCIe 4.0 x4 M.2 slots with support for M.2 drives up to 80mm in length. One of the M.2 slots communicates directly with the Tiger Lake-H processor, while the remaining two are attached to the PCH itself. There is support for RAID 0 and RAID 1 arrays. The unit is compatible with Intel’s Optane Memory as well.
NUC 11 Extreme Compute Element (Image credit: Chiphell)
If the Tiger Lake-H chip isn’t paired with a discrete graphics option, then the Xe LP graphics engine will do all the heavy lifting. The NUC 11 Extreme Compute Element provides one HDMI 2.0b port and two Thunderbolt 4 ports so you can connect up to three 4K monitors to the device.
Depending on the SKU, the NUC 11 Extreme Compute Element may sport 2.5 Gigabit Ethernet and/or 10 Gigabit Ethernet ports. It also offers Wi-Fi 6 and Bluetooth 5 connectivity.
The NUC 11 Extreme Compute Element may be a tiny device, but it supplies plenty of USB ports. There are a total of six USB 3.1 Gen 1 Type-A ports as well as two USB 3.1 headers and two USB 2.0 headers. Its audio capabilities include 7.1 multichannel audio that’s made possible through the HDMI or DisplayPort signals.
Since the NUC 11 Extreme Compute Element is based on Tiger Lake-H 45W processors, it’s reasonable to expect the device to hit the market once Intel officially launches the aforementioned chips. Tiger Lake-H 45W laptops are expected to land in the second quarter of this year so the NUC 11 Extreme Compute Element shouldn’t be far behind.
Asus has listed the ROG Maximus XIII Apex on its website, implying that the successor to the ROG Maximus XII Apex may be closer than we think. The new iteration to the Apex series has been engineered to tame Intel’s 11th Generation Rocket Lake processors.
Built around the new Z590 chipset and existing LGA1200 socket, the ROG Maximus XIII Apex comes equipped with an 18-phase power delivery subsystem. Each power stage, which can manage up to 90 amps, is accompanied by a MicroFine Alloy choke that can do 45 amps. Asus revamped the power design on the ROG Maximus XIII Apex completely by getting rid of the phase doublers. The motherboard also employs 10K Japanese black metallic capacitors that can take a beating. The VRM area is properly cooled with thick, aluminum passive heatsinks. The ROG Maximus XIII Apex feeds Rocket Lake chips through a pair of 8-pin EPS power connectors.
The overclocking toolkit on the ROG Maximus XIII Apex includes a double-digit debug LED, voltage read points, and a plethora of buttons and switches to aid in overclocking. There are also three condensation sensors that are placed strategically across the motherboard to notify you when condensation occurs around the processor, memory or PCIe slot. In total, the ROG Maximus XIII Apex has five temperature sensors, five 4-pin fan headers, two full-speed fan headers, and an assortment of headers for watercooling setups.
Like previous Apex motherboards, the ROG Maximus XIII Apex only provides two DDR4 memory slots. While memory capacity is limited to 64GB, the motherboard supports memory frequencies above DDR4-5000 with ease. The ROG Maximus XIII Apex sports Asus’ OptiMem III technology, featuring an optimized memory tracing layout to improve memory overclocking.
Image 1 of 3
ROG Maximus XIII Apex (Image credit: Asus)
Image 2 of 3
ROG Maximus XIII Apex (Image credit: Asus)
Image 3 of 3
ROG Maximus XIII Apex (Image credit: Asus)
The ROG Maximus XIII Apex offers numerous options for storage. It provides eight SATA III ports and up to four M.2 slots. The M.2 slots on the motherboard are PCIe 4.0 ready and come armed with an aluminium heatsink and embedded backplates to provide passive cooling. The other two M.2 slots reside on Asus’ ROG DIMM.2 module that connects to the motherboard through a DDR4-type interface beside the memory slots. The DIMM.2 module accommodates M.2 drives with lengths up to 110mm.
The expansion slots on the ROG Maximus XIII Apex consist of two PCIe x16 slots and one PCIe x8 slot. Wired and wireless networking come in the shape of a 2.5 Gigabit Ethernet port and Wi-Fi 6E connectivity with support for up to 6GHz bands. The audio system on the ROG Maximus XIII Apex uses Realtek’s ALC4080 audio codec complemented with a Savitech SV3H712 amplifier and high-end Nichicon audio capacitors.
In regards to USB ports, the ROG Maximus XIII Apex has four USB 3.2 Gen 1 ports, five USB 3.2 Gen 2 ports and one USB 3.2 Gen 2×2 Type-C port at the rear panel. There’s an additional USB 3.2 Gen 2×2 header on the motherboard. The ROG Maximus XIII Apex doesn’t supply any display outputs so it’s mandatory to pair it with a discrete graphics card.
ROG motherboards have a very rich software suite. On this iteration, Asus has directly implemented MemTest86 into the ROG Maximus XIII Apex’s firmware so overclockers can test memory stability without any hassles. Additionally, a one-year AIDA64 Extreme subscription is also included.
The pricing for the ROG Maximus XIII Apex is currently unknown. The previous Z490 version retailed for $356.99, so we can expect the Z590 followup to be price around that range if not a little bit higher.
The European Union wants to double its chip manufacturing output to 20 percent of the global market by 2030. The goal is part of its new Digital Compass plan, announced yesterday, which aims to boost “digital sovereignty” by funding various high-tech initiatives.
As well as doubling chip output, the EU also wants all households to have 5G access and gigabit internet connectivity by 2030; for “all key public services” to be available online in every member state; and for the bloc to have its first quantum computer. Funding for these and other projects will come from the EU’s €672.5 billion ($800 billion) coronavirus response fund, with 20 percent of this money ($160 billion) earmarked for tech investment.
The EU’s ambition to produce more semiconductor chips is particularly notable. Maintaining a steady supply of these chips has become a pressing concern for nations around the world as supply chain disruptions caused by the pandemic and the US-China trade war have affected global supplies. As with the flow of key resources like oil, access to cutting-edge chips is essential for many industries and products, from iPhones to cars. Currently, the bulk of production is concentrated in Asia, particularly in Taiwan and Korea.
In February, President Joe Biden signed an executive order to investigate how the US can further support its own chip manufacturing industry. “This is about making sure the United States can meet every challenge we face in this new era of pandemics, but also in defense cybersecurity, climate change, and so much more,” said Biden at a press conference.
US President Joe Biden pledged in February to secure the United States’ “critical” chip supply chains.Image: Saul Loeb/AFP via Getty Images
The EU’s worries mirror those of America’s. “We need to become less dependent on others when it comes to key technologies,” vice president of the European Commission, Margrethe Vestager, said at the launch of the Digital Compass plans, reports The Wall Street Journal.
This dependency extends beyond chip supply, too. The last few years have seen the EU grapple with its relationship with Chinese tech giant Huawei, which supplies essential 5G components but has been hamstrung by US sanctions introduced over national security concerns. In related news, Apple announced this week that it will invest more than €1 billion ($1.2 billion) in a silicon design center in Germany. Its employees will focus on “5G and future wireless technologies.”
Building up the EU’s chip production, though, will be a difficult task. Manufacturing semiconductors is an extremely expensive business, and the main success of European companies has been creating machinery used in this process rather than the chips themselves. Dutch firm ASML, for example, has a majority market share in the production of equipment known as “photolithographic machines” that are essential for chipmaking.
“As China has shown, throwing money at chips does not guarantee success,” Dan Wang, a technology analyst at Gavekal Dragonomics, told the WSJ. “For the last few decades, Europe has seen its number of semiconductor companies shrink, and it will require a mighty effort to wrest leadership from the US and Asia, which are also investing heavily.”
Supermicro’s 1023US-TR4 is a slim 1U dual-socket server designed for high-density compute environments in high-end cloud computing, virtualization, and enterprise applications. With support for AMD’s EPYC 7001 and 7002 processors, this high-end server packs up to two 64-core Eypc Rome processors, allowing it to cram 128 cores and 256 threads into one slim chassis.
We’re on the cusp of Intel’s Ice Lake and AMD’s EPYC Milan launches, which promise to reignite the fierce competition between the long-time x86 rivals. In preparation for the new launches, we’ve been working on a new set of benchmarks for our server testing, and that’s given us a pretty good look at the state of the server market as it stands today.
We used the Supermicro 1023US-TR4 server for EPYC Rome testing, and we’ll focus on examining the platform in this article. Naturally, we’ll add in Ice Lake and EPYC Milan testing as soon as those chips are available. In the meantime, here’s a look at some of our new benchmarks and the current state of the data center CPU performance hierarchy in several hotly-contested price ranges.
Inside the Supermicro 1023US-TR4 Server
Image 1 of 4
(Image credit: Tom’s Hardware)
Image 2 of 4
(Image credit: Tom’s Hardware)
Image 3 of 4
(Image credit: Tom’s Hardware)
Image 4 of 4
(Image credit: Tom’s Hardware)
The Supermicro 1023US-TR4 server comes in the slim 1U form factor. And despite its slim stature, it can host an incredible amount of compute horsepower under the hood. The server supports AMD’s EPYC 7001 and 7002 series chips, with the latter series topping out at 64 cores apiece, which translates to 128 cores and 256 threads spread across the dual sockets.
Support for the 7002 series chips requires a 2.x board revision, and the server can accommodate CPU cTDP’s up to 280W. That means it can accommodate the beefiest of EPYC chips, which currently comes in the form of the 280W 64-core EPYC 7H12 with a 280W TDP.
The server has a tool-less rail mounting system that eases installation into server racks and the CSE-819UTS-R1K02P-T chassis measures 1.7 x 17.2 x 29 inches, ensuring broad compatibility with standard 19-inch server racks.
The front panel comes with standard indicator lights, like a unit identification (UID) light that helps with locating the server in a rack, along with drive activity, power, status light (to indicate fan failures or system overheating), and two LAN activity LEDs. Power and reset buttons are also present at the upper right of the front panel.
By default, the system comes with four tool-less 3.5-inch hot-swap SATA 3 drive bays, but you can configure the server to accept four NVMe drives on the front panel, and an additional two M.2 drives internally. You can also add an optional SAS card to enable support for SAS storage devices. The front of the system also houses a slide-out service/asset tag identifier card to the upper left.
Image 1 of 7
(Image credit: Tom’s Hardware)
Image 2 of 7
(Image credit: Tom’s Hardware)
Image 3 of 7
(Image credit: Tom’s Hardware)
Image 4 of 7
(Image credit: Tom’s Hardware)
Image 5 of 7
(Image credit: Tom’s Hardware)
Image 6 of 7
(Image credit: Tom’s Hardware)
Image 7 of 7
(Image credit: Supermicro)
Popping the top off the chassis reveals two shrouds that direct air from the two rows of hot-swappable fans. A total of eight fan housings feed air to the system, and each housing includes two counter-rotating 4cm fans for maximum static pressure and reduced vibration. As expected with servers intended for 24/7 operation, the system can continue to function in the event of a fan failure. However, the remainder of the fans will automatically run at full speed if the system detects a failure. Naturally, these fans are loud, but that’s not a concern for a server environment.
Two fan housings are assigned to cool each CPU, and a simple black plastic shroud directs air to the heatsinks underneath. Dual SP3 sockets house both processors, and they’re covered by standard heatsinks that are optimized for linear airflow.
A total of 16 memory slots flank each processor, for a total of 32 memory slots that support up to 4TB of registered ECC DDR4-2666 with EPYC 7001 processors, or an incredible 8TB of ECC DDR4-3200 memory (via 256GB DIMMs) with the 7002 models, easily outstripping the memory capacity available with competing Intel platforms.
We tested the EPYC processors with 16x 32GB DDR4-3200 Samsung modules for a total memory capacity of 512GB. In contrast, we loaded down the Xeon comparison platform with 12x 32GB Sk hynix DDR4-2933 modules, for a total capacity of 384GB of memory.
The H11DSU-iN motherboard’s expansion slots consist of two full-height 9.5-inch PCIe 3.0 slots and one low-profile PCIe 3.0 x8 slot, all mounted on riser cards. An additional internal PCIe 3.0 x8 slot is also available, but this slot only accepts proprietary Supermicro RAID cards. All told, the system exposes a total of 64 lanes (16 via NVMe storage devices) to the user.
As one would imagine, Supermicro has other server offerings that expose more of EPYCs available 128 lanes to the user and also come with the faster PCIe 4.0 interface.
Image 1 of 2
(Image credit: Tom’s Hardware)
Image 2 of 2
(Image credit: Tom’s Hardware)
The rear I/O panel includes four gigabit RJ45 LAN ports powered by an Intel i350-AM4 controller, along with a dedicated IPMI port for management. Here we find the only USB ports on the machine, which come in the form of two USB 3.0 headers, along with a COM and VGA port.
Two 1000W Titanium-Level (96%+) redundant power supplies provide power to the server, with automatic failover in the event of a failure, as well as hot-swapability for easy servicing.
The BIOS is easy to access and use, while the IPMI web interface provides a wealth of monitoring capabilities and easy remote management that matches the type of functionality available with Xeon platforms. Among many options, you can update the BIOS, use the KVM-over-LAN remote console, monitor power consumption, access health event logs, monitor and adjust fan speeds, and monitor the CPU, DIMM, and chipset temperatures and voltages. Supermicro’s remote management suite is polished and easy to use, which stands in contrast to other platforms we’ve tested.
Test Setup
Cores/Threads
1K Unit Price
Base / Boost (GHz)
L3 Cache (MB)
TDP (W)
AMD EPYC 7742
64 / 128
$6,950
2.25 / 3.4
256
225W
Intel Xeon Platinum 8280
28 / 56
$10,009
2.7 / 4.0
38.5
205W
Intel Xeon Gold 6258R
28 / 56
$3,651
2.7 / 4.0
38.5
205W
AMD EPYC 7F72
24 / 48
$2,450
3.2 / ~3.7
192
240W
Intel Xeon Gold 5220R
24 / 48
$1,555
2.2 / 4.0
35.75
150W
AMD EPYC 7F52
16 / 32
$3,100
3.5 / ~3.9
256
240W
Intel Xeon Gold 6226R
16 / 32
$1,300
2.9 / 3.9
22
150W
Intel Xeon Gold 5218
16 / 32
$1,280
2.3 / 3.9
22
125W
AMD EPYC 7F32
8 / 16
$2,100
3.7 / ~3.9
128
180W
Intel Xeon Gold 6250
8 / 16
$3,400
3.9 / 4.5
35.75
185W
Here we can see the selection of processors we’ve tested for this review, though we use the Xeon Platinum Gold 8280 as a stand-in for the less expensive Xeon Gold 6258R. These two chips are identical and provide the same level of performance, with the difference boiling down to the more expensive 8280 coming with support for quad-socket servers, while the Xeon Gold 6258R tops out at dual-socket support.
Memory
Tested Processors
Supermicro AS-1023US-TR4
16x 32GB Samsung ECC DDR4-3200
EPYC 7742, 7F72, 7F52, 7F32
Dell/EMC PowerEdge R460
12x 32GB SK Hynix DDR4-2933
Intel Xeon 8280, 6258R, 5220R, 6226R, 6250
To assess performance with a range of different potential configurations, we used the Supermicro 1024US-TR4 server with four different EPYC Rome configurations. We outfitted this server with 16x 32GB Samsung ECC DDR4-3200 memory modules, ensuring that both chips had all eight memory channels populated.
We used a Dell/EMC PowerEdge R460 server to test the Xeon processors in our test group, giving us a good sense of performance with competing Intel systems. We equipped this server with 12x 32GB Sk hynix DDR4-2933 modules, again ensuring that each Xeon chip’s six memory channels were populated. These configurations give the AMD-powered platform a memory capacity advantage, but come as an unavoidable side effect of the capabilities of each platform. As such, bear in mind that memory capacity disparities may impact the results below.
We used the Phoronix Test Suite for testing. This automated test suite simplifies running complex benchmarks in the Linux environment. The test suite is maintained by Phoronix, and it installs all needed dependencies and the test library includes 450 benchmarks and 100 test suites (and counting). Phoronix also maintains openbenchmarking.org, which is an online repository for uploading test results into a centralized database. We used Ubuntu 20.04 LTS and the default Phoronix test configurations with the GCC compiler for all tests below. We also tested both platforms with all available security mitigations.
Linux Kernel and LLVM Compilation Benchmarks
Image 1 of 2
(Image credit: Tom’s Hardware)
Image 2 of 2
(Image credit: Tom’s Hardware)
We used the 1023US-TR4 for testing with all of the EPYC processors in the chart, and here we see the expected scaling in the timed Linux kernel compile test with the AMD EPYC processors taking the lead over the Xeon chips at any given core count. The dual EPYC 7742 processors complete the benchmark, which builds the Linux kernel at default settings, in 21 seconds. The dual 24-core EPYC 7F72 configuration is impressive in its own right — it chewed through the test in 25 seconds, edging past the dual-processor Xeon 8280 platform.
AMD’s EPYC delivers even stronger performance in the timed LLVM compilation benchmark — the dual 16-core 7F72’s even beat the dual 28-core 8280’s. Performance scaling is somewhat muted between the flagship 64-core 7742 and the 24-core 7F72, largely due to the strength of the latter’s much higher base and boost frequencies. That impressive performance comes at the cost of a 240W TDP rating, but the Supermicro server handles the increased thermal output easily.
Molecular Dynamics and Parallel Compute Benchmarks
Image 1 of 6
(Image credit: Tom’s Hardware)
Image 2 of 6
(Image credit: Tom’s Hardware)
Image 3 of 6
(Image credit: Tom’s Hardware)
Image 4 of 6
(Image credit: Tom’s Hardware)
Image 5 of 6
(Image credit: Tom’s Hardware)
Image 6 of 6
(Image credit: Tom’s Hardware)
NAMD is a parallel molecular dynamics code designed to scale well with additional compute resources; it scales up to 500,000 cores and is one of the premier benchmarks used to quantify performance with simulation code. The EPYC processors are obviously well-suited for these types of highly-parallelized workloads due to their prodigious core counts, with the dual 7742 configuration completing the workload 28% faster than the dual Xeon 8280 setup.
Stockfish is a chess engine designed for the utmost in scalability across increased core counts — it can scale up to 512 threads. Here we can see that this massively parallel code scales well with EPYC’s leading core counts. But, as evidenced by the dual 24-core 7F72’s effectively tying the 28-core Xeon 8280’s, the benchmark also generally responds well to the EPYC processors. The dual 16-core 7F52 configuration also beat out both of the 16-core Intel comparables. Intel does pull off a win as the eight-core 6250 processors beat the 7F32’s, though.
We see similarly impressive performance in other molecular dynamics workloads, like the Gromacs water benchmark that simulates Newtonian equations of motion with hundreds of millions of particles and the NAS Parallel Benchmarks (NPB) suite. NPB characterizes Computational Fluid Dynamics (CFD) applications, and NASA designed it to measure performance from smaller CFD applications up to “embarrassingly parallel” operations. The BT.C test measures Block Tri-Diagonal solver performance, while the LU.C test measures performance with a lower-upper Gauss-Seidel solver.
Regardless of the workload, the EPYC processors deliver a brutal level of performance in highly-parallelized applications, and the Supermicro server handled the heat output without issue.
Rendering Benchmarks
Image 1 of 8
(Image credit: Tom’s Hardware)
Image 2 of 8
(Image credit: Tom’s Hardware)
Image 3 of 8
(Image credit: Tom’s Hardware)
Image 4 of 8
(Image credit: Tom’s Hardware)
Image 5 of 8
(Image credit: Tom’s Hardware)
Image 6 of 8
(Image credit: Tom’s Hardware)
Image 7 of 8
(Image credit: Tom’s Hardware)
Image 8 of 8
(Image credit: Tom’s Hardware)
Turning to more standard fare, provided you can keep the cores fed with data, most modern rendering applications also take full advantage of the compute resources. Given the well-known strengths of EPYC’s core-heavy approach, it isn’t surprising to see the 64-core EPYC 7742 processors carve out a commanding lead in the C-Ray and Blender benchmarks. Still, it is impressive to see the 7Fx2 models beat the competing Xeon processors with similar core counts nearly across the board.
The performance picture changes somewhat with the Embree benchmarks, which test high-performance ray tracing libraries developed at Intel Labs. Naturally, the Xeon processors take the lead in the Asian Dragon renders, but the crown renders show that AMD’s EPYC can offer leading performance even with code that is heavily optimized for Xeon processors.
Encoding Benchmarks
Image 1 of 3
(Image credit: Tom’s Hardware)
Image 2 of 3
(Image credit: Tom’s Hardware)
Image 3 of 3
(Image credit: Tom’s Hardware)
Encoders tend to present a different type of challenge: As we can see with the VP9 libvpx benchmark, they often don’t scale well with increased core counts. Instead, they often benefit from per-core performance and other factors, like cache capacity.
However, newer encoders, like Intel’s SVT-AV1, are designed to leverage multi-threading more fully to extract faster performance for live encoding/transcoding video applications. Again, we can see the impact of EPYC’s increased core counts paired with its strong per-core performance as the EPYC 7742 and 7F72 post impressive wins.
Python and Sysbench Benchmarks
Image 1 of 2
(Image credit: Tom’s Hardware)
Image 2 of 2
(Image credit: Tom’s Hardware)
The Pybench and Numpy benchmarks are used as a general litmus test of Python performance, and as we can see, these tests don’t scale well with increased core counts. That allows the Xeon 6250, which has the highest boost frequency of the test pool at 4.5 GHz, to take the lead.
Compression and Security
Image 1 of 3
(Image credit: Tom’s Hardware)
Image 2 of 3
(Image credit: Tom’s Hardware)
Image 3 of 3
(Image credit: Tom’s Hardware)
Compression workloads also come in many flavors. The 7-Zip (p7zip) benchmark exposes the heights of theoretical compression performance because it runs directly from main memory, allowing both memory throughput and core counts to impact performance heavily. As we can see, this benefits the EPYC 7742 tremendously, but it is noteworthy that the 28-core Xeon 8280 offers far more performance than the 24-core 7F72 if we normalize throughput based on core counts. In contrast, the gzip benchmark, which compresses two copies of the Linux 4.13 kernel source tree, responds well to speedy clock rates, giving the eight-core Xeon 6250 the lead due to its 4.5 GHz boost clock.
The open-source OpenSSL toolkit uses SSL and TLS protocols to measure RSA 4096-bit performance. As we can see, this test favors the EPYC processors due to its parallelized nature, but offloading this type of workload to dedicated accelerators is becoming more common for environments with heavy requirements.
SPEC CPU 2017 Estimated Scores
Image 1 of 4
(Image credit: Tom’s Hardware)
Image 2 of 4
(Image credit: Tom’s Hardware)
Image 3 of 4
(Image credit: Tom’s Hardware)
Image 4 of 4
(Image credit: Tom’s Hardware)
We used the GCC compiler and the default Phoronix test settings for these SPEC CPU 2017 test results. SPEC results are highly contested and can be impacted heavily with various compilers and flags, so we’re sticking with a bog-standard configuration to provide as level of a playing field as possible. It’s noteworthy that these results haven’t been submitted to the SPEC committee for verification, so they aren’t official. Instead, view the above tests as estimates, based on our testing.
The multi-threaded portion of the SPEC CPU 2107 suite is of most interest for the purpose of our tests, which is to gauge the ability of the Supermicro platform to handle heavy extended loads. As expected, the EPYC processors post commanding leads in both the intrate and fprate subtests. And close monitoring of the platform didn’t find any thermal throttling during these extended duration tests. The Xeon 6250 and 8280 processors take the lead in the single-threaded intrate tests, while the AMD EPYC processors post impressively-strong single-core measurements in the fprate tests.
Conclusion
AMD has enjoyed a slow but steadily-increasing portion of the data center market, and much of its continued growth hinges on increasing adoption beyond hyperscale cloud providers to more standard enterprise applications. That requires a dual-pronged approach of not only offering a tangible performance advantage, particularly in workloads that are sensitive to per-core performance, but also having an ecosystem of fully-validated OEM platforms readily available on the market.
The Supermicro 1023US-TR4 server slots into AMD’s expanding constellation of OEM EPYC systems and also allows discerning customers to upgrade from the standard 7002 series processors to the high-frequency H- and F-series models as well. It also supports up to 8TB of ECC memory, which is an incredible amount of available capacity for memory-intensive workloads. Notably, the system comes with the PCIe 3.0 interface while the second-gen EPYC processors support PCIe 4.0, but this arrangement allows customers that don’t plan to use PCIe 4.0 devices to procure systems at a lower price point. As one would imagine, Supermicro has other offerings that support the faster interface.
Overall we found the platform to be robust, and out-of-the-box installation was simple with a tool-less rail kit and an easily-accessible IPMI interface that offers a cornucopia of management and monitoring capabilities. Our only minor complaints are that the front panel could use a few USB ports for easier physical connectivity. The addition of a faster embedded networking interface would also free up an additional PCIe slot. Naturally, higher-end Supermicro platforms come with these features.
As seen throughout our testing, the Supermicro 1023US-TR4 server performed admirably and didn’t suffer from any thermal throttling issues regardless of the EPYC processors we used, which is an important consideration. Overall, the Supermicro 1023US-TR4 server packs quite the punch in a small form factor that enables incredibly powerful and dense compute deployments in cloud, virtualization, and enterprise applications.
It may have a name that sounds like a 1960s spy plane, but the Geniatech XPI-3288 is a single board computer which looks like it has just stepped out of the Raspberry Pi factory, but it has a few differences that make it stand out from the crowd.
Image 1 of 4
(Image credit: Geniatech)
Image 2 of 4
(Image credit: Geniatech)
Image 3 of 4
(Image credit: Geniatech)
Image 4 of 4
(Image credit: Geniatech)
Measuring the same as a Raspberry Pi 4, 3.3 x 2.2 inches (85 x 56mm), the XPI-3288 looks very similar to the Pi, in fact the placement of the four USB 2.0 ports and Gigabit Ethernet is the same as the Raspberry Pi 3B. Powering the board is an Rockchip RK3200 SoC which features a quad core Arm Cortex A17 running at up to 1.6 GHz and a Mali-T764 GPU.
What is different to a Raspberry Pi is onboard eMMC flash storage, available in 8, 16 and 32GB capacities and a decidedly “tacked on” optional WiFi jack which covers part of the CPU and prevents the easy use of heatsinks to cool the unit. The WiFi Jack appears to have a PCB antenna, but there are also breakouts for external WiFi and Bluetooth antennas.
The placement of CSI / DSI connectors, used on the Raspberry Pi for cameras and the official display, is nearer to the USB ports and we assume that they are compatible with the official Raspberry Pi accessories.
The XPI-3288 comes with a single HDMI port, micro USB power input, micro SD support and a 40 pin GPIO which claims compatibility with the Raspberry Pi GPIO but that relies on the many software libraries that enable control of the GPIO to be ported and tweaked to work with the XPI-3288. Talking of software, the XPI-3288 is designed for use with Android 7.1 and “Linux”. Right now there are no details on which distros are supported, merely a reference to Linux Kernel 4.4.
The Geniatech XPI-3288 is being offered for sale at $75. The same spec Raspberry Pi 4 with 2GB of RAM is $40 cheaper!
Asus just announced the ProART B550-Creator, the first AM4 motherboard to come with Thunderbolt 4 support. The motherboard will be available next month with a price tag of $299.
Like its other ProART motherboards, the ProART B550-Creator arrives with a minimalistic design that features a black exterior accented by gold accents. The motherboard employs a powerful 12+2-phase power delivery subsystem that’s cooled by a pair of thick heatsinks. The motherboard feeds the processor with a combination of a 8-pin EPS and 4-pin ATX power connectors. On the memory side, the ProART B550-Creator is equipped with four DDR4 memory slots and accommodates up to 128GB of memory. However, Asus didn’t specify up to what memory frequencies are supported.
Storage-wise, the ProART B550-Creator offers four normal SATA III connectors for conventional hard drives and SSDs. There are also a pair of M.2 ports for high-speed storage. Logically, the primary M.2 port adheres to the PCIe 4.0 x4 interface, however, the nature of the secondary M.2 port is uncertain. The ProART B550-Creator’s M.2 ports utilize Asus’ new M.2 latch mounting system that facilitates SSD installation.
Image 1 of 3
ProART B550-Creator (Image credit: Asus)
Image 2 of 3
ProART B550-Creator (Image credit: Asus)
Image 3 of 3
ProART B550-Creator (Image credit: Asus)
The ProART B550-Creator provides three PCIe x16 and two PCIe x1 expansion slots. Since the motherboard’s product page isn’t available, the speed of the expansion slots are unknown.
Being a motherboard tailored to professionals and creative artists, the ProART B550-Creator naturally has the latest in connectivity. The motherboard flaunts two Thunderbolt 4 ports as well as dual 2.5 Gigabit Ethernet ports. The rear panel also shows a pair of USB 2.0 ports, four USB 3.0 ports, a combo PS/2 connector, a DisplayPort In connector and a standard HDMI port.
The ProART B550-Creator’s audio system is based on Realtek’s ALC1220A audio codec. The codec itself is isolated from the other components of the motherboard. The system also includes audio-grade capacitors and an integrated amplifier. The motherboard supplies five 3.5mm audio jacks and one SPDIF-Out connector for connecting your audio devices.
We use cookies on our website to give you the most relevant experience. By clicking “Accept”, you consent to the use of ALL the cookies.
This website uses cookies to improve your experience while you navigate through the website. Out of these, the cookies that are categorized as necessary are stored on your browser as they are essential for the working of basic functionalities of the website. We also use third-party cookies that help us analyze and understand how you use this website. These cookies will be stored in your browser only with your consent. You also have the option to opt-out of these cookies. But opting out of some of these cookies may affect your browsing experience.
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.
