Hardware stub

HW News - Intel 7nm Delays Through 2022+, NVIDIA Ampere Discussion & ARM Interest

Posted on July 26, 2020

This week’s news is anchored by yet more manufacturing woes from Intel, this time at 7nm. Intel has had to delay its 7nm products due to less than great yields, although the exact problem causing poor yields wasn’t disclosed. Joining Intel in manufacturing trouble is Samsung, as the company is reportedly struggling to improve its 5nm yields. 

We also have leaks that NVIDIA is internally working on Ampere launch timelines now, which would point toward early rumors that the cards would see an August/September launch window. Igor of Igor’s Lab may be on the mark with that one. Finally, we have some smaller AMD and Cooler Master news to go over, as well as ARM interest by Nvidia. 

News article and video embed are below, as usual.

01:17 | Intel’s 7nm is Delayed until at Least 2022

As if things couldn’t get any worse for Intel on the manufacturing side, it just announced that its 7nm process has been delayed. That means volume production won’t come online until 2022, or possibly 2023, as opposed to late 2021. 

Intel CEO Bob Swan and man who hates benchmarking delivered the disheartening news along with Intel’s Q2 earnings, while also touching on a few other technology points. According to Swan, Intel is “seeing an approximate six-month shift in our 7nm-based CPU product timing relative to prior expectations.” Furthermore, Intel notes that 7nm yields are approximately 12 months behind internal target goals. 

“We have identified a defect mode in our 7nm process that resulted in yield degradation. We’ve root-caused the issue and believe there are no fundamental roadblocks, but we have also invested in contingency plans to hedge against further schedule uncertainty,” says Swan. In order to maintain Intel’s roadmap and competitiveness, Swan says Intel will lean on other aspects such as die disaggregation and advanced packaging.

Part of Intel’s contingency plans will also see Intel leverage both internal and external process technologies, which is to say, Intel will be leaning on third party foundries where it makes sense. As such, Intel is now stating that Ponte Vecchio will combine both internal and external process technologies, combined with Intel’s packaging technology, and will ship late 2021 or early 2022.

Intel’s historic trouble at 10nm also led Intel to decouple its architecture advancements from process node advancements, thus, Sunny Cove was born. This allows Intel to move forward with architectural improvements without hanging them on smaller processes in the event of a roadblock. This is a disaggregated approach that makes Intel’s architectures portable between nodes, and means Intel could backport a newer architecture to an older process technology. 

Intel claims it’s still committed to its roadmap and annual product improvement cadence, and also states that “while process technology is very important, it is only one of the six technology pillars of innovation that drive differentiation in our products.” However, there is no way to cut this where it doesn’t sting; especially when AMD is executing so aggressively on its own roadmap and products. Intel is promising more updates at its upcoming Architecture Day, so stay tuned.

Source: https://newsroom.intel.com/wp-content/uploads/sites/11/2020/07/ceo-remarks-earnings-call-2Q2020.pdf

06:17 | NVIDIA Working on Ampere Scheduling

We have confirmed with board partners that NVIDIA is working on its Ampere launch scheduling and timelines starting the week of the 26th, lining-up with earlier rumors that the cards would likely launch in late August or early September. To be clear, this is NVIDIA working internally and with partners on launch timing, not necessarily anything that’ll go public. It may preempt more leaks, though. This whole Ampere series is the earliest that an NVIDIA product has leaked in recent history, and in speaking with sources close to the matter, our understanding is that NVIDIA is unpleased with the leaky ship it’s running. Jensen Huang, CEO of NVIDIA, personally tries to control leaks to such an extent that he’s often debating what will be shown at a press conference in the final moments before getting on stage, speaking from experience. This is unusual for NVIDIA, and it doesn’t seem to be one of the intentional leaks that the cynically minded of us would point toward. As for AMD and RDNA2, we don’t know quite as much at this time.

Source: GamersNexus

08:31 |  AMD Confirms (Again) that Zen 3 is Coming This Year

Speaking of AMD, the company is once again reaffirming that Zen 3 chips are coming this year. In a perfect world, we likely would’ve seen them by now, as AMD usually reserves processor announcements for Computex. But alas, we’re living in a world with human malware that has sidelined every trade show for the year.

So, for the last several months, AMD has been confirming (and re-confirming) that we will see Zen 3 this year. In some form, at least. To drive this point home once again, AMD’s Rick Bergman, Executive Vice President of Computing & Graphics, took to AMD’s blog.

The blog post mostly centered on AMD’s newest 7nm Ryzen 4000-series APUs. However, towards the end of the blog post, Bergman concluded with “So, what’s next for AMD in the PC space? Well, I cannot share too much, but I can say our high-performance journey continues with our first ‘Zen 3’ Client processor on-track to launch later this year. I will wrap by saying you haven’t seen the best of us yet…”

Bergman does offer the word “client” here, meaning we’ll get more Zen 3 than just Epyc Milan, which has been on the roadmap for this year for some time. Though, the client side of Zen 3 still encompasses both desktop and mobile, so we’re still left to speculate a bit. 

Source: https://community.amd.com/community/amd-corporate/blog/2020/07/21/consumer-and-commercial-pcs-two-sides-of-a-valuable-coin

10:25 | Samsung’s 5nm Node Suffering from Poor Yields

Intel isn’t the only one struggling with shrinking process nodes this week, it seems. A report out of Digitimes suggests that Samsung is currently high-centered on its own nanometer hurdle, this one sized at 5nm. 

Samsung’s 5nm process will rely on EUV for critical layers, and the volume production was slated for end of Q2 this year. However, according to SeekingAlpha, Samsung wasn’t able to get EUV equipment from ASML installed until the end of June, which already puts Samsung’s time table in jeopardy.

Further complicating that is the reportedly poor yields Samsung is seeing, of which Digitimes claims Samsung is “struggling to improve.” As such, it’s being reported that this not only affects Samsung’s own chip and product line production, but possibly Qualcomm’s as well. Samsung’s first 5nm chip was allegedly meant to be its own Exynos 992 SoC; however, upcoming Samsung products are rumored to be using the older Exynos 990 instead. 

As for Qualcomm, its upcoming Snapdragon 875G SoC and Snapdragon X60 5G modem could be delayed, as they may very well be caught in the crosshairs of Samsung’s manufacturing woes.   

Source: Digitimes (paywalled) -- https://www.digitimes.com/news/a20200720PD203.html

https://seekingalpha.com/news/3592705-samsungs-5nm-struggles-delay-qualcomm-5g-chips-report

11:40 | Nvidia Could Be Looking to Buy Arm

In a chip deal to end all chip deals, it seems the possible sale of Arm has courted attention from Nvidia. It likely goes without saying, but an Arm acquisition from Nvidia would have huge implications for the CPU market and would undoubtedly attract the gaze of regulators. 

Arm is currently owned by SoftBank, a holding company conglomerate that has stakes in a slew of technology companies -- and formerly owned a chunk of Nvidia. SoftBank acquired Arm in 2016 for $32B, but recent reports are suggesting SoftBank may be ready to list the company for sale. Two separate reports have come from Bloomberg on the sale: one mentioning Nvidia as a potential buyer, and another stating that Apple isn’t interested

Apple buying Arm would be a huge conflict of interest, and likely wouldn’t sit well with regulators. Being that Nvidia doesn’t have a horse in the mobile SoC race, it may stand a better chance of getting a deal approved, but that’s a big “maybe.” Reports are also suggesting that SoftBank may pursue an IPO for Arm. It’s also possible that SoftBank is simply gauging interest and may end up doing nothing with Arm.    

Source: https://www.bloomberg.com/news/articles/2020-07-22/softbank-s-chip-company-arm-is-said-to-attract-nvidia-interest

https://arstechnica.com/gadgets/2020/07/reports-arm-is-for-sale-and-nvidias-interested-apple-isnt/

13:15 | AMD’s Stock is Riding a 15-Year High

AMD’s stock is currently enjoying a 15-year high. As Tom’s Hardware points out, for the first time since 2006, AMD’s stock price has topped Intel’s. 

AMD’s stock price has been on the rise for months now, starting 2020 by breaking its own stock price record, and now, reaching a 30% improvement YoY according to MarketWatch. Intel’s stock has been more tumultuous, however. In fact, Intel’s stock price recently took a nosedive in after hours trading, following the news that its 7nm portfolio has been delayed. MarketWatch notes that Intel has only seen a 1% improvement YoY on its stock. 

It seems that part of AMD’s surge in stock is coming at the expense of Intel’s setbacks, even as AMD has certainly been executing exceptionally well on its roadmap and product release cadence. AMD’s stock is currency sitting at $68.48, while Intel’s has tumbled down to $50.67.         

Source:  https://www.tomshardware.com/news/amd-stock-price-surpasses-intels-for-first-time-in-15-years

15:15 | Cooler Master Pi Case 40

Cooler Master is currently crowdfunding a case for the Raspberry Pi 4 on Kickstarter, in what seems like a new venture for the company that could possibly spawn similar products. 

The case, known as the Pi Case 40, is attempting to differentiate itself from other Pi cases in a couple different ways. For starters, Cooler Master claims the case will passively cool a Raspberry Pi 4, even one that’s been heavily overclocked. The Pi Case 40 will feature a cast aluminum body acting as the heat sink, with added fins along the block for additional heat soaking. 

Secondly, the Pi Case 40 never needs to be removed, as the case will allow access to all of the I/O on the board, even the 40-pin GPIO headers, as the case redirects those to the side for access. The Pi Case 40 also features a power button to turn the Pi on or off without having to remove the power source. The power button is also remappable, so it can be configured for a number of other functions, such as scripts, overclocking, or opening applications.

The Pi Case 40 will also include a set of VESA mounts, which should make it easy to mount behind a display or on the wall. Cooler Master claims the Pi Case 40 is an open-source project, and they’re using Kickstarter as a means to receive feedback from the community. Additionally, Cooler Master will make the schematics and 3D model for the Pi Case 40 available for free via its website. 

Currently, the Pi Case 40 is expected to ship in September. You can pre-order through Kickstarter for $25. 

Source: https://www.kickstarter.com/projects/coolermaster/pi-case-40/

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