Category: Computers

  • HP EliteBook X G1i 14: Exceptional Battery Life Among Elite Models

    HP EliteBook X G1i 14: Exceptional Battery Life Among Elite Models

    Key Takeaways

    1. The HP EliteBook X G1i 14 boasts impressive battery life, lasting over 18 hours of continuous WLAN browsing at 150 nits brightness.
    2. Users can choose between a 56 Wh or 68 Wh battery, allowing for a balance between weight and runtime.
    3. The new Lunar Lake CPU enhances performance-per-watt, contributing to longer battery life compared to older processors.
    4. The device is limited to 1200p60 or 1600p120 IPS displays, which helps regulate power consumption compared to higher resolution OLED displays.
    5. Battery life is affected by the choice of battery size and display resolution, with lower configurations yielding better runtime.

    HP often promotes “AI” as a major feature of its new EliteBook X G1i 14 subnotebook, but the focus might be better placed on the device’s battery life. The runtimes are quite remarkable for such a compact design. Our evaluations show that the device can endure over 18 hours of continuous WLAN browsing at a brightness of 150 nits, outlasting many competitors and even other recent models like the EliteBook Ultra G1i 14 or EliteBook X G1a 14 AI.

    Factors Behind Battery Longevity

    The impressive battery life of the EliteBook X G1i 14 is due to three key reasons. To begin with, users can choose between a smaller 56 Wh or a larger 68 Wh battery, depending on whether they prioritize lighter weight or extended runtimes. The unit we tested is equipped with the larger battery to enhance its endurance.

    The Impact of the Lunar Lake CPU

    Another significant reason is the new Lunar Lake CPU, designed to enhance performance-per-watt far better than earlier processors. In fact, competing laptops featuring Lunar Lake chips, like the Dell Pro 14 Premium, also demonstrate longer battery life compared to most devices running on Meteor Lake or older models.

    Resolution Choices Affecting Power Consumption

    Lastly, a noteworthy aspect is that the EliteBook X G1i 14 is limited to either 1200p60 or 1600p120 IPS displays, while the EliteBook Ultra G1i 14 offers an option for 1800p OLED. By restricting the model to lower resolution IPS screens, power usage can be more effectively regulated since OLED displays are known for their greater energy needs.

    All three reasons mentioned contribute to the EliteBook X G1i 14’s exceptional runtimes. However, it’s important to note that battery life will decrease if the device is configured with the smaller battery and/or the higher resolution 1600p120 display.

    Source:
    Link

    Tags: , ,
  • Planet Devoured by Its Star at an Alarming Rate

    Planet Devoured by Its Star at an Alarming Rate

    Key Takeaways

    1. TOI 1227 b is a young exoplanet, only 8 million years old, making it the second-youngest transiting planet detected.
    2. The planet’s atmosphere is being stripped away by intense X-rays from its star, causing it to appear “puffed up.”
    3. The planet’s mass is likely closer to that of Neptune, despite its large appearance similar to Jupiter.
    4. The X-ray bombardment strips away an amount of mass equal to one Earth atmosphere every 200 years.
    5. In about a billion years, TOI 1227 b is expected to lose its atmosphere completely, shrinking to a lifeless rock.

    Astronomers from NASA’s Chandra X-ray Observatory have observed a young exoplanet undergoing a dramatic change. The planet, called TOI 1227 b, is experiencing its atmosphere being slowly stripped away due to a relentless stream of X-rays from the star it orbits.

    A Disturbing Transformation

    This intense exposure to radiation has caused the atmosphere of the planet to expand, making it appear “puffed up” like Jupiter, even though its mass is probably closer to that of Neptune.

    “It’s hard to comprehend the scale of devastation this planet is facing. The atmosphere just can’t hold up against the intense X-ray bombardment from its star,” said Attila Varga from the Rochester Institute of Technology, who was in charge of the study.

    Age and Orbits

    TOI 1227 b is only 8 million years old, which is about 625 times younger than Earth, marking it as the second-youngest transiting planet ever detected. It orbits a red dwarf star located roughly 330 light-years away at a very close range — less than a fifth of the distance from Mercury to our Sun. While the star is cooler than the Sun when seen in visible light, it shines brightly in X-ray wavelengths.

    The Grim Future Ahead

    The research team has worked out that the X-ray bombardment from the star strips away an amount of mass equal to one Earth atmosphere approximately every 200 years. “The outlook for this young planet is not very promising,” said Alexander Binks, a co-author from the Eberhard Karls University of Tübingen. The study suggests that in about a billion years, the planet will completely lose its atmosphere and shrink to about a tenth of its current size, ultimately becoming a small, lifeless rock with no chance of supporting life.

    Source:
    Link

    Tags:
  • Beelink EQi12: Affordable Mini PC with Built-in PSU Launches

    Beelink EQi12: Affordable Mini PC with Built-in PSU Launches

    Key Takeaways

    1. The Beelink EQi12 now features the Intel Core i5 1235U processor, offering 41.33% higher performance on Time Spy compared to the Core i5 12450H.
    2. Enhanced graphics with the 1235U version include the Xe Graphics G7 with 80 execution units, surpassing the Xe G4 with 48 EUs in the 12450H.
    3. The updated model operates with up to 8% less noise than the previous configuration.
    4. Port configuration remains similar to earlier models, with a Type-C port limited to data transfer, and it supports WiFi 6, Bluetooth 5.2, and up to 32 GB of RAM.
    5. The starting price for the EQi12 with the Core i5 1235U is $319, available for purchase at the global Beelink online store.

    The Beelink EQi12 has gotten an upgrade, featuring a fresh setup with the Intel Core i5 1235U. This processor is part of the Alder Lake-U lineup, and the company asserts that it’s a more powerful option compared to the Core i5 12450H. According to the provided comparison, the 1235U can score 41.33% higher on Time Spy and shows a 25.56% improvement on Fire Strike.

    Improved Graphics and Performance

    Additionally, Beelink mentions that the 1235U version of the EQi12 has a more advanced integrated GPU than the one with the Core i5 12450H. Specifically, it’s equipped with the Xe Graphics G7, which has 80 execution units (EUs), while the 12450H has the Xe G4 with only 48 EUs. There’s also a notable reduction in system noise.

    Reduced Operating Noise

    The company points out that the operating noise level can be up to 8% less when compared to the 12450H configuration. Apart from that, most features of the new model are largely unchanged. The updated version still includes an 85 W built-in power supply and the MSC 2.0 cooling system.

    Port Configuration and Pricing

    When it comes to ports, the EQi12 has a setup that is similar to previous models. Specifically, it includes several ports, though it’s important to highlight that the Type-C port is not a fully functional one; it is used solely for data transfers. Other notable features include WiFi 6, Bluetooth 5.2, support for up to 32 GB of RAM, and the capability to handle up to 4 TB of storage (4 TB Crucial P310 priced at $229.99 on Amazon). The starting price for the EQi12 with the Core i5 1235U is $319, and it’s currently on sale at the global Beelink online store.

    Source:
    Link

    Beelink EQi12: Affordable Mini PC with Built-in PSU Launches

     

    Tags: ,
  • Kubuntu Focus NX Gen 3 Mini PCs with Intel Arrow Lake-H Launched

    Kubuntu Focus NX Gen 3 Mini PCs with Intel Arrow Lake-H Launched

    Key Takeaways

    1. Processor Upgrade: The Kubuntu Focus NX Gen 3 models are now equipped with the latest Intel Arrow Lake mobile processors, offering significant performance enhancements over the previous Alder Lake generation.

    2. Rebranding and Custom Tools: The Gen 3 models are rebranded Asus NUC 15 Pro Slim devices, featuring Kubuntu pre-installed along with custom tools like System Rollback and a one-click app installer for easier transitions from Windows.

    3. Performance Improvements: The Arrow Lake processors provide a 22% increase in single-core performance and up to 50% in multi-core performance, along with integrated Arc GPUs that enhance graphics capabilities significantly.

    4. Specifications and Connectivity: The mini PCs support up to 96 GB of DDR5 RAM and have two M.2 slots for storage, with extensive connectivity options including Thunderbolt 4 ports, USB ports, and HDMI 2.1 outputs.

    5. Pricing Comparison: The Kubuntu Focus NX Gen 3 is priced higher than comparable Asus models, with costs starting at $895 for the entry-level version and reaching $1,090 for a model with enhanced specifications.

    The mini PC and desktop sector of Kubuntu is updating the Focus NX series with the newest Intel Arrow Lake mobile processors. The earlier generation utilized Alder Lake chips, but the fresh Gen 3 models are equipped with either the Core Ultra 5 225H or the Core Ultra 7 255H SoCs.

    Rebranding and Features

    It appears that the Gen 3 models are actually rebranded Asus NUC 15 Pro Slim devices that come with Kubuntu pre-installed and several customized tools. These include an integrated System Rollback to mitigate compatibility problems, power and fan management tools, a one-click app installer, and a user-friendly interface aimed at making the transition smoother for Windows users.

    Performance Improvements

    Internal evaluations by the Kubuntu team indicate that the new Arrow Lake processors offer a 22% boost in single-core performance, with multi-core performance improvements potentially reaching 50%. The most significant upgrade from the previous generation is the inclusion of integrated Arc GPUs, which can provide up to 3x the frames per second (FPS) and support advanced features such as XeSS. Additionally, the SoC is equipped with a 13 TOPS NPU for applications powered by AI.

    Specifications and Connectivity

    The design includes two SO-DIMM slots, allowing users to expand memory up to 96 GB DDR5-5600 RAM. For storage, there are two M.2 slots—one supporting PCie 5.0 and the other limited to PCie 4.0—offering a total capacity of up to 6 TB.

    The unit, measuring 117 x 112 x 37 mm (4.6 x 4.4 x 1.5 inches), provides a generous port selection. It includes 2x Thunderbolt 4 ports, 1x USB-C 3.2 Gen2X2, 3x USB-A 3.2 Gen 2, 1x USB 2.0, 2x HDMI 2.1 video outputs, and a 2.5 GbE jack, although it lacks an audio jack. Wireless connections are facilitated by a Wi-Fi 7 + BT 5.4 card.

    Pricing Considerations

    In terms of pricing, the Kubuntu Focus NX Gen 3 is somewhat pricier compared to the Asus models that come with Windows pre-installed. For context, the Asus NUC 15 Pro Slim barebones with the Core Ultra 7 255H is listed at $684, whereas the Kubuntu mini PC version with the same CPU, 16 GB RAM, and 512 GB SSD retails for $1,090. The entry-level Core Ultra 5 225H option starts at $895.

    Source:
    Link

    Kubuntu Focus NX Gen 3 Mini PCs with Intel Arrow Lake-H Launched
    Tags:
  • Samsung Rejected Nvidia CEO Jensen Huang’s HBM Proposal in 2018

    Samsung Rejected Nvidia CEO Jensen Huang’s HBM Proposal in 2018

    Key Takeaways

    1. Samsung Foundry is struggling with low production yields and complications in HBM production, impacting its advanced process nodes.
    2. Samsung received approval to supply Nvidia with HBM3E memory for AI chips, but it still trails behind SK Hynix, the main supplier for Nvidia.
    3. Samsung missed an opportunity to collaborate with Nvidia in 2018, which could have positioned them as a key partner in HBM production and AI chip development.
    4. Nvidia’s reliance on SK Hynix for HBM has contributed to a significant increase in SK Hynix’s stock price, while Samsung has lost out on AI chip manufacturing opportunities.
    5. Samsung is now working to secure HBM business from Nvidia and AMD, aiming to provide HBM4 by early 2026.

    It’s no secret that Samsung Foundry is having a tough time. The company is said to be facing challenges with low production yields in its advanced process nodes, which has also affected the manufacturing of its upcoming 1.4 nm node. Additionally, there are complications surrounding High-Bandwidth Memory (HBM) production.

    Partnership Dynamics

    Recently, Samsung has received approval to provide Nvidia with HBM3E memory for AI chips after missing out on the approval last year. Despite this, the company still lags behind SK Hynix, which is the primary supplier of HBM for Nvidia. The situation could have been quite different if Samsung had not turned down Nvidia’s offers back in 2018.

    A new report reveals that Nvidia’s CEO, Jensen Huang, sought to expand the HBM partnership with Samsung in 2018. Huang reportedly visited Samsung to discuss collaboration in three main areas: HBM production, the development of new process nodes, and cooperation on CUDA. Such a partnership could have positioned Samsung as a pivotal ally for Nvidia.

    Missed Opportunities

    However, Samsung declined Nvidia’s proposals. Huang expressed his frustration, saying, “there was no one at Samsung to discuss long-term strategy with me.” This inability to engage in strategic discussions may have been linked to an ongoing criminal investigation involving Samsung Electronics Chairman Lee Jae-Yong at that time.

    Nvidia has significantly benefitted from the AI boom due to its AI chips, which have seen remarkable demand. HBM is critical for these AI superchips, as components like the H200 require substantial high-throughput memory for AI inference tasks.

    Market Impact

    A large portion of Nvidia’s HBM chips comes from SK Hynix, which has greatly benefitted from this arrangement. At the present moment, SK Hynix’s stock price has surged more than 220% over the past five years.

    Samsung’s choice to not collaborate with Nvidia may have repercussions for Samsung Foundry as well. Currently, TSMC manufactures all of Nvidia’s AI chips, including the latest B200 GPU. There have been speculations that Nvidia could engage Samsung to fabricate its upcoming gaming chips using the 2 nm process, but there is no concrete news about Samsung making Nvidia’s AI accelerators.

    With Intel facing its own troubles and Samsung struggling to keep pace with TSMC, TSMC has gained a monopoly on Nvidia’s AI GPUs.

    Future Prospects

    While it may be too late for Samsung to fully counteract the consequences of its past decisions, the company is now working to secure some HBM business from Nvidia and AMD. Reports suggest that Samsung is aiming to provide HBM4 to both companies in early 2026.

    Source:
    Link

    Tags:
  • Microsoft Unveils Surface Laptop 5G with eSIM and Nano SIM from $1799

    Microsoft Unveils Surface Laptop 5G with eSIM and Nano SIM from $1799

    Key Takeaways

    1. Microsoft has launched the Surface Laptop 5G, a 5G-enabled version of the 7th Edition, aimed at business users.
    2. The laptop features a built-in 5G modem, eSIM module, nano SIM card slot, and a powerful Neural Processing Unit (NPU).
    3. It includes a dynamic antenna system with six antennas that optimize signal strength based on usage and can switch between Wi-Fi and 5G.
    4. Pricing starts at $1799.99 for the base model with Intel Core Ultra 5, 16GB RAM, and 256GB storage, while the high-end model is priced at $2,699.99.
    5. The Surface Laptop 5G will initially be available only in Platinum Silver, with shipping starting on August 26.

    Microsoft has introduced a new model of the Surface Laptop, named the Surface Laptop 5G, which is a 5G variant of the 7th Edition. This latest laptop is essentially a 5G-enabled version of the 13.8-inch Surface Laptop tailored for business users and runs on Intel’s Ultra Processor (Series 2).

    Features and Specifications

    The Surface Laptop 5G comes with a built-in 5G modem, an eSIM module, and a nano SIM card slot located on the right side. As stated in Microsoft’s press release, this laptop is equipped with a Neural Processing Unit (NPU) that can perform over 40 trillion operations per second (TOPS), enabling quick, on-device intelligence.

    Connectivity and Performance

    Additionally, the Surface Laptop 5G includes a dynamic antenna system featuring six antennas placed in key positions. These antennas adjust their signal paths and power automatically based on how the laptop is held or used, ensuring reliable connectivity. Moreover, the laptop can switch effortlessly between Wi-Fi and 5G networks and can also function as a mobile hotspot for other devices.

    Pricing Options

    The starting price for the Surface Laptop 5G is $1799.99, which is $300 more than the non-5G model that launched in February. For this price, you get an Intel Core Ultra 5, 16GB of RAM, and 256GB of storage. There’s also a higher-end model priced at $2,699.99, featuring an Intel Core Ultra 7, 32GB of RAM, and a 1TB SSD. Initially, Microsoft will offer the laptop only in Platinum Silver, with shipping beginning on August 26.

    Microsoft has not yet announced any plans for a 5G version of the Surface Laptop 7 that would utilize Snapdragon processors (Snapdragon X Plus 16GB + 256GB is available on Amazon for $840).

    Source:
    Link

    Tags:
  • Nvidia Adds CUDA Support for RISC-V Architecture

    Nvidia Adds CUDA Support for RISC-V Architecture

    Key Takeaways

    1. Nvidia announced native support for CUDA on RISC-V, highlighting its potential as a competitor to x86 and ARM architectures.
    2. The announcement was made by Nvidia’s VP, Frans Sijstermans, at a RISC-V summit in China, showcasing how RISC-V can manage CUDA drivers.
    3. The integration of CUDA with RISC-V may help Nvidia expand its ecosystem in China, where RISC-V is growing rapidly.
    4. AMD is developing its own alternative to CUDA, called ROCm, which has also added support for RISC-V, indicating competition in the market.
    5. The adoption of ROCm is slow, suggesting it may take time before there is significant competition against Nvidia’s CUDA in computing software.

    Back in 2021, researchers focusing on RISC-V showed that Nvidia’s CUDA code could operate on non-proprietary hardware like a Vortex GPGPU that is based on RISC-V, utilizing an OpenCL translator. While this might not be the most effective method to implement CUDA support for RISC-V, it clearly indicated a rising interest in RISC-V as a serious competitor to x86 and ARM processing architectures. Nvidia is now formally recognizing the potential of RISC-V in the computing field by announcing the native support of CUDA for RISC-V.

    Significant Announcement

    The announcement was made by Frans Sijstermans, Nvidia’s Vice President of Hardware Engineering, during a RISC-V summit held in China. Sijstermans is also a director on the RISC-V board. A diagram showcased during the event illustrated how the RISC-V processor can manage CUDA drivers at the operating system level, while Nvidia’s GPUs run the CUDA kernels. There is a Data Processing Unit (DPU) mentioned, presumably from Nvidia as well, which indicates that the diagram depicts a computing system for high-performance computing (HPC) and data centers.

    Industry Reactions

    According to Tom’s Hardware, this step to provide CUDA support for open-source architectures like RISC-V might allow Team Green to broaden its ecosystem in China, where RISC-V is experiencing rapid growth, despite the limitations on marketing AI accelerators like the GB200 and GB300 in that area.

    In contrast, AMD is pushing its own alternative to Nvidia’s CUDA through ROCm, which is now in its seventh version and has already added support for the RISC-V architecture. Numerous companies have voiced their opposition to the dominance of CUDA, and Team Red is making efforts to challenge the monopoly. However, the adoption of ROCm remains gradual, and it may take a few more years before we witness true competition in the realm of computing software.

    Source:
    Link

    Tags:
  • YMTC Establishes Homegrown NAND Production to Avoid U.S. Sanctions

    YMTC Establishes Homegrown NAND Production to Avoid U.S. Sanctions

    Key Takeaways

    1. YMTC aims for 150,000 wafer starts per month and 15% of the global NAND market by 2026, despite being on the U.S. Entity List since late 2022.
    2. The company plans to increase bit output faster than the industry’s expected growth of 10-15% in 2025, with 130,000 WSPM projected by late 2024.
    3. Upcoming products include a 1 TB TLC device and a 3D QLC X4-6080, with future innovations expected to exceed 300 layers for better bit maximization.
    4. YMTC intends to launch a pilot line using only Chinese-made tools by late 2025 to reduce dependence on foreign machinery, potentially doubling output.
    5. Significant challenges remain in achieving production consistency and efficiency, particularly in extreme-ultraviolet lithography, critical for successful semiconductor manufacturing.

    Yangtze Memory Technologies Co. (YMTC) has been on the U.S. Commerce Department’s Entity List since late 2022. Despite this, the firm based in Wuhan is moving forward with plans to expand its operations. The goal is to achieve about 150,000 wafer starts per month (WSPM) this year and capture 15 percent of the global NAND market by the end of 2026.

    Ambitious Growth Plans

    By late 2024, YMTC was already on track for approximately 130,000 WSPM, which would represent around 8 percent of the total global capacity. The company has also started shipping its 232-layer TLC parts (X4-9070), which are made by combining two layers to reach a total of 294 layers. While many of its competitors are cutting back on capital spending, YMTC aims to increase its bit output faster than the industry’s expected growth of 10-15 percent in 2025.

    Product Innovations on the Horizon

    The company’s roadmap includes a 1 TB TLC device and a 3D QLC X4-6080 expected later this year, along with a 2 TB TLC X5-9080 featuring a 4.8 GT/s interface planned for 2026. Future products are anticipated to exceed 300 layers by stacking three decks, which would shift the focus from throughput to maximizing bits per wafer.

    Aiming for Self-Sufficiency

    In a bid to lessen reliance on foreign machinery, YMTC intends to launch a pilot line using only Chinese-made tools in the latter half of 2025. Analysts suggest that if this initiative is successful, it could potentially double bit output and increase market share beyond 15 percent. However, they also note that this facility is still in the experimental phase, and achieving full production will take time.

    Reports from TechInsights indicate that YMTC’s newest “Xtacking 4.0” chips perform comparably to those of leading competitors. However, the company has had to reduce the number of layers due to domestic equipment that still struggles with yield. Significant challenges remain, particularly in extreme-ultraviolet lithography, which is not yet accessible to China. Progress will depend on bridging the equipment and yield gap, even as local vendors like AMEC, Naura, and Piotech work to improve their technologies.

    Overall, YMTC’s plans for capacity expansion, multi-deck NAND development, and efforts toward localization demonstrate the shift of China’s semiconductor self-sufficiency goals from mere policy statements to actual production. Whether these initiatives will result in a sustainable 15 percent market share by 2026 will rely more on the development of domestic fabrication tools and consistent production yields than on ambition alone.

    Source:
    Link

    Tags:
  • Sixunited STHT1 Mini-ITX Motherboard Launches with AMD Strix Halo

    Sixunited STHT1 Mini-ITX Motherboard Launches with AMD Strix Halo

    Key Takeaways

    1. Sixunited launched a new laptop with an OLED display powered by AMD’s Ryzen AI Max Strix Halo processors.
    2. The STHT1 motherboard is designed for DIY PC enthusiasts, measuring 170 mm x 170 mm and compatible with standard mini-ITX cases.
    3. The STHT1 motherboard features soldered LPDDR5X memory with a maximum capacity of 128 GB and includes two M.2 2280 NVMe SSD slots.
    4. Connectivity options include various USB ports, an HDMI 2.1 port, a DP/VGA/COM port, a 3.5 mm audio jack, and a gigabit Ethernet port.
    5. Pricing for the STHT1 motherboard has not yet been announced, and it is primarily aimed at OEMs for integration into their products.

    Chinese company Sixunited has recently launched a new laptop that features an OLED display and is powered by AMD’s Ryzen AI Max Strix Halo processors. Prior to this, they showcased a mini-PC containing the same components. Now, they’ve introduced the STHT1 motherboard, aimed at DIY PC enthusiasts. This motherboard measures 170 mm x 170 mm and fits into any standard mini-ITX case.

    Memory and Storage Options

    Like other Strix Halo motherboards, the LPDDR5X memory on the STHT1 is soldered, meaning it cannot be upgraded. The motherboard supports a maximum of 128 GB of memory and includes two M.2 2280 NVMe SSD slots. There’s also an extra M.2 slot available for adding a Wi-Fi module. It requires a 19-Volt (2×2 ATX) DC input for power.

    Connectivity Features

    For its I/O options, the Sixunited STHT1 provides two configurations: either 2x USB 3.2 Gen2 Type-C and 2x USB 2.0 Type-A, or 1x USB 3.2 Gen2 Type-C (which supports data plus 15W PD) along with 1x USB 3.2 Gen1 Type-A and 2x USB 2.0 Type-A. Additionally, users can expect an HDMI 2.1 port, a DP/VGA/COM port, a 3.5 mm audio/microphone jack, and a gigabit Ethernet port.

    Pricing Information

    Sadly, Sixunited has yet to announce the pricing for the STHT1 motherboard. Although it can be utilized by individual PC builders for a mini-ITX setup, it’s mainly designed for OEMs to include in their products.

    Source:
    Link

    Sixunited STHT1 Mini-ITX Motherboard Launches with AMD Strix Halo
    Tags:
  • Kioxia Launches First 245.76 TB NVMe SSD for AI Workloads

    Kioxia Launches First 245.76 TB NVMe SSD for AI Workloads

    Key Takeaways

    1. Kioxia has launched the LC9 Series NVMe SSD with a capacity of 245.76 TB, the highest available for both 2.5-inch and EDSFF E3.L formats.
    2. The new SSD doubles the previous maximum capacity of 122.88 TB, catering to data centers with high space efficiency needs.
    3. The drive supports demanding applications like generative AI and vector databases by reducing I/O bottlenecks during data processing.
    4. It features advanced technology with 32 BiCS FLASH QLC dies and a PCIe 5.0 x4 NVMe 2.0 controller, offering fast speeds of up to 12 GB/s for reads.
    5. The drive includes reliability features like die-level recovery and power-loss protection, achieving around 1.3 million IOPS for reads and 50,000 IOPS for writes.

    Kioxia has introduced a new model with a capacity of 245.76 TB in its LC9 Series, making it the highest-capacity NVMe SSD available in both 2.5-inch and EDSFF E3.L formats. This new version doubles the previous maximum capacity of 122.88 TB in the series and is designed for data centers where space efficiency is crucial.

    Demand for High-Capacity Storage

    The needs of generative AI training, vector databases, and various retrieval-augmented generation tasks require large, efficient, and speedy storage solutions. Kioxia emphasizes that the LC9’s large scale is meant to alleviate I/O bottlenecks, which can cause costly GPUs to sit idle during data ingestion and inference processes.

    Advanced Technology and Performance

    This drive utilizes 32 of Kioxia’s 2 Tbit BiCS FLASH QLC dies, employing CMOS-direct-bonded-to-array (CBA) technology that allows for 8 TB in a single NAND package, marking another industry milestone. It features a PCIe 5.0 x4 (dual-port) NVMe 2.0 controller that offers up to 12 GB/s for sequential reads and 3 GB/s for writes, effectively balancing signal integrity with the high package density and extreme height.

    Specifications and Reliability

    In terms of random throughput, the drive achieves around 1.3 million IOPS for reads and 50,000 IOPS for writes, with an endurance rating of 0.3 DWPD suitable for typical hyperscale workloads. To ensure reliability, it includes die-level recovery, parity-based error management, and power-loss protection, in addition to Flexible Data Placement that helps minimize write amplification.

    Kioxia is currently sampling the 245 TB LC9 to selected partners, with larger volume units to be available in 2.5-inch U.2 (up to 122.88 TB), E3.S (up to 122.88 TB), and E3.L (245.76 TB) formats. A public demonstration is scheduled for the Future of Memory and Storage conference in August 2025.

    Source:
    Link

    Tags: