According to a report from Tom’s Hardware, while major tech companies are investing heavily in datacenter GPUs, the lifespan of these GPUs may only be 1 to 3 years, depending on their utilization rates.

The report, citing a general architect at Alphabet, noted that because GPUs are under heavy workload of AI training and inference, they tend to wear out more quickly than other components.

According to the report, in datacenters operated by cloud service providers (CSPs), the utilization rate of GPUs for AI workloads ranges from approximately 60% to 70%.

The report indicated that, citing the words from the general architect at Alphabet, at this utilization rate, a GPU can typically survive for 1 to 2 years, or up to 3 years. While the report stated that this claim cannot be considered 100% accurate and requires further confirmation, it highlighted that modern datacenter GPUs for AI and HPC applications consume and dissipate 700W of power or more, which is significant stress for chips.

One way to extend the life of the GPUs is to reduce the utilization rate, according to the report. However, to reduce the utilization rate implies that the GPUs will lose value more gradually and it will take longer to return their capital, which isn’t ideal for business. Therefore, the report pointed out that most cloud service providers will use their GPUs at a high utilization rate.

The report also references a study conducted by Meta, which describes training its Llama 3 405B model on a cluster powered by 16,384 NVIDIA H100 80GB GPUs. According to the report, in that study, the model flop utilization (MFU) rate of the cluster was about 38% (using BF16), while during a 54-day pre-training snapshot, out of 419 unforeseen disruptions, 148 (30.1%) were caused by GPU failures (including NVLink fails) and 72 (17.2%) were due to HBM3 memory failures.

This result carried out by Meta, according to the report, is quite favorable for NVIDIA’s H100 GPUs. If GPUs and their memory fail at Meta’s rate, the annualized failure rate will be about 9%, and in 3 years, it will be about 27%. However, GPUs will likely fail more frequently after a year of heavy use, as the report pointed out.

(Photo credit: NVIDIA)

Recently, “Black Myth: Wukong” created by Chinese team was released and quickly became a smash hit among game players worldwide with its stunning graphics, rich gameplay, and unique storyline.

Behind this game lies the application of several cutting-edge technologies, including virtual reality and artificial intelligence. These technologies are supported by the underlying hardware, enabling the game to deliver realistic and breathtaking visual effects while ensuring smooth gameplay and responsiveness. Notably, GPU and memory devices represent two key components in this process.

• High-Performance GPU to Meet Game Demands

“Black Myth: Wukong” features high-definition visuals and complex effects, supporting ray tracing technology that simulates real-world light behavior to provide more realistic lighting and shadow effects. This requires high-performance GPU to offer powerful graphics processing capabilities, which thus ensures a smooth gaming experience.

NVIDIA provided significant technical support for “Black Myth: Wukong.” Its RTX 40 series GPU supports full panoramic ray tracing and Deep Learning Super Sampling (DLSS) 3 technology, delivering smoother and higher-quality game visuals.

NVIDIA showcased “Black Myth: Wukong” at Gamescom, the world’s largest gaming convention, highlighting the game’s performance at the highest visual settings. NVIDIA also announced that “Black Myth: Wukong” is now part of the GeForce Now cloud gaming service, allowing subscribers to play the game without lengthy download times.

Additionally, Colorful collaborated with NVIDIA to launch a special edition “iGame GeForce RTX 4080 SUPER Advanced Black Myth: Wukong” graphics card, based on Colorful’s iGame GeForce RTX 4080 SUPER Advanced OC model.

This card features deep customization in appearance, 16GB of large memory, a powerful cooling system, and DLSS 3 technology. With local AI computing power of up to 836 TOPS, it can accelerate creative work even further.

Other manufacturers like GALAX, MSI, and Gigabyte have also recently rolled out custom or co-branded graphics cards based on the RTX 40 series for “Black Myth: Wukong.”

• SSD is in vogue

In light of the official PC specifications, “Black Myth: Wukong” requires a minimum of 16GB RAM and at least 130GB of memory space for the game’s main files, updates, and potential temporary files or cache data.

Compared to HDD (hard disk drives), SSD (solid-state drives) offers faster read/write speeds, lower latency, and greater durability, making them more suitable for large-scale games.

SSD can significantly reduce waiting times during game startup, scene loading, and resource switching, enabling them to become a sought after in the market. And this is why “Black Myth: Wukong” recommends PC setups equipped with SSD.

Recently, Chinese memory brand ZhiTai collaborated with “Black Myth: Wukong” to launch a co-branded version of the TiPlus7100 solid-state drive.

TiPlus7100 “Black Myth: Wukong” edition solid-state drive features an aesthetic design rooted in Eastern culture, which is equipped with high-quality original NAND and Yangtze Memory’s Xtacking® technology architecture.

It offers read speeds of up to 7000MB/s and comes in 1TB and 2TB capacities. The HMB mechanism and SLC Cache intelligent cache design ensure a stutter-free gaming experience. With its M.2 2280 single-sided PCB design, this co-branded edition can meet expansion needs for PS5, laptops, desktops, and more.

In addition to ZhiTai, other manufacturers like UNIC Memory, SCY, and YXSC have also showcased their products recently, which meet the configuration requirements for “Black Myth: Wukong” and offer excellent gaming experiences.

UNIC Memory’s new Chinese-inspired brand UniWhen announced that its MoYunZangJing series uses original memory and offers various frequencies ranging from 6400 to 8000 MT/s.

It supports Intel XMP 3.0 and AMD EXPO dual-platform overclocking techniques, with capacities of 16GB2 and 24GB2, matching the demands for multi-tasking and running multiple games simultaneously.

SCY revealed that its DDR5 UDIMM RGB gaming edition memory, along with the S7000, S7000 Pro, and C7000 series SSD, can provide a buff for “Black Myth: Wukong” players.

The DDR5 UDIMM RGB gaming edition offers up to 32GB (16GBx2) to avoid lag or disconnection during intense gameplay. The S7000 series offers sequential read/write speeds of up to 7400MB/s and 6600MB/s, with up to 4TB of storage.

The S7000 Pro series reaches sequential read/write speeds of 7100MB/s and 6400MB/s, with storage options from 512GB to 4TB. The C7000 series offers sequential read/write speeds of 7000MB/s and 6100MB/s, with up to 2TB of storage, designed for handheld consoles and Steam gaming devices.

At elexcon Shenzhen International Electronics Exhibition 2024, YXSC Semiconductor demonstrated the performance of its consumer-grade PCIe SSD from the F800E/F series running “Black Myth: Wukong.” The F800E/F series features 3D TLC NAND chips, with capacities ranging from 128GB to 2048GB, and adopts the M.2 standard form factor.

Read more

• [News] “Black Myth: Wukong” Sparks Trend, Boosts Demand for Memory and Graphics Cards
• [News] Intel Unveils Two New GPUs, Manufactured on TSMC’s 4nm Process while Reportedly Targeting NVIDIA RTX 40 Series

(Photo credit: Stram)

China’s GPU company Lisuan Technology, based in Shanghai, has averted the crisis of bankruptcy, as it secures around 328 million yuan (nearly USD 46 billion) in financing from domestic NAND/ DRAM manufacturer Dosilicon and others, according to a report by Chinese media outlet Sina.

On August 20^th, Dosilicon made an announcement, stating that it plans to invest 200 million yuan of its own funds to increase the stake in Lisuan Tech. By subscribing to an additional 5 million yuan of Lisuan’s newly registered capital, the memory company will hold approximately 37.88% of Lisuan’s equity.

On the other hand, the report notes that other investors plan to inject a total of 128 million yuan to Lisuan, subscribing to a total of 3.2 million yuan in its newly added registered capital. In total, Lisuan Tech has received 328 million yuan in financing from Dosilicon and others.

Regarding the reasons behind the investment, the report indicates that there is a certain level of synergy between Dosilicon and its target company, Lisuan Tech. As Dosilicon has already established a portfolio of both standard and niche DRAM products, its R&D team can further engage in technical collaboration with the graphics rendering chip design team at Lisuan to enhance the design capabilities of both parties.

The report, citing public information, states that Lisuan Tech, with 20 years of experience in GPU development and design, is one of the few domestic companies in China capable of providing customized high-performance GPU solutions.

The firm’s first 6nm GPU, based on its self-developed ‘Pangu’ architecture, is ready for tape-out, the report suggests. The product even boasts to offer performance on par with NVIDIA’s high-end graphics cards.

However, due to delays in securing financing, the company has fallen into difficulties, with rumors circulating that it was facing bankruptcy.

According to the data cited by the report, in 2023, Lisuan had no revenue and a net loss of 145 million yuan. In the first half of 2024, it reported no revenue and a net loss of 97.9 million yuan. The bulk of the losses was said to stem from R&D investments.

Read more

• [News] ASML Claims the World Counts on China for Legacy Chips; So Do Chip Equipment Giants
• [News] China Makes Progress in Chip Tool Self-Sufficiency, Yet Lithography Remains a Key Bottleneck

(Photo credit: Lisuan Tech)

Super Micro Computer, Inc. (Supermicro), a provider of servers and storage solutions, released its financial results for the fourth quarter of the fiscal year 2024 (ending June 30, 2024) on August 6. The revenue increased by 143.6% year-over-year (37.9% quarter-over-quarter) to USD 5.31 billion. The Non-GAAP diluted earnings per share (EPS) rose by 78.1% year-over-year (decreased by 6% quarter-over-quarter) to $6.25.

Supermicro forecasts that for the first quarter of the fiscal year 2025 (ending September 30, 2024), revenue will be between USD 6 billion and 7 billion (midpoint of USD 6.5 billion), and the Non-GAAP diluted EPS is expected to be between $6.69 and $8.27 (midpoint of $7.48).

Additionally, for the fiscal year 2025 (ending June 30, 2025), revenue is projected to be between USD 26 billion and 30 billion (midpoint of USD 28 billion). A report from MoneyDJ further cite sources, indicating that Supermicro’s fiscal year 2025 revenue is expected to reach USD 23.4 billion.

According to the Q4 financial report for the fiscal year 2024 released by Supermicro, the gross margin decreased from 17.0% in the same period last year to 11.2%, the lowest since the company started reporting quarterly results in May 2007 , and below the 15.5% reported in the third quarter of the fiscal year 2024.

Per wccftech’s report, the Super Micro earnings call marked the first opportunity to engage with a company integrated into NVIDIA’s AI ecosystem. Consequently, the company’s management faced numerous questions about potential GPU delays impacting its financial performance.

In response, Charles Liang, Supermicro’s Chief Executive Officer and Chairman of the Board, acknowledged that his company had heard about potential delays from NVIDIA.

However, he emphasized that Supermicro considers such delays as a normal possibility, noting that technology companies often experience slight advancements or postponements in their schedules.

In this instance, it seemed to Liang that NVIDIA had pushed out its timeline slightly, which he stated would not impact Supermicro’s ability to offer its customers new solutions like the H200 cooling system, given their extensive customer base. He concluded by saying that the overall impact of this delay on Supermicro should be minimal.

Read more

• [News] Foxconn Remains Unfazed Despite NVIDIA’s GB200 Delay, with Customers Aim to Secure Their Orders
• [News] Supermicro President Charles Liang Bullish on Direct Liquid Cooling, Targeting a 15% Market Share within Next Year

(Photo credit: Supermicro)

Due to challenges in exporting high-performance processors based on x86 and Arm architectures to China, the country is gradually adopting domestically designed operating systems.

According to industry sources cited by Tom’s hardware, Tencent Cloud recently launched the TencentOS Server V3 operating system, which supports China’s three major processors: Huawei’s Kunpeng CPUs based on Arm, Sugon’s Hygon CPUs based on x86, and Phytium’s FeiTeng CPUs based on Arm.

The operating system optimizes CPU usage, power consumption, and memory usage. To optimize the operating system and domestic processors for data centers, Tencent has collaborated with Huawei and Sugon to develop a high-performance domestic database platform.

Reportedly, TencentOS Server V3 can run GPU clusters, aiding Tencent’s AI operations. The latest version of the operating system fully supports NVIDIA GPU virtualization, enhancing processor utilization for resource-intensive services such as Optical Character Recognition (OCR). This innovative approach reduces the cost of purchasing NVIDIA products by nearly 60%.

TencentOS Server is already running on nearly 10 million machines, making it one of the most widely deployed Linux operating systems in China. Other companies, such as Huawei, have also developed their own operating systems, like OpenEuler.

Read more

• [News] Huawei’s Largest Global R&D Center in Shanghai Unveiled, Invests Over USD 1.4 Billion
• [News] Hurdles in Acquiring NVIDIA’s High-End Products: Assessing the Progress of Eight Chinese AI Chip Companies in Self-Development

(Photo credit: Tencent Cloud)