Tag Archives: Silicon

Samsung Introduces the World to 2nm, Set to Storm the World 2025 Onward

Samsung just had their 7th annual Samsung Foundry Forum, and it was interesting. As the name suggests, the forum focuses more on Samsung’s in-house chip, semiconductor, and silicon development. They are, after all, the largest and own the world’s most advanced semiconductor manufacturing plant. They introduced the world to the 3nm process early on this year and is speculated to bring the first of the 3nm processed chips to smartphones later this year. The most powerful Android devices today are powered by Qualcomm’s 4nm Snapdragon 8 Gen2. Apple is also expected to update their latest iPhone launching this year to a 3nm chip.

This year’s Foundry Forum focuses not just on what is, but also what is to be in the future for Samsung’s semiconductor business. Samsung stresses the importance of artificial intelligence (AI) and its requirements, moving forward into the future. They spoke about how their own gate-all-around (GAA) advanced node technology would be the driving force of AI applications in the future. Of course, they also pointed to growth of their semiconductor business in the form of an expansion in production capacity, stretching their lead in the semiconductor production business over their competitors. However, they have not mentioned anything on their own Exynos chips made for smartphones. To be fair, there is still plenty to unpack at the Foundry Forum, especially when it comes to 2nm process.

Samsung detailed their plans to mass produce their 2nm process chips by 2025, first for mobile applications. That also means that you can expect Apple to introduce their iPhone with 3nm silicon in that same year. You can also expect Qualcomm to jump in line to introduce their Snapdragon 8 processor with 2nm process technology in late 2025 or early 2026 in this case. Most of all though, you can be hopeful on the return of the Samsung Exynos chips on Samsung’s flagship devices.

Samsung’s Exynos processing chips for mobile applications are still a thing, just not on the current line of flagship devices that Samsung pushes like the Galaxy S23 line-up and the Galaxy Z Fold. The Exynos chips are now powering the mid-range smartphones instead which, to be fair, is probably the most profitable segment in the industry. Still, that means that the Samsung Galaxy flagships devices gets only the Qualcomm Snapdragon treatment. While that is not a bad thing, it is a sad thing to see, especially when Samsung just renewed and extended its partnership with AMD to bring Radeon graphics power to Samsung’s in-house developed chips.

Still, Samsung’s commitment to pushing their production capacity and technology just means that they are not out of the fight in silicon supremacy just yet. With their renewed partnership with AMD, they did mention that they are looking to bring a new Exynos chip to market and into their own flagship smartphones or handheld devices. We could see an iteration of Samsung’s 3nm Exynos with AMD’s graphical prowess later in the year when they launch their lates Samsung Galaxy Z Fold. In the Forum, they also said that they are increasing their production capacity for foundry products made for mobile devices and other applications, which could be a direct result of them pushing the next-generation Exynos chips on to their latest flagship devices. It could also be because their customers are demanding even more chips too.

The Samsung 2nm processor is not just for mobile applications like smartphones or gaming handheld though. Since the Forum stressed AI applications, Samsung aims to put the 2nm chips into high-performance-computing (HPC) solutions that is meant to push AI technologies as well in 2026. Eventually, the technology will also be available for the automotive industry in 2027. In 2027 as well, Samsung is planning to start the mass production of 1.4nm chips.

The promise of 2nm process seems to be significant too. They claim that with the current development, the 2nm chip has proven to be 12% more powerful than the current 3nm chip. It is 5% smaller than the 3nm chip and is up to 25% more power efficient too, making the jump from 3nm to 2nm a rather large one in terms of performance and efficiency.

Samsung also promises that in 2025, the 2nm process technology will find its way into gallium nitride (GaN) power applications for mobile devices, data centers, and automotive industry. They are also advancing networking technology in 2025 with the 5nm Radio Frequency (RF) technology that supports 6G networks. They say that the 5nm RF process shows a 40% in power efficiency while taking up half the space of a traditional 14nm process.

We said earlier that you can be hopeful to see Samsung’s Exynos chips make a comeback in their flagship devices. That is, at least for now, just speculations and hopeful statements. There is no confirmation on Samsung reintroducing Exynos flagship class processors currently. You can also expect Apple to be the first to introduce a 3nm processor for their iPhones later in the year.

[WWDC 2022] Apple Launches the M2 for the New MacBook Air and 13-inch MacBook Pro

Let us talk about Apple’s first foray into the silicon market – the Apple M1 chip. It was a somewhat successful endeavor. The M1 powered MacBooks proved to be rather popular among fans and even to new users. It is also the first successful implementation of a System on a Chip (SoC) solution for a major computing device with a desktop interface that is the MacOS in this case.

Of course, the success of the M1 chip also has something to do with the fact that Apple has somehow engineered the MacOS around the integrated silicon. But it also has something to do with the Apple M1’s specific build to optimise and be optimized with apps designed for MacOS.

Of course, while we expected Apple to come up with more powerful versions of their silicon last year, we were a little wrong to expect the second-generation of Apple’s silicon to be announced then. Instead, Apple launched variants of the M1 chip. You get larger footprints of the M1 that also comes with a lot more grunt from the SoC. Everything changes for 2022 though with Apple’s M2.

Welcome to the Apple M2 silicon, supposedly an evolution of the M1 chip. Obviously, it will be more powerful than the regular M1 chip. As with the M1 chip as well, the M2 will start out its life in the MacBook Air and 13-inch MacBook Pro.

Why Should You Care?

Apple WWDC22 M2 chip CPU perf vs power 01 220606
Source: Apple

PC users might snicker at the thought of an SoC running their Windows system. It is not too far-fetched of an idea though. Windows 11 seems to be more SoC friendly, and Qualcomm is working on a solution for that as we write this piece.

But the M1 is prove that an SoC can power a desktop class computing solution. The M2 just solidifies that position for Apple as the only manufacturer that dare to go beyond conventions. While Apple did point out that the M2 can game better with Metal accelerator, it still will not game like Windows PC can.

Instead, the M2 is tailor built to ensure that the MacOS is fully optimized and runs as fast as it can. The M2 chip is designed specifically to ensure that the MacOS and its apps are getting enough processing power in all the right places to ensure smooth, snappy experience. It just does everything that the M1 does, but better.

Bigger, Badder, Better

  • Apple WWDC22 M2 SoC 220606
  • Apple WWDC22 M2 chip hero 220606
  • Apple WWDC22 M2 chip M1 chip 2up 220606
  • Apple WWDC22 M2 chip CPU perf vs power 01 220606
  • Apple WWDC22 M2 chip CPU perf vs power 02 220606
  • Apple WWDC22 M2 chip CPU perf vs power 03 220606
  • Apple WWDC22 M2 chip GPU perf vs power 01 220606
  • Apple WWDC22 M2 chip GPU perf vs power 02 220606
  • Apple WWDC22 M2 chip GPU perf vs power 03 220606

Footprint wise, the M2 is slightly larger than the M1 chip. It is built with he second-generation 5nm technology, technically the current most advanced transistor technology. The new construction allows Apple to push the envelope a little bit with the M2.

It is supposed to be about 18% faster than the M1 in terms of CPU performance. In terms of GPU performance, it is supposed to 35% better than the older chip. The M2 is also supposed to pack a dedicated Neural Engine that is about 40% faster than before. While the M1 tops out at 16GB of memory bandwidth, the basic M2 can support up to 24GB in unified memory. You can find all 24GB powering the updated 13-inch MacBook Pro.

The larger footprint can be contributed to the fact that Apple has squeezed 25% more transistors than the M1. More transistors also means that the chip can deliver up to 100GB/s in memory bandwidth with the speedy unified memory. At the same time, with a larger cache, the M2 does not require any more power than the M1 demanded. At the same maximum power consumption. The M2 still beats out the M1 in all processor aspects.

Faster Workflow = Better Productivity

Apple WWDC22 M2 chip hero 220606
Source: Apple

The result of all that can be seen in all the Apple silicon optimized apps. Thanks to the wider bandwidth, the M2 supports up to 8K H.264 and HEVC video encoding. It is supposed to also be able to stream more 4K and 8K videos on more displays, if you can find enough dongles or displays to plug the MacBook Air or the 13-inch MacBook Pro.

At the same time, while not spoken at length, the M2 features better security all around. Apple’s Secure Enclave ensures that your data is locally stored and protected instead of it being uploaded to iCloud. Of course, to fully take advantage of the chip’s power, you need Apple’s latest MacOS Ventura.

The MacBook Air – Now a Proper Baby MacBook?

  • Apple WWDC22 MacBook Air 4up colors hero 220606
  • Apple WWDC22 MacBook Pro 13 multitasking demo 220606
  • Apple WWDC22 MacBook Pro 13 Baldurs Gate III 220606
  • Apple WWDC22 MacBook Pro 13 Affinity Photo 220606
  • Apple WWDC22 MacBook Air ports 220606
  • Apple WWDC22 MacBook Air MagSafe 220606
  • Apple WWDC22 MacBook Air M2 performance 220606
  • Apple WWDC22 MacBook Air Liquid Retina Display 220606
  • Apple WWDC22 MacBook Air keyboard 220606
  • Apple WWDC22 MacBook Air iPhone 13 Continuity 220606
  • Apple WWDC22 MacBook Air hero 220606
  • Apple WWDC22 MacBook Air headphone jack 220606
  • Apple WWDC22 MacBook Air Final Cut Pro 220606
  • Apple WWDC22 MacBook Air Adobe Photoshop 220606

For the first time ever, the MacBook Air got a redesign. Instead of the wedge-shaped aluminium slate, it now looks like a thinner version of the regular MacBook. It is not a bad thing.

It now features a more updated “MacBook” look in a 13.6-inch Liquid Retina display with the weird notch in the top middle section of the display. That notch houses an 1080p Full HD FaceTime camera and other sensors.

The US 1,199 (starting from) device now looks a little more up to par to a regular MacBook too, at least on paper. It now features dour speakers and MagSafe charging has made its way back into the thinnest and lightest version of the MacBook family. There are still no fans in the notebook for any significant cooling performance, but the M2 is an SoC that does not really need a dedicate cooling fan. The MacBook Air also comes in four different colour finishes now instead of the usual two that was offered before.