HP, Asus and Acer begin using Chinese CXMT chips amid memory shortage

Bad news! President Trump may not like it! 😊 

How much pricier are personal computers or computing devices now?

How many memory chips does a data center need? "... large-scale AI data centers collectively consume up to 70% of global memory chip production, driven by components like high-bandwidth memory (HBM) and enterprise SSDs.

Individual high-end AI server racks can pack tens of terabytes of memory, with single AI accelerators utilizing anywhere from 180 GB to 288 GB of high-bandwidth memory per processor chip." (Google Search)

"China's ChangXin Memory Technologies (CXMT) is benefiting from the global memory chip supply crunch. 

Leading PC makers HP, Asus and Acer have started to use small amounts of chips from China's ChangXin Memory Technologies amid an unprecedented memory shortage fueled by demand for artificial intelligence infrastructure ..."

HP, Asus and Acer begin using CXMT chips amid memory shortage - Nikkei Asia (behind paywall) "But PC makers careful to avoid angering top suppliers Micron, Samsung and SK Hynix"

Source

2D quantum memory device reaches single-electron limit of information storage

Amazing stuff!

"... a design strategy for making individual-electron states easier to discern at room temperature ..."

From the editor's summary and abstract:

"Editor’s summary

Realizing stable and distinguishable quantum memory at the single-electron level has long been challenged by increased parasitic gate-channel fringe capacitance, which substantially lowers the threshold voltage shift (ΔVth) caused by stored electrons as device dimensions are scaled down.

Using atomically thin two-dimensional materials and edge-contacted metal electrodes, Liu et al. designed a coplanar drain channel source structure that enabled a nonvolatile ΔVth of 0.5 volts at room temperature. Their work demonstrates the ability to control single-electron quantum behavior, offering promising prospects for nanoscale memory device engineering. ...

Abstract

The ultimate goal of information storage is single-electron memory.

Quantum mechanics predicts that two distinguishable quantum states can be realized by confining a single electron within an ultrasmall space.

However, scaling down such devices paradoxically amplifies fringe capacitance effects, which hinders the experimental observation of single-electron memory.

We report a two-dimensional single-electron memory device based on a coplanar drain-channel-source structure that suppressed fringe capacitance, exhibiting a nonvolatile threshold voltage shift of 0.5 volts after the change of a single electron.

Two intriguing quantum behaviors have also been verified regarding the programming voltage.

Additionally, we have predicted and observed a distinctive quantum memory effect: A quantum state is cut off by density of states scissors."

2D quantum memory device reaches single-electron limit of information storage

Robust single-electron memory with quantum states manipulation (no public access)

Room-temperature 2D single-electron memory device

Huge Memory AI Server Aims to Shatter the Memory Wall with up to 128 TB of DRAM per server

This is huge! Mind boggling!

"Memory is arguably the most serious constraint on modern AI large language models (LLMs). According to one influential paper, LLM token generation is an inherently memory-bound task, meaning the rate at which models output text is limited by how quickly data can be read in from memory. The severity of this bottleneck grows with model size. This creates a “memory wall” that holds back LLM inference performance.

AI hardware startup Majestic Labs is taking a direct—and comprehensive—approach to solving this problem. It’s developing a new AI server, Prometheus, with up to 128 terabytes of memory. That’s over 60 times more than Nvidia’s DGX B300 server, a cutting-edge AI processing rack. ..."

Huge Memory AI Server Aims to Shatter the Memory Wall - IEEE Spectrum "Majestic Labs’ Prometheus packs up to 128 TB of DRAM per server"

Source

Antiferromagnets could be better than ferromagnets for some ultrafast, high-density memories

Good news!

"While antiferromagnets show much promise for spintronics applications, they have proved more difficult to control compared to ferromagnets. Researchers in Japan have now succeeded in switching an antiferromagnetic manganese–tin nanodot using electric current pulses as short as 0.1 ns. Their work shows that these materials can be used to make efficient high-speed, high-density, memories that operate at gigahertz frequencies, so outperforming ferromagnets in this range. ..."

"Advances in spintronics have led to the practical use of magnetoresistive random-access memory (MRAM), a non-volatile memory technology that supports energy-efficient semiconductor integrated circuits. Recently, antiferromagnets─magnetic materials with no net magnetization─have attracted growing attention as promising complements to conventional ferromagnets. While their properties have been extensively studied, clear demonstrations of their technological advantages have remained elusive.

Now, researchers ... have provided the first compelling evidence of the unique benefits of antiferromagnets. Their study shows that antiferromagnets enable high-speed, high-efficiency memory operations in the gigahertz range, outperforming their ferromagnetic counterparts. ..."

From the editor's summary and abstract:

"Editor’s summary

Antiferromagnetic spintronics hold the promise of high speed and high efficiency unachievable with ferromagnets. However, reaching these goals in experiments has proven tricky. Takeuchi et al. realized all-electrical driving of an antiferromagnetic Mn3Sn (manganese-tin) nanodot. The researchers used electric-current pulses as short as 0.1 nanoseconds at current densities insensitive to pulse width. ...

Abstract

Electric current driving of antiferromagnetic states at radio or higher frequencies remains challenging to achieve. In this study, we report all-electrical, gigahertz-range coherent driving of chiral antiferromagnet manganese-tin (Mn3Sn) nanodot samples. High coherence in multiple trials and threshold current insensitive to pulse width, in contrast to results observed with ferromagnets, were achieved in subnanosecond range, allowing 1000/1000 switching by 0.1-nanosecond pulses at zero field. These features are attributed to the inertial nature of antiferromagnetic excitations. Our study highlights the potential of antiferromagnetic spintronics to combine high speed and high efficiency in magnetic device operations."

Antiferromagnets could be better than ferromagnets for some ultrafast, high-density memories – Physics World

Antiferromagnets Outperform Ferromagnets in Ultrafast, Energy-Efficient Memory Operations (original news release)

Electrical coherent driving of chiral antiferromagnet (no public access)

(a) Schematic illustration of memory device consisting of chiral antiferromagnet Mn3Sn / nonmagnetic metal heterostructure (b) A scanning electron microscope image of the fabricated device with Mn3Sn nanodot and nonmagnetic metal channel. ©Yutaro Takeuchi et al.

Major energy savings with new memory chip breakthrough

Good news!

"... Now, researchers have developed layered material that can reduce the energy consumed by memory devices by a factor of 10 by doing away with the need for power-hungry external magnetic fields.

The alloy is made from the magnetic elements cobalt and iron, and nonmagnetic elements germanium and tellurium. It allows 2 opposing magnetic forces to coexist in the same thin material.

Until now, this has only been possible by stacking different ‘ferromagnetic’ and ‘antiferromagnetic’ materials in multilayer structures. ..."

"... To store information, memory devices must switch the direction of electrons within a material. With conventional materials, this typically requires an external magnetic field to alternate the electron orientation. ... new material, however, features a built-in combination of opposing magnetic forces that create an internal force and tilted overall magnetic alignment.

“This tilt allows electrons to switch direction rapidly and easily without the need for any external magnetic fields. By eliminating the need for power-hungry external magnetic fields, power consumption can be reduced by a factor of ten,” ..."

From the abstract:

"The discovery of van der Waals (vdW) magnetic materials exhibiting non-trivial and tunable magnetic interactions can lead to exotic magnetic states that are not readily attainable with conventional materials.

Such vdW magnets can provide a unique platform for studying new magnetic phenomena and realizing magnetization dynamics for energy-efficient and non-volatile spintronic memory and computing technologies.

Here, the coexistence of ferromagnetic and antiferromagnetic orders in vdW magnet (Co0.5Fe0.5)5-xGeTe2 (CFGT) above room temperature, inducing an intrinsic exchange bias and canted perpendicular magnetism is discovered.

Such non-trivial intrinsic magnetic order enables to realize energy-efficient, magnetic field-free, and deterministic spin-orbit torque (SOT) switching of CFGT in heterostructure with Pt.

These experiments, in conjunction with density functional theory and Monte Carlo simulations, demonstrate the coexistence of non-trivial magnetic orders in CFGT, which enables field-free SOT magnetization dynamics in spintronic devices."

Major energy savings with new memory chip breakthrough

Material breakthrough paves way for major energy savings in memory chips (original news release) "It is anticipated that, within just a few decades, the surging volume of digital data will constitute one of the world’s largest energy consumers. Now, researchers ... have made a breakthrough that could shift the paradigm: an atomically thin material that enables two opposing magnetic forces to coexist – dramatically reducing energy consumption in memory devices by a factor of ten. This discovery could pave the way for a new generation of ultra-efficient, reliable memory solutions for AI, mobile technology and advanced data processing."

Coexisting Non-Trivial Van der Waals Magnetic Orders Enable Field-Free Spin-Orbit Torque Magnetization Dynamics (open access)

Figure 1

Coexistence of ferro- and antiferro-magnetic orders in a single atomic stacking nanolayers of (Co0.5Fe0.5)5-xGeTe2. 

A multi-level breakthrough in optical computing—a faster, more efficient, and robust memory cell

Amazing stuff!

"Until now, researchers have been limited in developing photonic memory for AI processing – gaining one important attribute like speed while sacrificing another like energy usage. In the article, the international team demonstrates a unique solution that addresses current limitations of optical memory that have yet to combine non-volatility, multibit storage, high switching speed, low switching energy, and high endurance in a single platform. ...

The authors propose a resonance-based photonic architecture which leverages the non-reciprocal phase shift in magneto-optical materials to implement photonic in-memory computing. ...

By using magneto-optic memory cells comprised of heterogeneously integrated cerium-substituted yttrium iron garnet (Ce:YIG) on silicon micro-ring resonators, the cells cause light to propagate bidirectionally, like sprinters running opposite directions on a track. ...

The team is now working to scale up from a single memory cell to a large-scale memory array which can support even more data for computing applications. They note in the article that the non-reciprocal magneto-optic memory cell offers an efficient non-volatile storage solution that could provide unlimited read/write endurance at sub-nanosecond programming speeds. ..."

From the abstract:

"Processing information in the optical domain promises advantages in both speed and energy efficiency over existing digital hardware for a variety of emerging applications in artificial intelligence and machine learning. A typical approach to photonic processing is to multiply a rapidly changing optical input vector with a matrix of fixed optical weights. However, encoding these weights on-chip using an array of photonic memory cells is currently limited by a wide range of material- and device-level issues, such as the programming speed, extinction ratio and endurance, among others. Here we propose a new approach to encoding optical weights for in-memory photonic computing using magneto-optic memory cells comprising heterogeneously integrated cerium-substituted yttrium iron garnet (Ce:YIG) on silicon micro-ring resonators. We show that leveraging the non-reciprocal phase shift in such magneto-optic materials offers several key advantages over existing architectures, providing a fast (1 ns), efficient (143 fJ per bit) and robust (2.4 billion programming cycles) platform for on-chip optical processing."

A multi-level breakthrough in optical computing—a faster, more efficient, and robust memory cell "For the first time, an international cadre of electrical engineers has developed a new method for photonic in-memory computing that could make optical computing a reality in the near future."

A Multi-Level Breakthrough in Optical Computing (original news release) "Engineers at Pitt, UC Santa Barbara, University of Cagliari, and Institute of Science Tokyo Demonstrate Faster, More Efficient, and Robust Memory Cell"

Integrated non-reciprocal magneto-optics with ultra-high endurance for photonic in-memory computing (open access)

Fig. 1: Non-reciprocal photonic in-memory computing.

New type of computer memory could greatly reduce energy use and improve performance

Good news!

"... This method of changing the electrical resistance in computer memory devices, and allowing information processing and memory to exist in the same place, could lead to the development of computer memory devices with far greater density, higher performance and lower energy consumption. ...
One potential solution to the problem of inefficient computer memory is a new type of technology known as resistive switching memory. Conventional memory devices are capable of two states: one or zero. A functioning resistive switching memory device however, would be capable of a continuous range of states – computer memory devices based on this principle would be capable of far greater density and speed. ..."

From the abstract (written exclusively for subject matter experts):

"A design concept of phase-separated amorphous nanocomposite thin films is presented that realizes interfacial resistive switching (RS) in hafnium oxide–based devices. The films are formed by incorporating an average of 7% Ba into hafnium oxide during pulsed laser deposition at temperatures ≤400°C. The added Ba prevents the films from crystallizing and leads to ∼20-nm-thin films consisting of an amorphous HfOx host matrix interspersed with ∼2-nm-wide, ∼5-to-10-nm-pitch Ba-rich amorphous nanocolumns penetrating approximately two-thirds through the films. This restricts the RS to an interfacial Schottky-like energy barrier whose magnitude is tuned by ionic migration under an applied electric field. Resulting devices achieve stable cycle-to-cycle, device-to-device, and sample-to-sample reproducibility with a measured switching endurance of ≥104 cycles for a memory window ≥10 at switching voltages of ±2 V. Each device can be set to multiple intermediate resistance states, which enables synaptic spike-timing–dependent plasticity. The presented concept unlocks additional design variables for RS devices."

New type of computer memory could greatly reduce energy use and improve performance | University of Cambridge Researchers have developed a new design for computer memory that could both greatly improve performance and reduce the energy demands of internet and communications technologies, which are predicted to consume nearly a third of global electricity within the next ten years.

Thin-film design of amorphous hafnium oxide nanocomposites enabling strong interfacial resistive switching uniformity (open access)

IBM's Roaring Come Back (In AI)

Posted: 12/8/2018

Remembered For Typewriters & Mainframes

In boxing, they say heavyweight champions once defeated never come back! IBM was for many years counted down for being a staid IT behemoth of former mainframe glory.

Their database technology is rather closed up and not well known outside of Fortune 500 companies or so. Even Oracle database technology is more accessible than that of IBM.

Was NIPS 2018 The Stage For A Comeback

IBM has for some time tried to come out with new innovative chip technologies or promote IBM Watson, but so far it appears these attempts were not exactly business success stories. Not sure, how much the IBM cloud offering is doing compared to its competitors, but it appears Microsoft oder Amazon are doing so much better.

However, at least since 2015 IBM was pushing for better hardware for AI (see e.g. Deep Learning with Limited Numerical Precision, Approximate Computing: Challenges And Opportunities). Certainly, AI needs urgently better and more powerful hardware!

At NIPS 2018, IBM came out showcasing their advances in hardware for AI, taunting Nvidia that GPUs had come to their limits (“post GPU era”):

1. IBM Scientists Demonstrate Mixed-Precision In-Memory Computing for the First Time; Hybrid Design for AI Hardware
2. 8-Bit Precision for Training Deep Learning Systems
3. Dual 8-Bit Breakthroughs Bring AI to the Edge
4. The future of hardware is AI
5. Training Deep Neural Networks with 8-bit Floating Point Numbers
6. Mixed-Precision In-Memory Computing

Are we entering a post von Neumann architecture era? IBM wants us to believe it.

We’ll see what the future holds for phase change memory; in memory computing; analog hardware; 8-bit precision; mixed precision computing etc. Stay tuned for more exciting things to come!

Let’s see how the competition will respond to this latest foray by IBM!