Month: December 2020

The iCue H115i Elite Capellix is the latest CPU chiller from Corsair and boy is it colourful. You’d better like lights if you’re planning on picking up this all-in-one cooler. This thing has a lot of lights. OK, the RGB functionality is far from the only thing Corsair’s latest all-in-one water cooling solution has going for it. But it is very much central to the proposition.

The model currently roasting our retinas is the H115i and thus a 280mm member of a model line also available in 240mm and 360mm form factors. The aforementioned RGB-ness concerns the fans and the top cover of the pump head. The latter comes with an alternative transparent cover if you want to maximize the RGB goodness in that area.

The clever bit, as implied by the wordy moniker, is the use of Corsair’s Capellix LEDs. Corsair reckons they’re not only 90 per cent smaller than conventional LEDs. They’re also more efficient, brighter and run cooler. Add that all up and you have a lighting solution that gives you more mounting options, increased visual impact and fewer downsides and concerns around power use and heat.

The net result is that Corsair has squeezed no fewer than 33 LEDs into the pump head, plus eight per fan. The lighting effects are certainly a clear cut above the RGB normal, with slick transitions and animations and generally a more complex and detailed overall effect. However, the pump head itself looks a little old fashioned, especially compared to the most advanced alternatives such as the NZXT Kraken Z73 with its funky pump-top screen, albeit that’s a much more expensive product.

Anywho, that RGB stuff is controlled by the iCue Commander Core lighting and fan-speed controller box. Along with the pump head, it supports up to six RGB-enabled fans and hooks up neatly to the motherboard via a USB header. Less visually edifying—or physically reassuring, for that matter, is the cable connecting the pump to the motherboard header to initiate operation. It feels fragile and looks cheap.

The other half of the equation is obviously the cooling on offer. Here Corsair has equipped the H115i with a low-noise centrifugal pump with a flow-rate spec of 0.82L/min and a claimed 20dB of noise. The radiator, meanwhile, clocks in at 322x137x27 millimetres and as standard comes with a pair of 140mm mag-lev fans running at speeds from 400 to 2,000rpm.

The full package includes mounting for both Intel and AMD CPU sockets covering all the latest kit. The pump head comes primed with a layer of thermal paste, but there’s no extra tube. So, either bring your own paste or don’t muck the installation up.

Speaking of mucking things up, the instruction booklet is a little light on details and the package labelling, which includes a bag unhelpfully labelled ‘miscellaneous hardware’, could be clearer. But it’s not truly an impediment to assembly. More of an issue, in our kit at least, was the tightness of the screws for attaching the fans to the radiator. They were stiff enough to risk rounding off the heads, which isn’t ideal.

Once up and running, the H115i runs super quiet at idle. In fact, you’ll struggle to detect it with normal hearing. Better still, in even the most extreme performance mode noise levels are extremely well controlled. To give the H115i the most demanding, we tried it out with the flaming inferno that is Intel’s 14nm Core i9-10900K 10-core CPU. The result was tangibly lower temps than several other AIO coolers we’ve test-driven with the 10900K, including the NZXT Kraken Z63. If it’s maximum performance and minimum racket you seek, therefore, you could do an awful lot worse.

At least you could if reliably controlling the Corsair iCue H115i Elite Capellix was a given. In our experience and with the current software and firmware, that’s not quite the case. Initial installation and setup with Corsair’s iCue software went smoothly and the interface offers more RGB configuration options than you could humanly desire.

The cooling configs include quiet, balanced, extreme or zero RPM modes, while the cooling and RGB functionality come together neatly via emergency triggers that can be set to not only ramp-up fan speed but also indicate an issue with an LED colour change when temps blow past a user-defined limit. Nice.

Less edifying is the fact that we had numerous issues with software and driver. After running fully stable at first, a system reboot had the iCue software initially refusing to recognise the cooler was present and then indicating it was present but with a warning symbol and essentially zero functionality. A quick perusal via the world’s de facto search tool suggests ours was not an isolated experience. It seems issues with Corsairs iCue platform are not uncommon. Which is a pity, because we otherwise reckon there’s plenty to appreciate. Assuming, that is, you really do like lights.

AMD’s Threadripper Pro processors for workstation desktops are finally here. These chips disrupt a market that Intel has dominated since AMD discontinued its Opteron lineup in 2017. In fact, they have the potential to upend it by offering up to an incredible 64 processing cores on a single CPU die. That’s far more in core count than an Intel Xeon W-class chip, and still more than a match for two Xeon scalable chips, while obviating the complexity and cost of multiple-CPU implementions.

Read on to find out if a Ryzen Threadripper Pro is worth considering in your next workstation.

Threadripper Pro: A Lenovo Exclusive (for Now)

Unlike the non-“Pro” consumer Ryzen Threadripper, AMD has no plans to offer Ryzen Threadripper Pro chips as standalone purchases. For its debut, the Pro chip is available exclusively in the Lenovo ThinkStation P620, a workstation tower that we are reviewing in parallel with this story. AMD says it will make the chip available to other workstation vendors after the exclusivity arrangement ends, though it did not provide a timeline for that.

Now, for the technical details. Let’s start with the basic specifications of the four Ryzen Threadripper Pro chips that were announced this past summer…

Per-chip pricing isn’t available, as you can see. These chips can only be purchased as part of a prebuilt system. A little online shopping, however, shows that the ThinkStation P620’s pricing is in line with a comparable Intel-based workstation when it comes to cost per CPU core. (More on that in our ThinkStation P620 review.)

Two of the Ryzen Threadripper Pro chips, the Ryzen Threadripper Pro 3975WX and the 3995WX, have the same core and thread counts as their non-Pro counterparts, the Ryzen Threadripper 3970X and the Ryzen Threadripper 3990X. The other two chips, the Ryzen Threadripper Pro 3945WX and 3955WX, do not have current-generation Ryzen Threadripper counterparts, as their 12-core/24-thread and 16-core/32-thread footprints are filled by the Ryzen 9 3900X and Ryzen 9 5900X (at the 12-core level), and the Ryzen 9 3950X and Ryzen 9 5950X (at the 16-core). The Ryzen Threadripper Pro 3945WX is notable for its high 4GHz base clock.

WRX80: A New Chipset for Threadripper

The Ryzen Threadripper Pro shares many specifications with the non-Pro version, including a 280-watt thermal design power (TDP) rating and AMD’s Zen 2 architecture. (See our Ryzen 9 3900X review for a primer on Zen 2.) However, the Pro chips have slightly lower clock speeds, likely for stability or thermal reasons. For example, the Ryzen Threadripper 3970X has a base clock of 3.7GHz and boost clocks of 4.5GHz, while the core-equivalent Ryzen Threadripper Pro 3975WX lowers those to 3.5GHz and 4.2GHz, respectively.

The Ryzen Threadripper Pro also uses the new WRX80 chipset, which brings on a bunch of improvements over the TRX40 chipset. (TRX40 debuted to support third-generation Threadrippers.) The WRX80 supports up to 128 PCI Express 4.0 lanes, up from 88 for the Ryzen Threadripper and its TRX40 chipset. It also doubles the memory channels from four to eight and works exclusively with error correcting code (ECC) memory.

The interior of the ThinkStation P620 and its WRX80 motherboard.

The ECC support is a crucial distinction in the workstation world. ECC is essential technology for fields, such as science, architectural design, and finance, in which unimpeachable accuracy and every decimal place matters. Non-Pro Ryzen Threadripper chips, in contrast, support ECC memory only in conjunction with certain motherboards.

For business, the Ryzen Threadripper Pro offers a Trusted Platform Module (TPM) for hardware encryption, along with AMD’s Ryzen Pro line of security and management technologies, its answer to Intel’s vPro.

One last curiosity about the Ryzen Threadripper Pro is that it does not require Windows 10 Pro for Workstations, like some Intel Xeon chips do. Thanks to a licensing agreement between AMD and Microsoft, Ryzen Threadripper Pro systems can use straight Windows 10 Pro, which will help keep costs down. The arrangement makes sense, since Windows 10 Pro for Workstations supports technologies that Ryzen Threadripper Pro workstations will not use, such as multi-CPU implementations.

AMD Ryzen Threadripper Pro vs. Intel Xeon

Intel’s competition for Ryzen Threadripper Pro chips with 16 or fewer cores comes from its Xeon W-class line. Here are the closest comparisons Intel’s Xeon offers for the 12-core and 16-core Threadripper Pros.

The Intel chips are largely outclassed on paper, only outdoing AMD’s offerings with their higher boost clocks and lower TDP ratings. However, given AMD and Intel chips are built on different architectures, little can be assumed from clock speed and other differences, such as cache amounts. In addition, the higher maximum memory ceilings of the AMD chips are largely on-paper advantages, since most workstations with the chips above would not be ordered with more than 1TB of memory. (That said, some Xeon scalable chips can support 1.5TB or more.)

What can’t be ignored is the Ryzen Threadripper Pro’s massive advantage in PCI Express lane support. It also supports PCI Express 4.0, a potentially big deal in workflows where maximum storage throughput is beneficial. (See our Samsung SSD 980 Pro review to see how fast a PCIe 4.0 SSD can get.)

Cooling on the Threadripper Pro 3955WX is implemented via a twin-stack air cooler with two fans. No single Xeon W-class chip matches the 32-core/64-thread Ryzen Threadripper Pro 3975WX in core and thread count, so that leaves the job up to Intel’s scalable Xeon line, specifically the Xeon Gold and Platinum chips with 16 or more cores. Two of them can be used in a dual-CPU workstation to achieve equivalent core and thread counts, though as of this writing, I was unable to find a dual-CPU Intel tower workstation that supported more than a total of 56 cores and 112 threads (two 28-core/56-thread chips), leaving the monster 64-core/128-thread Ryzen Threadripper Pro 3995WX in a class of its own.

Testing the Ryzen Threadripper Pro 3955WX

Our testing focuses on the one Ryzen Threadripper Pro processor we’ve managed to get our mitts on so far, the 3955WX, a 16-core/32-thread chip that is a sensible choice for any kind of demanding workflow. Our test platform is the Lenovo ThinkStation P620 with Windows 10 Pro, 32GB of DDR4-3200 ECC memory, a 512GB PCI Express 3.0 SSD boot drive, and a 5GB Nvidia Quadro P2200 graphics card.

The memory is two 16GB ECC modules. That’s a somewhat unfortunate arrangement, since it makes use of only two out of the Ryzen Threadripper Pro’s possible eight memory channels. That said, none of our testing is particularly memory-bandwidth-sensitive, so the impact on our testing should be minimal.

Use the drop-down or the arrows at the top of the chart, below, to switch among various benchmarks. We ran our usual CPU testing suite, which comprises a host of core-crunching, media-processing, and single-threaded tasks. We’ve included, for context, recent Ryzen 7, Ryzen 9, Core i7, and Core i9 mainstream CPUs, plus the latest Threadripper, Xeon, and Core X-Series high-end chips we’ve tested.

For the Threadripper Pro 3955WX, pay special attention, across the charts, to the comparison numbers served up by these four chips:

The Intel Xeon W-2295: Intel’s closest Xeon W-series competition, with 18 cores.

The AMD Ryzen Threadripper 3960X: AMD’s 24-core non-Pro Threadripper CPU of the current generation.

The AMD Ryzen 9 3950X: The top Zen 2 chip of the mainstream Ryzens from the previous generation (the Ryzen 3000 series), with 16 cores. (We are working on testing the newer Ryzen 9 5950X.)

The Intel Core i9-10980XE Extreme Edition: The top (18-core) CPU in Intel’s Core X-Series high-end desktop line, around $1,000, mainly still relevant due to its support for more PCI Express lanes than Intel’s mainstream Core CPUs.

For starters, the Ryzen Threadripper Pro 3955WX is neck-and-neck with the Ryzen 9 3950X in almost every test, a small surprise given considering that they share the same core and thread count and have similar clock speeds. Likewise, the Threadripper 3960X consistently dominates, no surprise at all given its substantial core/thread advantage.

Next, let’s look at the Intel Xeon silicon. The Ryzen Threadripper Pro 3955WX has a slight disadvantage next to the Xeon W-2295 in decidedly single-threaded scenarios, such as the geriatric iTunes 10.6 and in the single-core run of POV-Ray 3.7. However, the bigger news is that the Ryzen Threadripper Pro 3955WX matches or beats it just about everywhere else, especially in the fully threaded Cinebench tests, an impressive stat considering the AMD chip has two fewer cores. This indicates that AMD’s Zen 2 architecture offers better per-core performance than the Intel chip’s aging 14nm “Cascade Lake” process.

Meanwhile, the Core i9-10980XE, also in the Cascade Lake generation, trades blows with the Threadripper Pro 3955X but generally comes in a bit behind it on the multi-core-heavy trials. It does fare a bit better than the Xeon in most cases.

Watch Out, Intel! CPU Competition on Workstations, Too, Is Heating Up

Though the Ryzen Threadripper Pro’s Zen 2 architecture isn’t AMD’s latest and greatest—Zen 3 came out just in the final quarter of 2020—it still compares more than favorably with Intel’s “Cascade Lake” Xeon chips. Our results show that it punches above its weight, offering more per-core performance than Intel’s chips, and, as you can see in our Lenovo ThinkStation P620 review, at a comparable price-per-core.

The excitement is somewhat dampened by Lenovo’s exclusive hold on the Ryzen Threadripper Pro, though at least it’s a temporary arrangement. Before we see Threadripper Pro, if ever, in DIY or consumer hands, we’d have to see at least a small ecosystem of WRX80-based motherboards emerge. Before and beyond that, though, we hope the chip is swiftly adopted by rival workstation makers, such as Dell and HP, to bring some competition to a market that very much needs it.

Windows 10 includes a custom volume mixer, which sounds like a great idea on paper—who wouldn’t a way to control the volume of individual apps? It’s incredibly useful when you need to dial back your game audio a bit so you can hear your guildmates on Discord (or the YouTube video you’re listening to in the background). It’s not so useful when you forget it exists.

For the past week or two, some YouTube videos I’ve watched have sounded a little faint. I didn’t think much of it at first, assuming I simply bumped the dial on the external DAC I use with my headphones. I cranked that up a little bit, which solved the YouTube problem, but I then had to dial back Windows 10’s overall volume because my games and system sounds were coming in a bit too hot.

I didn’t really think about the relationship between my apps and volume. Instead, I kept moving my system’s volume up and down as needed when switching between apps. I figured some of the YouTube videos were just encoded at a lower volume. It happens. The difference between those videos’ volumes, my system volume, and the volumes of my other apps wasn’t jarring enough to make me think something was wrong.

Something was wrong.

After around two weeks of this, it finally clicked. Maybe there was something weird going on with my audio levels. And right before I was about to bury myself in Windows 10’s sound settings, I remembered that damn Windows 10 volume mixer.

You can see it in action yourself by right-clicking on the volume icon in the lower-right corner of your taskbar, and then left-clicking on “Open volume mixer.” In a perfect world, all of the sliders would be at the exact same level, like so:

If you’re seeing some sliders set much differently, or finding that some apps are muted even though you’re still hearing sounds in others (or hearing general Windows system sounds) you’ve found your problem.

To fix this, you’ll have to engage in a fun dance of moving all the sliders either to the top or the bottom of their ranges so they all match. Then, move your Device volume (the left-most slider) to a comfortable position, and it should move all of the other sliders at the same rate. Getting everything to equal out can be a little annoying, but that’s how you fix it.

I haven’t tried it, but there’s another technique to consider: I believe you can also pull up Windows 10’s Sound mixer options via a Start Menu search. In the screen that appears, scroll down a bit, and then click on the “Reset” button in the lower-left corner. This should turn all of your apps up to 100%—that’s one-hundred percent of the system volume, since it’s a relative measurement—without messing with your actual system volume and blowing out your eardrums in the process.

Thanks to Apple’s new M1 processor, the latest MacBook Air, MacBook Pro, and Mac Mini machines are breaking benchmarks left, right, and center. That’s enough for many of the geekrati to level up to the ARM-based hardware, but Apple has something spectacular planned for 2021.

Building on his previous reports, noted Apple analyst Ming Chi-Kuo lays out a stronger case for Apple switching to mini-LED displays in the MacBook hardware next year. With Apple’s suppliers gearing up for a massive switch in production, it’s clear that the LED panels offering improved colors, higher contract, and better brightness, are on the way. Tim Hardwick reports:

“According to the report, Apple is set to launch its first mini-LED iPad Pro in the first quarter of 2021 and mass produce mini-LED MacBook Pro models in the second quarter, and a number of Taiwan manufacturers are building a “pre-emptive presence” in the Mini-LED segment in anticipation.”

consideration of the silicon. The M1 chip is posting some great numbers of benchmarking sites, and the raw processing power is already matching top tier chips from Intel. While there is more to a computer than the processor, it makes a pretty big difference.

While many are pointing out that Intel still has the superior chip, that misses two pieces of positioning.

The first is that you have to look at the chips being compared. The entry level MacBook Air, with the lowest specced M1, is outperforming everything bar Intel’s top tier of chips. Put the M1 chip against one of Intel’s entry level chips, say in the Core i3 range, and you have a more honest comparison.

Yet the M1 chip is the first generation chip. Literally any MacBook released after this point is going to have more performance in it. When Apple rolls out an M2, or perhaps even the presumptively named M1X for faster desk-bound machines, I would expect Intel to lose that the performance lead.

Then there is Apple’s hardware and software. For all of the slick presentation and joyful initial reviews the M1 and the associated Mac machines are still first generation hardware; coupled with a signification update to macOS to allow it to run on Apple’s ARM-based silicon. While some apps work wonderfully, others are causing problems, running slower than on the Intel Macs, and significantly impacting on battery life.

As with any computing platforms, many of these problems will be addressed with software updates in the majority of cases, although the possibility remains that issues could be found in the hardware. Much as the lore of Apple is rock solid, in practice this is first generation hardware and software. The cautious should wait.

And finally, as typified by this report on the upcoming use of mini-LED screens in next year’s MacBook machines, the current MacBook Air and MacBook Pro machines are old designs. While they have the M1 chip inside, and the necessary updates circuitry to accomodate the changes, apple declined to make any external changes.

That certainly makes the sell to the average consumer a little easier (“look,” the theory goes, “it’s the same Mac you know and love, but its faster and has a longer lasting battery”), yet the brave new world of ARM surely demands a brave new design with the latest technology? That’s coming next year.

A faster processor that will out-gun most if not all of the x86 options; iterations to improve the first generation hardware and software; and the latest technology. All of these are expected in next year’s new MacBooks. 

Tech enthusiasts gleaned some pre-launch insight into the AMD Radeon RX 6800/XT graphics cards ahead of launch thanks to developer Patrick Schur sharing an AIB card BIOS with veteran tech reviewer Igor Wallossek. Now it looks like history is repeating, with the dynamic duo teaming up to produce insightful analysis of an ‘AMD Radeon RX 6700 XT’ before it becomes official.

The Igor’s Lab blog, spotted via VideoCardz, says that Schur sent along two AMD Navi22 AIB BIOS files for analysis. Before we go on to that it is worth stopping and considering the questions – qhat is an AMD Radeon RX 6700 XT? According to this leaked information it is, nice and neatly, about half-an-RX 6900 XT. Compared to that known quantity flagship it has half the Shaders/TMUs/ROPs/RT Units at 2560/192/64/40. Other spec downgrades are the 3MB L2 cache (down from 4MB), and that the RX 3700 XT has 12GB of GDDR6 on a 192-bit bus.

Checking over the BIOS with his More PowerTool, Igor reckons that “enormous clock rates are expected” from the RX 6700 XT partner cards. The top GFX clock rates supported by the partner BIOS is 2,854MHz – that is higher than the RX 6800 XT limit of 2,800MHz, but less than the RX 6900 XT limit of 3,000MHz. It is much faster than its predecessor, the RX 5700 XT / Navi 21, could muster.

The two BIOS files under scrutiny had been set to different TGPs; one a moderate 186W, the other pushed it to 211W – expected to be a BIOS from a heavily overclocked sample. Another interesting insight derived from the BIOS settings is that the lower TGP model might well be a single fan design (looking at the very high cooling fan RPM). The OC model, meanwhile, has a 15 degree higher temperature max fan target of 95°C.

The rumour mill points to the AMD Radeon RX 6700 XT arriving in Q1 next year, and it is possible a non-XT version will arrive at the same time. Reports suggest that Nvidia is imminently launching the GeForce RTX 3060 Ti so it will be addressing the biggest GPU market first, if stocks are available.

Sometimes for a company to get its projects over the line, it requires the help of crowdsourcing ideas. One such company is Eve Tech, or EVE for short, and looking to build on the success of its first-generation EVE V 2-in-1 dubbed the “Surface Pro Killer”. Today the company has officially lifted the lid on its second-generation EVE V. The new EVE V 2-in-1 includes a whole host of impressive features and specs including Intel’s Tiger Lake mobile processors with Xe integrated graphics, Thunderbolt 4, Wi-Fi 6, and a 13.4″ 4K (3840 x 2400) with 10-point multitouch. The design of the device has been through deep interactions with interested end-user customers to build a device with the specifications that they want.

The EVE V 2nd generation has crammed all of its technology into a very high specification 2-in-1 with consultation of potential customers. For a convertible designed to compete with the latest and greatest, Eve Tech has lifted the lid on the EVE V 2nd Gen specifications, the technologies under the hood, as well as an impressive array of features.

Focusing on the screen, the EVE V is using a 13.4-inch Indium Gallium Zinc Oxide (IGZO) panel with a native 3840 x 2400 pixel (4K 16:10) resolution. This has a 16:10 aspect ratio, with two different variations including the higher-spec 4K resolution, or an FHD 1920 x 1080 resolution at 16:9. Some of the finer specifications of the screen include a 1500:1 contrast ratio, up to 500 nits of brightness, and is using edge-to-edge Corning Gorilla Victus glass, announced recently as one of the most scratch resistant glass surfaces.

The Tiger Lake processor comes with four cores and eight threads, as well as Xe graphics and Thunderbolt 4 support. For a convertable, the EVE V uses a lot of that IO, with three Type-C ports, two of them Thunderbolt 4. The two Thunderbolt 4 Type-C ports also include DisplayPort 1.4 and HDMI 1.4b video support, with a power output of up to 5V or 15 watts as well as charging capabilities. Other connectivity includes a USB 3.2 G2 Type-C port, with a single 3.5 mm audio jack, a microSD reader with compatibility for SDXC up to a maximum of 2 TB. The sound capabilities are also impressive with dual 0.7-watt front-facing speakers, and a pair of 1 W front-facing low-frequency speakers that resemble subwoofer qualities. Some variations of the EVE V will come with a Nano Sim slot for 4G capabilities.

Included with the EVE V is its V pen, which has been upgraded over the EVE V 2016 model. It is based on Wacom’s AES 2.0 technology and features 4096 pressure levels which is designed to benefit from the 10-point touchscreen. The new V keyboard supplied includes a folio cover with a soft touch feel, with a Windows Precision touchpad that includes a textured glass surface for optimal precision. It also includes an adjustable backlight for low light situations. There is a 5 Megapixel rear-facing camera, as well as a 720p front-facing camera for conference calls, as well as a fingerprint reader integrated into the power button.

Although Eve Tech hasn’t fully announced the availability of the V, it estimates that it’ll start shipping by the end of Q3 2021, and will be available globally across various retailers. Users can reserve the EVE V with a fully refundable down payment of $300, with the remaining expected to be paid prior to shipping.

The EVE V will start at $1399 for the base model with a Core i5, 16 GB of memory and a 512 GB NVMe SSD, with the top-spec model with the Core i7, 32 GB of RAM, and a 2 TB NVMe SSD for around $2000. This is inline with some of the premium Tiger Lake notebooks we have seen being launched recently with 3840×2400 screens, such as the Dell XPS13 and the Razer Book 13.