TechnoSports Media Group

Tag: NVIDIA

NVIDIA Climbs to Second Spot as TSMC’s Largest Client, Trailing Apple

TSMC traditionally does not disclose details of its relationships with customers, but under US law it is required to disclose data about customers who account for more than 10% of its revenue. Financial analyst Dan Nystedt predicts that NVIDIA is expected to contribute 11% of TSMC’s revenue in the year 2023.

NVIDIA Claimed Second Spot as Apple Client

In the year Apple, identified as “Client A”, in TSMC’s filing with the US Securities and Exchange Commission (SEC) accounted for 25% of the company’s revenue amounting to $17.52 billion. On the other hand, NVIDIA labeled as “Client B” paid TSMC $7.73 billion representing an 11% share of the company’s revenue.

According to Dan Nystedt, TSMC’s top ten customers collectively made up 91% of its revenue year an increase from 82% in 2022. These customers include MediaTek, AMD, Qualcomm, Broadcom, Sony, and Marvell.

Apple has been a client of TSMC for a period and is expected to retain that position in the future. While no other client has contributed more than 10% of TSMC’s revenue in a significant amount of time companies like AMD, MediaTek, and Qualcomm have been scaling up their orders recently. The surge in demand for NVIDIA H100 and A100 accelerators due to advancements in intelligence technologies has driven increased order volumes from them. In fact, NVIDIA secured part of TSMCs production capacity between 2021 and 2022.

NVIDIA produces products, like the H100 and A100 AI accelerators, which are made and packaged at TSMC factories using their CoWoS technology. With the rising demand, for AI equipment NVIDIAs portion of TSMC’s earnings is projected to rise in 2024. The company has secured production and packaging capabilities to maintain a supply of its products.

It is not yet known whether AMD’s share of TSMC’s revenue will exceed 10%. The company is actively selling EPYC server processors, and its Instinct MI300 AI accelerators are also in high demand, so the share of “red” ones in the Taiwanese contractor’s income may also increase.

FAQs

March 4, 2024
NVIDIA RTX 1000 & RTX 500 Ada Laptop GPUs: Democratizing AI-Ready Performance, Outpacing NPUs & CPUs
NVIDIA has introduced two new entry-level Ada GPUs, the RTX 1000 and the RTX 500 to its laptop lineup, aiming to democratize AI-readiness across user segments.

All About the NVIDIA RTX 1000 & RTX 500 GPU

The RTX 1000 and RTX 500 Ada GPUs, from NVIDIA are designed for entry-level laptops focusing on enhancing AI performance. By harnessing Ada’s hardware capabilities, these processors claim to deliver up to 14 times faster AI speed and three times AI-driven photo editing. Moreover, they boast an increase in graphics performance for tasks such as rendering and content creation compared to using CPUs.

While CPUs have incorporated AI features through NPU technologies GPUs continue to outperform them in terms of performance. NVIDIA’s introduction of these two products expands access to performance levels for a wider range of users. While NPUs excel in low-power AI operations the RTX 1000 and RTX 500 GPUs from NVIDIA provide performance enhancements over systems lacking dedicated GPUs ensuring quicker task completion. Notable features of this lineup include:
- Third-generation RT Cores that offer double the ray tracing performance for rendering.
- Fourth-generation Tensor Cores that provide twice the throughput for deep learning tasks and AI-driven creative tasks.
- Ada Generation CUDA cores that deliver up to 30% single precision floating point throughput, for graphics and computational performance.
- The RTX 500 comes with 4GB while the RTX 1000 offers 6GB enabling smooth operation of demanding 3D modeling and AI applications. Both GPUs come with core setups, CUDA cores, RT cores, and Tensor Cores.
- The RTX 1000 Ada GPU packs 6GB of VRAM whereas the RTX 500 Ada GPU carries 4GB. In terms of power consumption, the RTX 1000 Ada ranges from 35 140W while the RTX 500 Ada falls within the range of 35 60W.
In terms of performance, the RTX 1000 Ada offers 12.1 TFLOPs of FP32 and up to 193 TOPS (INT8), while the RTX 500 Ada offers 9.2 TFLOPs of FP32 and up to 154 TOPS (INT8). Both GPUs are already available in laptops from Dell, Lenovo, and MSI, offering AI-infused capabilities at reasonable price points.

FAQs
February 28, 2024
Nvidia Anticipates Supply Constraints for Next-Gen Blackwell GPUs in 2024

Shortly after an analyst reported a significant reduction in lead times for Nvidia’s Hopper-based H100 GPUs, utilized in artificial intelligence (AI) and high-performance computing (HPC), Nvidia itself announced expectations of supply constraints for its forthcoming Blackwell-based GPU products.

More About Nvidia Supply of Next-Gen GPU

During Nvidia’s call with analysts and investors Colette Kress, the company CFO disclosed that there are expected supply constraints, due to high demand for Nvidia’s upcoming Blackwell architecture, which will power the B100 products. These new products are projected to offer improved AI computing performance compared to the Hopper architecture. It is likely that Nvidia’s existing customers have already placed orders for some B100 products given the demand for high-performance AI processors in the market.

The main focus is on how Nvidia can ramp up production of B100 SXM modules B100 PCIe cards and DGX servers based on the architecture. These products utilize components. Represent a significant shift. Rumors are suggesting that Blackwell could be Nvidia’s venture into chiplet designs. If this turns out to be true it could simplify the production of GPUs at the silicon level by improving chip yields. However, incorporating chiplet solutions also introduces additional complexities.

Nvidia is getting ready to launch a B40 GPU for business and training purposes alongside the GB200 product that pairs the B100 GPU with an Arm-based Grace CPU. They are also working on the GB200 NVL designed specifically for training language models. In November Nvidia unveiled the H200 compute GPU suited for AI and HPC tasks, which is now, in the production phase.

Based on the Hopper architecture, the H200 essentially refreshes Nvidia’s existing product lineup with increased memory capacity and bandwidth. Given Nvidia’s refined Hopper supply chain since 2022, the product ramp-up should progress relatively swiftly. However, Nvidia’s CEO, Jensen Huang, acknowledges that meeting 100% demand for this new product immediately is not feasible.

Huang emphasized during the quarterly earnings call with analysts and investors that the transition from zero to significant production levels cannot occur overnight, especially with new product generations like the H200.

FAQs

February 24, 2024
NVIDIA’s Q4 Report: 769% Annual Profit Growth Sparks 10% Share Surge

Before its anticipated earnings report, which was expected to have an impact, on the tech stock market and AI sector investors approached NVIDIA Corporation cautiously. This caution was reflected in drops in the company’s share prices. However, NVIDIA’s fourth-quarter earnings report again confirmed its leading position in the field of intelligence.

The company reported $22 billion in revenue and $5.16 in diluted GAAP earnings per share surpassing analysts’ predictions of $20.6 billion in revenue and $4.64 in EPS. This outstanding performance exceeded optimistic expectations driving NVIDIA’s stock to achieve double-digit percentage gains this year.

More About NVIDIA’s Profit Growth

The latest quarterly results underscore NVIDIA’s growth trajectory which has been consistent since the beginning of the year. Notably, the company saw a 769% increase in bottom line profit generating $12.3 billion in GAAP income during the period – marking its third consecutive quarter of record-breaking profit and revenue. The main catalyst for this growth was NVIDIA’s Data Center division, which reported $18.4 billion in revenue – a 409% growth compared to the previous year.

While aftermarket trading showed enthusiasm for NVIDIA’s performance there were results, from its Gaming division. Despite a 56% increase in Gaming sales the sequential numbers stayed flat. NVIDIA credited the boost in Gaming revenue to sales to partners after inventory adjustments and growing demand.

The company’s continued triumph in the Data Center sector driven by the use of its Hopper GPU in diverse AI applications like language models and recommendation systems played a role, in its success. Moreover, NVIDIA’s upbeat guidance for the next quarter, projecting revenue potential of up to $24 billion, further bolstered investor confidence.

In response to the impressive earnings report and outlook, NVIDIA’s shares surged by 10% in aftermarket trading, reflecting Wall Street’s positive reception of the results and future prospects.

FAQs

February 22, 2024
NVIDIA App Beta: The Ultimate Companion for Gamers and Creators

NVIDIA is known for its continuous innovation in enhancing gaming and content creation experiences. This time, it’s leaping forward with the launch of the latest NVIDIA Game Ready Driver, coinciding with the beta release of their new NVIDIA app. Designed to be the essential companion for gamers and creators equipped with NVIDIA GPUs on their PCs and laptops, this app aims to modernize and unify the NVIDIA Control Panel and GeForce Experience.

‘NVIDIA App’ Beta Kicks-off with a New Game Ready Driver that Optimizes ‘Nightingale’ with DLSS 3 and Reflex

The newly launched Game Ready Driver not only supports Nightingale, the exciting new PvE open-world survival crafting game with DLSS 3 and Reflex, but also optimizes settings for four more games – Granblue Fantasy: Relink, Pacific Drive, and Skull and Bones.

The NVIDIA app is designed to simplify the process of keeping your PC updated with the latest NVIDIA drivers. It also enables quick discovery and installation of NVIDIA applications like GeForce NOW, NVIDIA Broadcast, and NVIDIA Omniverse.

Featuring a unified GPU control center, the NVIDIA app allows for fine-tuning of game and driver settings from a single place. It introduces a redesigned in-game overlay for convenient access to powerful gameplay recording tools, performance monitoring overlays, and game-enhancing filters, including innovative new AI-powered filters for GeForce RTX users.

This initial beta release of the NVIDIA app integrates top features from existing apps, optimizes the user experience, includes an optional login for redeeming bundles and rewards, and introduces new RTX capabilities to elevate your gaming and creative experiences.

The NVIDIA app also showcases new AI-powered Freestyle Filters. RTX HDR seamlessly adds High Dynamic Range (HDR) to Standard Dynamic Range (SDR) games on HDR displays, while RTX Dynamic Vibrance enhances the NVIDIA Control Panel Digital Vibrance feature, further improving visual clarity in games.

GeForce gamers can download the NVapp beta here.

February 22, 2024
Raja Koduri Highlights the Importance of PC GPUs from NVIDIA, AMD, and Intel in AI and Data Center Success

Raja Koduri, an executive at Intel and the founder of Mahira AI, brings an interesting perspective to the conversation about how AI intersects with data centers. He emphasizes the importance of having GPUs specifically designed for PCs to drive success in today’s ecosystem.

More About Raja Koduri’s Statement about PC GPUs

Koduri mentions the popularity of gaming GPUs among consumers highlighting their affordability and accessibility compared to workstation versions. These GPUs are well-received by users and play a role, in providing easy access to AMD Radeon or NVIDIA GeForce GPUs for developers worldwide.

However, Koduri raises concerns about the structure of GPU offerings potentially limiting accessibility and adoption among developers. He suggests that tech giants like AMD and Intel may need to reconsider their strategies for consumer GPUs since these tools are essential for PC developers. Koduri believes that platforms such as AMD ROCm and Intels SYCL stacks might be overshadowing PC-oriented GPUs causing developers to overlook opportunities.

He also notes that NVIDIA and AMD hold a position in this area compared to Intel, which ironically hinders the developer community’s willingness to embrace Intels consumer-grade GPUs. Developers typically seek a balance between high-quality gaming performance and advanced AI capabilities so they may need to reconsider their preference, for consumer-grade GPUs if manufacturers do not address these concerns.

Raja Koduri attributes this situation to the dominance of the AI ecosystem, where GPU manufacturers prioritize AI accelerators over an audience. ZLUDA provides a way to use NVIDIAs CUDA libraries, on the ROCm stack. The performance on GPUs may vary in today’s setups creating potential hurdles, for developers.

Despite recent developments such as NVIDIA’s TensorRT-LLM support and AMD’s ROCm support for certain Radeon GPUs, Raja Koduri suggests that manufacturers must rethink the software ecosystem’s progression to meet the evolving needs of developers in the AI PC era. While these advancements may not concern the average gamer, they are crucial for developers navigating the modern landscape of AI capabilities and software stacks.

FAQs

February 21, 2024
RTX 4070 Ti Super vs RX 7900 XT: Which Graphics Card Should You Choose?

In the evolving world of graphics cards, in 2024 two unexpected rivals have emerged; the NVIDIA RTX 4070 Ti Super vs RX 7900 XT. These GPUs have caught attention due to their pricing. These GPUs target an audience of gamers looking for top-notch 4K gaming experiences without breaking the bank.

RTX 4070 Ti Super vs RX 7900 XT

Choosing between the NVIDIA RTX 4070 Ti Super and AMD RX 7900 XT is no task as both cards offer features and capabilities. To determine which one is the option in 2024 thorough evaluations were conducted to analyze their strengths and weaknesses.

The RTX 4070 Ti Super, set to be released on January 24 2024 with a price tag of $800 faces competition, from the RX 7900 XT. Since its launch in December 2022, the RX 7900 XT has seen price reductions. It is currently available for as low as $730. When it comes to specifications both GPUs have differences in terms of memory capacity, power consumption levels, and connectivity options.

The RX 7900 XT boasts a memory capacity of 20GB. It supports DisplayPort 2.1 which brings advantages in usage scenarios. In terms of power usage, the RTX 4070 Ti Super requires around 285 watts while the RX 7900 XT consumes up to 300 watts.

Both these GPUs perform excellently in gaming at resolutions of up to 4K with over 60 frames per second (fps) achieved. While the RTX model may offer ray tracing capabilities the RX 7900 XT generally outperforms it by a margin. Notably due to its price point the RX 7900 XT provides value for budget-conscious gamers.

Ultimately deciding between the RTX 4070 Ti Super and RX 7900 XT depends on preferences and priorities. While most users will find performance and value, with the RX model overall enthusiasts seeking premium gaming experiences may still opt for the RTX model.

The RTX 4070 Ti Super and RX 7900 XT are both contenders, in the graphics card market of 2024 despite their costs. They come with features and performance capabilities that cater to the demands of gamers.

February 17, 2024
NVIDIA GeForce RTX 4070 Ti and RTX 4070 GPUs Receive Price Drops in the US

NVIDIA has recently introduced price promotions for the GeForce RTX 4070 Ti and RTX 4070 Non SUPER GPUs, which might attract gamers looking for graphics cards in this range.

All About NVIDIA GeForce RTX 4070 Ti and RTX 4070 Price Drop

A weeks ago the NVIDIA GeForce RTX 4070 Ti GPU was replaced by the GeForce RTX 4070 Ti SUPER. Both GPUs were priced at the manufacturer’s suggested price (MSRP). The original RTX 4070 Ti faced criticism, for its performance-to-price ratio offering 12 GB of VRAM at a price of $799 USD. On the other hand, the SUPER variant came with improved specs, including 16 GB of VRAM making it an appealing option and reportedly achieving strong sales in the gaming market.

Though the GeForce RTX 4070 Ti has been replaced there are still models available in the market. In response to this NVIDIA board partners have initiated price reductions and promotions to attract gamers. The first wave of these discounts is now visible with the RTX 4070 Ti Non-SUPER being sold for $699 US on Newegg.

Among the models benefiting from these price reductions are the MSI Ventus 2X variants available, in both white versions. Both models come with specifications and a fan cooling system. The standard OC variant is priced at $699 US while the White OC variant is priced higher at $729 US. This means that the price has been reduced by $100 compared to the manufacturer’s suggested price (MSRP) of $799. We made this decision strategically considering the challenges of maintaining both card’s prices after the release of the SUPER variant.

When it comes to value proposition AMD remains ahead with its Radeon RX 7900 XT GPUs also priced at $699 US. They now offer 20 GB of VRAM and superior performance. Despite NVIDIA’s advantages, in features and efficiency, most people still find the 7900 XT a choice. If you’re looking for a slight performance upgrade the RTX 4070 Ti SUPER is an option even though it comes with a $100 USD premium.

On another note, the NVIDIA GeForce RTX 4070 Non SUPER has also received a price reduction. It is now available at $519 US for the GALAX reference model on Amazon. This model features a fan and dual-slot cooler configuration. Uses a standard 8-pin power connector.

Although the RTX 4070 Non SUPER models haven’t been replaced by the RTX 4070 SUPER they are now priced around $549 US to reflect their MSRP. However, NVIDIA has indicated that gamers can expect to find these cards for less, than that in some cases. There is a possibility that prices might drop to, than $500 USD in the months.

February 10, 2024
NVIDIA Unveils Budget-Friendly GeForce RTX 3050 6 GB GPU at $169

NVIDIA has recently unveiled its budget option, the GeForce RTX 3050 6 GB GPU, designed specifically for gamers on a budget who are looking for a great 1080p gaming experience. This GPU comes at a price of $169 US.

The All New NVIDIA GeForce RTX 3050 GPU

This launch introduces the version of the GeForce RTX 3050 GPU to the market, with the 6 GB model aimed at meeting the needs of budget-conscious PC gamers. The main goal, behind introducing this model is to make RTX gaming accessible to an audience by providing features like RTX Ray Tracing and AI computing at a price. This is especially important considering that many gamers still rely on GPUs like GTX 1650 and GTX 1050 as shown by Steams Hardware Survey.

Compared to the existing NVIDIA GeForce RTX 3050 8 GB GPU for $250 $280 US on Newegg the price point of $100 US for the RTX 3050 6 GB GPU is expected to attract a significant number of gamers who have been using older GPUs. NVIDIA claims that this new model offers four times the performance compared to its predecessors while also supporting features, like DLSS Super Resolution.

The RTX 3050 6 GB has an advantage in terms of its setup. You can easily power it up using the PCIe slot without the need for any connectors, making it hassle-free. Known brands like MSI and ASUS have already listed models on Newegg.

In terms of specifications, the upcoming NVIDIA GeForce RTX 3050 GPU will be based on the GA107 325 Kx SKU. Will use the PG173 SKU16 PCB. It will have 2048 cores, which is a 20% reduction compared to the 8 GB model. Its core clock will remain at a steady 1470 MHz. In terms of memory, it will come with 6 GB of VRAM and a 96-bit bus interface, compared to the model 8 GB of VRAM and 128-bit bus.

Despite these reduced specs, certain designs of the RTX 3050 with 6 GB won’t need power input, resulting in lower overall power consumption. The card will still support quad-display output with three DP ports and one HDMI port. The official price for the NVIDIA GeForce RTX 3050 GPU with 6 GB of VRAM is set at $169 US dollars, while OC (overclocked) models may be priced higher.

If you’re already familiar with the NVIDIA GeForce ecosystem on your PCs, the card could be a choice for an upgrade. However, it’s worth considering options like the Intel Arc A580 and the AMD RX 6600 8 GB, which offer similar performance at similar price points. For gamers on a budget, the 6 GB of memory might be sufficient for 1080p gaming. If you’re just getting started with entry-level 1080p PC gaming, an 8 GB card would be a good choice.

February 3, 2024
Samsung’s Upcoming GDDR7 Memory: A Leap Forward with 37 Gbps Pin Speeds, Surpassing GDDR6X by 54%
Samsung is gearing up to introduce its latest and fastest GDDR7 memory modules next month, showcasing speeds of up to 37 Gbps tailored for next-generation GPUs. TechRadar reports that Samsung will unveil its next-gen GDDR7 memory at the 2024 IEEE International Solid-State Circuit Conference in San Francisco, during a session titled “A 16Gb 37 Gb/s GDDR7 DRAM with PAM3-Optimized TRX Equalization and ZQ Calibration.”

This session will serve as the platform for Samsung to reveal its cutting-edge memory solutions designed for the upcoming generation of GPUs.

More About Samsung GDDR7 Memory

The reported 37 Gbps DRAM speed represents a significant leap, boasting a 50%+ improvement over the current GDDR6X DRAM, which tops out at 24 Gbps. While Samsung initially planned to introduce GDDR6W with doubled capacities and 64-bit DRAM, the company is now shifting its focus towards the true next-gen standard. Although Samsung already offers GDDR6 memory with speeds of up to 24 Gbps, recent graphics card releases, such as NVIDIA’s RTX 4080 SUPER, have only reached speeds of 23.5 Gbps.

In a recent announcement, Samsung revealed that they had internally achieved speeds of up to 36 Gbps with GDDR7 memory. This indicates the company’s continuous efforts to maximize the potential of next-gen memory modules.

While it remains uncertain if these higher speeds will be readily available in sufficient quantities to meet the demand for next-gen gaming and AI GPU lineups, it is anticipated that speeds ranging from 32 to 36 Gbps will be prominent in the upcoming GPU generation. The bandwidth offered by the 37 Gbps pin speeds across various bus configurations is as follows:
- 512-bit: 2368 GB/s (2.3 TB/s)
- 384-bit: 1776 GB/s (1.7 TB/s)
- 320-bit: 1480 GB/s (1.5 TB/s)
- 256-bit: 1184 GB/s (1.2 TB/s)
- 192-bit: 888 GB/s
- 128-bit: 592 GB/s
Additionally, GDDR7 memory is expected to deliver a 20% increase in efficiency, a noteworthy improvement considering the substantial power consumption of high-end GPUs. Samsung’s GDDR7 DRAM will incorporate technology optimized for high-speed workloads, with a low-operating voltage option designed for power-conscious applications like laptops.

Addressing thermal concerns, the new memory standard will utilize an epoxy molding compound (EMC) with high thermal conductivity, reducing thermal resistance by up to 70%. Reports from August indicated that Samsung had provided samples of its GDDR7 DRAM to NVIDIA for early evaluation in the development of its next-gen gaming graphics cards.
January 30, 2024