HW News - Dell's Scummy Tactics, RTX 3070 Ti Rumor, NVIDIA Challenged on ARM Purchase

By Published April 24, 2021 at 3:29 pm

This week, we see VESA issue a rare press release squashing a run-away rumor claiming a new VESA DisplayHDR 2000 specification. We also have new developments regarding the potential Nvidia-Arm deal, as the UK is becoming increasingly more scrutinous of the deal, now citing national security concerns. AMD also finally got around to releasing its big Radeon Software update, after delaying it out of the usual end-of-year/holiday release window. There’s plenty more to go over, of course, like RTX 3070 Ti rumors.

At GN, we recently dove into the dangers of using cheap cables, and documented the various fire hazards and false marketing therein. We also looked at how Intel has become AMD, in a sense, and looked at Intel’s Phantom Canyon NUC11PHKi7C.

01:10 | Advisory: Dell Ripping-Off Buyers with Warranty

The first story of the week is more of a public service announcement and is based off of our own recent experience dealing with Dell and Alienware. We'd advise watching the news video for the full story, but we'll add a quick recap here: After AMD published a video detailing a new Alienware PC, we went to Dell’s website to buy one. Looking at the options available for the PC on Dell’s website, we nearly missed that buying an Alienware PC on their site auto-opted you into a $10 a month service subscription. Dell clarified to us that this being auto-included with the purchase was OK, because the first month was free, but we strongly disagree. This feature which, again, is a recurring monthly payment, is buried under all of the other options you make when buying a system and is included as default. We wanted to include this advisory ahead of that independent piece because we saw while buying a system how quickly other people were buying them. This also comes not long after Linus’ own issues with Dell’s service where they included the service despite LTT explicitly opting out of it. 

06:24 | The UK Has Stepped In on the Nvidia-Arm Deal

As expected, the Nvidia-Arm deal is facing an increasing amount of opposition and skepticism. The proposed Nvidia-Arm merger is already set to face a great deal of scrutiny in China, and the deal is currently under a developing investigation in the US, where the FTC is currently probing for more information regarding the acquisition. 

Now, the UK has made it clear it will intervene on behalf of national security concerns. The UK’s Competition and Markets Authority (CMA) already launched an investigation into the merger back in January, and it did so on grounds of antitrust and monopoly concerns. However, the UK’s Secretary of State for Digital, Culture, Media, and Sport has now called for the CMA to also look into national security implications the deal may have.

According to the Intervention Notice filed by the Secretary of State for Digital, Culture, Media, and Sport, the CMA will have until July 30, 2021, to report its findings regarding the national security aspects of the proposed Nvidia-Arm deal. 

Arm has long been based in Cambridge and is deeply integrated in the security of devices using Arm technology, and it stands to reason that the UK is worried about a foreign company acquiring it. Hermann Hauser, Arm co-founder, has fervently opposed the deal from the onset, calling on the UK to get involved and even launching a website (savearm.co.uk) to halt or at least alter the deal. 

In the meantime, Nvidia recently took the wraps off of its first datacenter CPU design, named Grace. Nvidia’s Grace is an Arm-based CPU that will use next-gen Neoverse (presumably from Arm’s new ArmV9 architecture) cores, as well as Nvidia’s NVLink, a proprietary CPU-to-GPU interconnect. All of this raises more questions as to how Nvidia can be a neutral, even-handed third-party home for Arm, as well as the hundreds of Arm licensees that Nvidia is competing with. 

It’ll likely be several months before we have a better idea of how this deal will play out. However, Nvidia is still confident it can push the deal through.         

Source: https://www.gov.uk/government/publications/proposed-acquisition-of-arm-limited-by-nvidia-corporation-public-interest-intervention


08:43 | RTX 3070 Ti Rumored

With the 3080 Ti now functionally confirmed to exist in a 12GB SKU, and with a freshly rumored May 26th release date (and 25th review date), we can all declare our boredom of this GPU and move on to the next rumor. The next one is the RTX 3070 Ti, alleged to exist by HKEPC. HKEPC’s Chinese-language coverage cited AIB partners in Taiwan (most of them) as the source for the information. 

HKEPC says that the RTX 3070 Ti will be available in early June. The media outlet said that the GPU will be GA104, which is currently used in the RTX 3070, the 3060 Ti, and some mobile devices. The 3070 uses GA104-300, the 3060 Ti uses GA104-200, and the 3070 Ti is rumored to use GA104-400. HKEPC says that this will run 6144 CUDA cores and 8GB of GDDR6X, with the G6X inclusion making it a higher value target for miners.

Source: https://translate.google.com/translate?sl=zh-CN&tl=en&u=https://www.hkepc.com/20177/

10:25 | VESA Confirms, There Is No “DisplayHDR 2000”

After several other outlets erroneously -- and embarrassingly -- ran with an unfounded rumor regarding a “DisplayHDR 2000” certification, VESA released an official statement on the matter. “There is no ‘DisplayHDR 2000’ tier in the VESA DisplayHDR specification and logo program at this time,” says VESA in its press release. 

“VESA has no knowledge of the origins of the DisplayHDR 2000 logo currently posted on these display listings on the Taobao website. However, VESA takes any misuse of our trademarks and logos seriously. VESA does not endorse the use of this logo unless and until a DisplayHDR 2000 tier has been officially announced by VESA, and any products claiming to meet this tier level have been officially certified by VESA and are listed on our website at https://displayhdr.org/certified-products/.”

VESA’s statements are in response to a pair of monitors -- Samsung Odyssey G9 and Acer EI491CRG9 -- that were listed over at Toabao. Both monitors were listed with an illegitimate logo claiming support for DisplayHDR 2000. As it stands, DisplayHDR 1400 is currently the latest VESA certification -- and VESA has only certified a handful of displays for that spec, as it stands currently.   

Source: https://displayhdr.org/pr/vesa-addresses-use-of-illegitimate-displayhdr-2000-logo/

12:01 | TSMC Q1’ 2021 Revenue 

TSMC recently released its quarterly revenue numbers for Q1’2021, and with them, an overview of how much revenue each process brings in. To start, TSMC disclosed that for Q1’2021, it raked in $12.92B in net revenue, which is a 1.9% increase QoQ and a 25% increase YoY. For the quarter, TSMC had a gross margin of 52.4% and an operating margin of 41.5%. 

All told, TSMC shipped a total of 3.36M wafers throughout the quarter. TSMC noted that its advanced technologies, which it defines as 7nm and below, accounted for 49% of its wafer revenue. Specifically, TSMC’s 5nm node accounted for 14% of the quarter’s wafer revenue, while 7nm accounted for 35%. TSMC’s 5nm revenue is actually down a bit QoQ, while 7nm is up. This isn’t a huge surprise, as TSMC doesn’t have nearly as many customers at 5nm as it does at 7nm -- not yet, anyway.

As for other process nodes that aren’t 7nm or 5nm, TSMC lists 20nm and 10nm as having 0% impact on Q1’2021 revenue. However, some of TSMC’s more mature processes are still seeing activity, as 28nm accounted for 11% of the quarter’s revenue. Furthermore, 40/45nm came in at 7%, 65nm at 5%, and 90nm at 3%, respectively.

Breaking revenue down by platform, TSMC notes that 45% of the company’s Q1’2021 revenue came from smartphones. Additionally, HPC accounted for 35%, IoT at 9%, automotive at 4%, and DCE and “Others” at 4% and 3%, respectively.      

Source: https://investor.tsmc.com/english/quarterly-results/2021/q1


14:30 | AMD Adrenalin 21.4.1 Update

AMD typically follows an annual cadence when it comes to updating its Radeon Software/Adrenalin suite, usually opting for an end of year or holiday rollout for big updates. Considering the anomaly that 2020 was, and how 2021 is already shaping up to be, it’s not surprising that we’re just now talking about this update in April. 

AMD has officially released Adrenalin 21.4.1, available immediately. The new update comes with a number of new features and enhancements, as well as the usual list of bug fixes and stability improvements. Most notably, AMD is trying to offer users more granular control over install options, as noted by “Full Installation, Minimal Installation, and Driver only” installation options. 

On top of overhauling the installer, AMD is now offering the ability to monitor Ryzen CPUs, whereas previously the Radeon software only monitored GPUs. According to reports and previews, AMD has tweaked the performance metrics tab to better accommodate both CPU and GPU monitoring, and there’s a new stress test feature as well. 

AMD has also improved its AMD Link, rolling out version 4.0 with this release. AMD Link has evolved quite a bit over the past few years, and now offers a full Windows 10 client/app. This means users who are using AMD Link no longer require a SmartTV or smartphone to use the feature. AMD also notes that AMD Link now offers support for 4K/144 FPS streaming. 

Elsewhere, AMD has also made changes to ReLive, rehoming all of the streaming and broadcast tools into one tab, and there’s also a new setup wizard (Streaming Wizard) available for first time configurations. AMD is also adding Microsoft PlayReady AV1 Decode support, and is extending EyeFinity support to its Ryzen 4000 and 5000 Mobile series of chips.    

Source: https://www.amd.com/en/support/graphics/amd-radeon-6000-series/amd-radeon-6900-series/amd-radeon-rx-6900-xt

16:52 | Discord-Microsoft Deal May Have Fallen Through

According to The Wall Street Journal, who cites sources familiar with the matter, Discord’s talks with Microsoft over a sale have reportedly been shelved for the time being. According to multiple reports, Discord had been in advanced talks with Microsoft regarding a deal that could’ve seen Microsoft acquire Discord for more than $10B. 

While the sales talks seem to be halted for now, WSJ notes that a deal isn’t entirely off the table. It seems for now, though, Discord is content with being independent, and is more interested in exploring an IPO in the future. According to WSJ, Discord’s user base is currently at 140M, and the company earned $130M in revenue in 2020. However, it seems Discord isn’t yet profitable, according to reports. 

Microsoft’s interest in Discord would’ve likely been to help bolster its presence in social media and gaming, as well as expanding its obsession with chat software. Microsoft has also been on a spending spree, with this year alone seeing the company spending $7.5B to acquire ZeniMax Media and more than $19B to buy Nuance Communications.  

Source: https://www.wsj.com/articles/discord-ends-deal-talks-with-microsoft-11618938806

18:10 | Intel Wins Second Patent Trial Against Patent Troll

Intel has been locked in a bitter patent trial with VLSI Technology, where VLSI is suing Intel over various accounts of alleged patent infringement. The lawsuit will spill out over the course of three trials, one of which Intel has already lost. While Intel plans to appeal the first verdict, it won’t have to appeal the second one, as the courts have ruled in Intel’s favor this time. 

A federal jury in Texas -- in the same court that ordered Intel to pay VLSI $2.18B in the first trial -- has decided that Intel has not infringed on at least two patents held by VLSI. These patents were previously owned by NXP Semiconductors, and were issued two decades ago. The patents deal with MP3 player technology, and VLSI was targeting Intel’s Speed Shift technology, claiming products using such technology infringes on the patents.

“Intel is pleased that the jury rejected VLSI’s meritless claims that Intel’s cutting-edge processors infringe expired patents on MP3 player technology. VLSI is a shell company created by Fortress, a Softbank-owned hedge fund, for the sole purpose of extracting billions from innovators like Intel,” said Intel to Tom’s Hardware. 

Intel had previously said that the patents weren't exactly cutting edge, even two decades ago when they were issued. Additionally, Intel noted that the $3B in damages that VLSI was seeking was more than 3,000 times what the patents were valued at in past acquisitions. A third trial over similar patents is set to take place in June.

Should you require more proof that VLSI and Fortress are patent trolls, consider that Apple and Intel forged a rare alliance to throw in against them.     

Source: https://www.bloomberg.com/news/articles/2021-04-21/intel-wins-billion-dollar-patent-trial-over-former-nxp-chips


20:31 | New Storage-Based Cryptocurrency

A new storage-based cryptocurrency by the name of Chia is on the horizon, after being in development since 2017. Chia will be a new coin that users “farm,” rather than mine, and users will do so with storage space. Whereas other coins, like Bitcoin, use a proof-of-work model, Chia will use a proof of space-time model. 

What this means is that users on the Chia blockchain will allocate unused hard drive space to be “seeded.” Seeded HDD space will then be used to store cryptographic numbers, known as plots, on the Chia blockchain. 

“Chia blockchain will ‘seed’ unused space on their hard-disk drive by installing software which stores a collection of cryptographic numbers on the disk into ‘plots.’ These users are called ‘farmers.’ When the blockchain broadcasts a challenge for the next block, farmers can scan their plots to see if they have the hash that is closest to the challenge. A farmer’s probability of winning a block is the percentage of the total space that a farmer has compared to the entire network,” according to Chia.

“Proof of time requires a small period of time to pass between blocks. Proof of time is implemented by a Verifiable Delay Function that takes a certain amount of time to compute, but is very fast to verify. The key idea of a VDF is that they require sequential computation, and since having many parallel machines does not yield any benefit, electricity waste is minimized,” Chia explains. 

Overall, Chia hopes to create a better cryptocurrency that doesn’t require specialized hardware (GPUs or ASICs) and reduces energy waste (i.e., electricity). That said, in anticipation of Chia’s arrival on currency exchanges, it seems many users are panic-buying storage ahead of the coin becoming publicly tradable or farmable.

According to a report from HKEPC (via Tom’s Hardware), prospective Chia farmers are already snatching up HDDs ranging in capacity from 4TB to 18TB. As a result, HDD and SSD prices are expected to increase by as much as HKD$200 to HKD$600 (~$26 to $77) in Hong Kong. Furthermore, Jia1he2 Jin4wei1, one of China’s biggest domestic memory manufacturers, has already sold through all of its Gloway and Asgard-branded 1TB and 2TB NVMe M.2 SSDs. This is despite the fact that Chia is ostensibly better suited for enterprise-grade HDDs due to its heavy read/write operations.  

Time will tell what impact Chia will have on the cryptocurrency market, and what kind of knock-on effects it will have on consumer hardware.  

Source: https://www.techspot.com/news/89350-emerging-storage-based-cryptocurrency-chia-could-jack-up.html


23:49 | Uni. of Minnesota Banned From Linux Kernel   

Phoronix is reporting that Greg Kroah-Hartman, the Linux kernel maintainer for the stable branch, has banned the University of Minnesota from Linux development contributions. The ban comes after two graduate students at the university intentionally submitted vulnerable and faulty code to the mainline Linux kernel. 

This was allegedly done in the name of research, using Use-After-Free (UAF) bugs to purportedly show how unsavory individuals may introduce vulnerabilities into open-source projects. Well, you’d be hard pressed to find a bigger open-source project than Linux. As you might imagine, this experiment didn’t go over well.  

The graduate students documented their “research” in a paper titled "On the Feasibility of Stealthily Introducing Vulnerabilities in Open-Source Software via Hypocrite Commits."  

The students apparently tried injecting multiple patches into the mainline Linux kernel, resulting in not only them, but the entire University of Minnesota being banned from future Linux development. According to Kroah-Hartman: “I will now have to ban all future contributions from your University and rip out your previous contributions, as they were obviously submitted in bad-faith with the intent to cause problems."

The Linux development community seemed to overwhelmingly agree with the ban, with many prominent Linux developers and programmers chiming in on Kroah-Hartman’s comments. Jered Floyd, who serves as Red Hat Technology Strategist, may have said it best regarding this so-called research. 

“This is worse than just being experimented upon; this is like saying you’re a ‘safety researcher’ by going to a grocery store and cutting the brake lines on all the cars to see how many people crash when they leave,” said Floyd in a tweet.   

The University of Minnesota has since responded to the ban, and has stated it is looking into the matter and has suspended that line of research. Additionally, the school will “investigate the research method and the process by which this research method was approved, determine appropriate remedial action, and safeguard against future issues, if needed." 

Whether or not any of that is enough to reverse the ban remains to be seen. 

Source:  https://www.phoronix.com/scan.php?page=news_item&px=University-Ban-From-Linux-Dev&s=09

Update: https://www.tomshardware.com/news/university-of-minnesota-responds-linux-ban

Editorial: Eric Hamilton
Host, Additional Writing: Steve Burke
Video: Keegan Gallick, Andrew Coleman

Last modified on April 24, 2021 at 3:29 pm

We moderate comments on a ~24~48 hour cycle. There will be some delay after submitting a comment.

  VigLink badge