Hardware arşivleri - Lavega Games

Blackwell B200, the World’s Most Powerful AI Chip

Onur Bahadır — Tue, 19 Mar 2024 09:13:01 +0000

Nvidia has officially unveiled its next-generation Blackwell GPU architecture, which offers up to 5 times the performance increase compared to the Hopper H100 GPUs that propelled the company to trillion-dollar status and the position of the world’s third-most valuable company. As competitors strive to catch up with Nvidia, the company has also announced the Blackwell-based new products, the Blackwell B200 GPU and the GB200 Superchip. Here’s everything you need to know:

Nvidia’s Blackwell Architecture and the New Blackwell B200:

Nvidia detailed its new generation AI & Tensor Core GPU architecture codenamed Blackwell at a recent event. As anticipated, Blackwell GPUs represent Nvidia’s first Multi-Chip Module (MCM) design, combining two GPUs on the same package.

The Blackwell architecture, produced on the TSMC 4NP process, boasts a total of 104 billion transistors per compute die. The B100 GPUs feature a high-bandwidth interface of 10 TB/s for super-fast chip-to-chip communication. By combining multiple GPUs on the same package, a total of 208 billion transistors are accessed, offering complete GPU cache coherence.

Compared to Hopper, Nvidia’s Blackwell GPU offers 128 billion more transistors, 5 times the AI performance per chip at 20 petaFlops, and 4 times the embedded memory. The Blackwell GPU itself is equipped with an 8 TB/s memory bandwidth interface and up to 192 GB of 8-layer HBM3e memory.

Nvidia claims that the GB200, utilizing a dual GPU and single CPU solution, offers 30 times more performance for LLM inference workloads and reduces costs and energy consumption by up to 25 times compared to the H100. Nvidia has clearly made a significant leap from one generation to the next. For instance, training a 1.8 trillion-parameter AI model previously required 8,000 Hopper GPUs and 15 megawatts of power. Today, the same task can be accomplished with 2,000 Blackwell GPUs and just 4 megawatts of power.

An important advancement is Nvidia’s second-generation Transformer Engine, which doubles computation, bandwidth, and model size by utilizing four bits per neuron instead of eight. Another significant improvement is evident when multiple GPUs are interconnected: the new 5th generation NVLink enables communication between 576 GPUs at 1.8 terabytes per second bidirectional bandwidth.

In terms of product offerings, Nvidia is introducing the GB200 NVL72 package, which combines 36 CPUs and 72 GPUs on a single liquid-cooled rack, providing a total of 720 petaflops of AI training performance or 1,440 petaflops of inference. The company plans to offer NVL72 to cloud service subsidiaries of Amazon, Google, Microsoft, and Oracle. Additionally, Nvidia will provide DGX Superpod for DGX GB200, which combines eight systems to deliver a total of 288 CPUs, 576 GPUs, 240 TB of memory, and 11.5 exaflops of FP4 computation for HPC applications. Nvidia also announced the HGX B200 platform for HPCs, which includes Blackwell Tensor Core GPUs and provides up to 15 times more inference performance compared to the previous generation.

In summary, Nvidia has announced two new solutions: GB200, a combination of Grace CPU and Blackwell B100 GPU, and Blackwell B200 GPUs. These solutions are expected to be released in 2024. Each GB200 Grace Blackwell Superchip is supported by two B200 AI GPUs and a single Grace CPU with 72 ARM Neoverse V2 cores. The platform will offer 40 petaflops of computing performance (INT8) and a massive 864 GB memory pool with 16 TB/s memory bandwidth. In terms of performance, the GB200 Grace Blackwell Superchip platform will offer a 30-fold increase in AI performance compared to the H200 Grace Hopper platform.

Nvidia has not shared specific release dates and pricing information yet. However, the H100 GPU was priced between $25,000 and $40,000 per chip, and the cost of a system based on these chips could reach up to $200,000. Demand for these systems was so high that wait times remained around 11 months for a considerable period. Demand for the Blackwells also appears to be high already.

Stay plugged into the gaming pulse, and anticipate the thrill of new adventures and the nostalgia of departing classics. For the freshest updates, keep an eye on Lavegagames.c om – where gaming news comes to life.

Blackwell B200, the World’s Most Powerful AI Chip yazısı ilk önce Lavega Games üzerinde ortaya çıktı.

World’s First Error-Tolerant Quantum Computer Set to Launch

Onur Bahadır — Tue, 13 Feb 2024 08:11:01 +0000

The world’s first commercial error-tolerant quantum computer with “logical qubits” could start operating before the end of the year. Logical qubits, which connect physical quantum bits or qubits through quantum entanglement, reduce errors in quantum computers by storing the same data in different locations.

Representatives from QuEra, the company building the machine, announced that the new device, with 256 physical and 10 logical qubits, will be released to the market by the end of 2024.

Harvard, QuEra, and researchers from other institutions previously showcased a quantum computer with the highest number of logical qubits tested to date, featuring 48 logical qubits. The recent announcement followed this achievement. Harry Zhou, a co-author of the study and a physicist at QuEra and Harvard University, stated, “This is the first machine with quantum error correction.”

Zhou mentioned that while this computer may not have enough power to be useful on its own, it provides a platform for software programmers to start testing code for future quantum computers.

Why is Error Correction Needed?

Traditional computers store information in bits with values of 0 or 1, whereas a qubit’s general state, according to quantum mechanics, can exist in a quantum superposition of both 0 and 1. Qubits can also be brought together in multiple states simultaneously using quantum entanglement, allowing them to perform many calculations much faster than classical computers. However, qubits are easily prone to errors, making them error-prone.

While quantum computers with millions of qubits could outperform the best supercomputers, natural errors limit scalability. Moreover, the largest quantum computer built to date has only about 1,000 qubits. This is where error correction comes into play, with logical qubits being one of the methods used in quantum error correction.

Logical Qubits

The new error correction system relies on data redundancy, where the same data is stored in multiple locations. Logical qubits perform the same calculations on multiple physical qubits, significantly reducing error rates if one or more physical qubits fail, allowing calculations to continue.

These new qubits have made significant progress compared to their predecessors. Google made a groundbreaking step in quantum computers last year by introducing logical qubits. In their research, Google demonstrated a system with 49 physical qubits that achieved an error rate of 2.9%. QuEra’s error rate, with 48 logical qubits, has now reached 0.5%.

IBM also demonstrated error correction technology in its 127-qubit Heron chip last year, reducing error rates five times compared to other chips. However, the production of the first commercial error-tolerant machine is not expected until 2029.

In the coming years, QuEra plans to release several quantum computers starting with a machine featuring 30 logical qubits and 3,000 physical qubits in 2025. The company also states that a machine with over 10,000 physical qubits and 100 logical qubits will be launched in 2026, capable of accurate calculations surpassing the capacities of today’s supercomputers.

World’s First Error-Tolerant Quantum Computer Set to Launch yazısı ilk önce Lavega Games üzerinde ortaya çıktı.

AMD Fluid Motion Frames Released: Here Are the Supported Games

Onur Bahadır — Mon, 02 Oct 2023 15:06:00 +0000

AMD had previously announced that its FidelityFX Super Resolution (FSR) 3.0 technology, introduced last year, would provide frame interpolation similar to DLSS 3 technology. With preparations now complete, the red team has announced that Fluid Motion Frames (AFMF) technology is accessible in more than 12 games.

The wait for FSR 3.0 is finally over, with the technology now available in two games: Forspoken and Immortals of Aveum. However, the most significant development is the availability of AMD Fluid Motion Frames technology. In this regard, AFMF, much like Nvidia’s Frame Generation technology, creates a new frame by looking at the images of previous frames.

AMD Fluid Motion Frames

Looking at the list of AFMF-supported games, it’s evident that the list is quite extensive. Among them, you’ll find many AAA titles such as Starfield, Star Wars Jedi: Survivor, and The Last of Us Part 1. It’s worth noting that these games currently do not have FSR 3.0 support. However, since Fluid Motion Frames is integrated exclusively with Radeon graphics drivers, third-party hardware has been bypassed.

AFMF is only compatible with the Radeon RX 7000 series (Radeon RX 7900/7800/7700/7600 series). Therefore, support is not provided for older generation graphics cards. However, it’s worth mentioning that Phoenix APUs, commonly used in handheld consoles, are included in this list. Here is the list of AFMF-supported games:

A Plague Tale – Requiem
Deep Rock Galactic
Hitman 3
Red Dead Redemption 2
Star Wars Jedi: Survivor
Borderlands 3
Dying Light 2
Hogwarts Legacy
Resident Evil 3
Starfield
Control
Far Cry 6
Horizon Zero Dawn
Resident Evil 4
The Last of Us Part 1
Dead Space
Ghostwire: Tokyo
Metro Exodus Enhanced Edition
Shadow Of The Tomb Raider
The Witcher 3: Wild Hunt

Final Thoughts

Hello #LavegaCrew,

At Lavega Games, we’re passionate about gaming and technology, and we’re excited to share the latest news with you. We understand that a high-quality gaming experience is important, which is why we strive to provide valuable insights and recommendations based on our extensive industry knowledge and connections. Don’t forget to connect with us on our social media channels where you can stay up to date on our latest projects and share your feedback with us. (Instagram, Reddit, Twitter, and GitHub).

Thank you for being a part of our community!

If you are interested in NFTs and Meta world, don’t forget to check out our LavegaPro website.

AMD Fluid Motion Frames Released: Here Are the Supported Games yazısı ilk önce Lavega Games üzerinde ortaya çıktı.

First performance benchmarks shared for RTX 4060: Here are the results

Onur Bahadır — Thu, 22 Jun 2023 13:18:25 +0000

Excitement continues to build as users eagerly anticipate the arrival of the GeForce RTX 4060, which will be available very soon. In this context, we are hearing new leaks and rumors. Most recently, as the launch of the card approaches, performance tests have started to surface on the internet. So, what does the mid-range model of the Ada Lovelace RTX 4000 series, the RTX 4060, have to offer?

RTX 4060 spotted in performance tests
As a reminder, the GeForce RTX 4060 was introduced with 3072 CUDA cores, a 1.83GHz base clock speed, a 2.46GHz boost clock speed, 8GB of memory, a 128-bit memory interface, and a 110W TDP. During the launch, Nvidia stated that the card is respectively 1.2 times and 1.6 times faster than the RTX 3060 and RTX 2060 when DLSS is turned off, and when DLSS is enabled, this difference increases to 1.7 times and 2.3 times faster, respectively.

RTX 4060

Looking at the results obtained from Geekbench, we see that the RTX 4060 scored 105,630 points in Geekbench’s OpenCL test and 99,419 points in the Vulkan test. Based on these results, we can say that the RTX 4060 is approximately 17% to 18% faster than the RTX 3060. When compared to AMD, it achieves a higher score than the Radeon RX 7600 by approximately 4% (Vulkan) to 32% (OpenCL). Although this may seem like a modest result, it is worth noting that the transition from the RTX 2060 to the RTX 3060 resulted in a 10% to 15% increase in performance. However, it is important to remember that these results are based on Geekbench, and there may be differences between benchmark scores and real-world performance.

Final Thoughts

Hello #LavegaCrew,

Thank you for being a part of our community!

If you are interested in NFTs and Meta world, don’t forget to check out our LavegaPro website

First performance benchmarks shared for RTX 4060: Here are the results yazısı ilk önce Lavega Games üzerinde ortaya çıktı.

Intel pulls the plug on 11th Gen ‘Tiger Lake’ processors: Here’s the list

Onur Bahadır — Thu, 08 Jun 2023 12:44:58 +0000

Intel is preparing to phase out several processor models that are part of the 11th Gen Tiger Lake family. The company has announced the processors that will be discontinued.

Intel, which recently released 13th Gen Raptor Lake architecture processors, has started to gradually withdraw support for older generation architectures. In this context, the chip giant previously announced the discontinuation of certain 11th Gen Rocket Lake processors, and now it is preparing to discontinue more models.

intel 11th gen

For those unfamiliar, Intel initiated the End of Life (EOL) process for desktop 11th Gen Core “Rocket Lake” and Tiger Lake-H series CPUs back in April. Among the first processors included in the list were five mobile 11th Gen Core-H models, including the popular gaming laptop CPU Core i7-11800H, and four 11th Gen Core-B desktop models.

Looking at the recently updated list, we can see that over 40 processors used in Mini PCs or motherboards with embedded CPUs have been added. Additionally, Intel will also halt the production of RM590E, HM570E, and QM580E chipsets for these processors.

Based on this information, we can say that the Tiger Lake series will join the previously discontinued Rocket Lake desktop CPUs in recent months. This indicates the ultimate end of the 11th Gen Intel Core family. Indeed, it is possible that other models will be added to the list in the coming months.

Final Thoughts

Hello #LavegaCrew,

Thank you for being a part of our community!

If you are interested in NFTs and Meta world, don’t forget to check out our LavegaPro website

Intel pulls the plug on 11th Gen ‘Tiger Lake’ processors: Here’s the list yazısı ilk önce Lavega Games üzerinde ortaya çıktı.

Increased Demand for NVIDIA AI GPUs Leads to 40% Rise in Chip Prices and Potential Stock Shortages

Onur Bahadır — Tue, 23 May 2023 09:28:18 +0000

NVIDIA’s strategic focus on artificial intelligence (AI) has led to a thriving GPU business. The company’s investment in technologies like ChatGPT and other generative AI models has paid off, as it is now receiving orders from various sectors of the tech industry for its cutting-edge GPUs, including the A100 and H100 models.

However, recent reports indicate that NVIDIA may struggle to meet the overwhelming demand for AI chips, despite its best efforts. According to a recent report from DigiTimes, as orders for NVIDIA’s latest GPUs continue to surge, the company has ramped up chip production at TSMC (Taiwan Semiconductor Manufacturing Company).

NVIDIA GPU

NVIDIA’s A100 GPU utilizes the TSMC 7nm process node, while the H100, which began shipping in May, is based on the TSMC 4N process node—an optimized version of the 5nm node specifically designed for NVIDIA. Since their release, both the H100 and DGX H100 GPU-powered servers have experienced tremendous demand. Despite NVIDIA’s attempts to keep up, the recent AI boom has placed significant strain on the company’s ability to meet the demand.

In China, where NVIDIA is offering modified versions of the A100 and H100 GPUs to circumvent US sanctions (known as A800 and H800), these chips are being sold at prices 40% higher than the original manufacturer’s suggested retail price (MSRP). The demand for AI GPUs in China is so high that even the V100, a GPU released in 2018 featuring the first-ever Tensor Core architecture, is priced at around 10,000 USD or 69,000 RMB. The NVIDIA A800 is priced at approximately 252,000 RMB or 36,500 USD.

Delivery times have also been affected, as NVIDIA prioritizes fulfilling orders for its non-Chinese tech partners. Consequently, the delivery time in China, which was originally around 3 months for NVIDIA GPUs, has now been extended to three months or even longer in some cases. It is reported that certain new orders may not be fulfilled until December, resulting in a waiting time of over 6 months.

Final Thoughts

Hello #LavegaCrew,

Thank you for being a part of our community!

If you are interested in NFTs and Meta world, don’t forget to check out our LavegaPro website

Increased Demand for NVIDIA AI GPUs Leads to 40% Rise in Chip Prices and Potential Stock Shortages yazısı ilk önce Lavega Games üzerinde ortaya çıktı.

AMD RX 7900 XTX Matches RTX 4090 with 700W Power Consumption

Onur Bahadır — Wed, 17 May 2023 12:20:45 +0000

A Reddit user known as Jedi95 decided to push the limits of the RX 7900 XTX, one of the top GPUs, by removing the power limit to see how fast it could go. With a high-end custom water-cooling setup and low coolant temperatures of 10 degrees Celsius, the modified RX 7900 XTX achieved remarkable performance, reaching RTX 4090 levels in 3DMark with an impressive clock speed of 3.33GHz. However, this achievement came at a cost, as the card had to consume nearly 700W to reach such high performance.

The performance achieved by this modified RX 7900 XTX demonstrates its capabilities in new territory. In 3DMark Time Spy Extreme, the GPU scored an impressive 18,335 points, surpassing the results of some RTX 4090 cards. For comparison, PCGamer’s RTX 4090 review recorded a GPU index score of 16,654 points, making Jedi95’s RX 7900 XTX benchmark result 10% faster.

AMD RX 7900 XTX Matches RTX 4090

This indicates that AMD’s RX 7900 XTX can compete with Nvidia’s flagship RTX 4090, surpassing the expected performance of an RTX 4080, albeit with higher power consumption. It’s important to note that 3DMark may not fully reflect real-world gaming performance.

Power throttling is a common occurrence on AMD’s RX 7900-series cards, as most games tend to reach the maximum power rating, causing the GPU to throttle or not reach peak clock speeds. This differs from Nvidia’s RTX 40-series GPUs, which often operate below their maximum power consumption figures during gaming.

Jedi95’s modified RX 7900 XTX also matched the best benchmark results of the stock RX 7900 XTX in the 3DMark browser, competing with GPUs that were likely cooled with liquid nitrogen.

The modification employed by Jedi95 was a “shunt-like” mod that tricked the VRM controller into reporting false power consumption figures to the GPU, bypassing AMD’s power limitations. This allowed the GPU to operate without any power restrictions, as it remained unaware of its actual power consumption. Jedi95 mentioned that this mod was superior to traditional shunt mods since power reporting could be controlled through software.

It’s not surprising that pushing voltage and frequency limits, along with improved cooling, can result in stable performance. However, power consumption scales with frequency and voltage squared, leading to extreme power usage, as demonstrated in this case.

This experiment provides insight into the capabilities of AMD’s RDNA 3 GPUs when power limits are disregarded. Operating without constraints, Navi 31 can achieve extremely high clock speeds well above 3.3 GHz, yielding significantly improved performance. Based on Jedi95’s findings, AMD may have sufficient headroom to potentially develop a theoretical RX 7950 XTX that could compete with the RTX 4090. However, accomplishing this would require a substantial increase in power draw by AMD.

Final Thoughts

Hello #LavegaCrew,

Thank you for being a part of our community!

If you are interested in NFTs and Meta world, don’t forget to check out our LavegaPro website

AMD RX 7900 XTX Matches RTX 4090 with 700W Power Consumption yazısı ilk önce Lavega Games üzerinde ortaya çıktı.

End of the Road! Windows 10 Support is Coming to an End

Onur Bahadır — Fri, 28 Apr 2023 12:43:19 +0000

Microsoft’s Windows 10 operating system has been a game-changer since its release in 2015, quickly gaining popularity and winning over users’ hearts. However, like all versions, it’s time to bid farewell to Windows 10 as well.

Windows 10

According to Microsoft’s recent announcement, the latest update, Windows 10 22H2, will be the final update for this version. While we may not see any more new features or performance enhancements for this popular operating system, security updates will continue for a while.

For Home, Pro, Education, and Enterprise versions, security updates will be available until October 14, 2025. This means that if any security-related issues or vulnerabilities arise in the operating system, we will still receive updates. After the designated date, however, there will be no more updates.

Limited support will continue for LTSC versions used by businesses and organizations if any security issues arise. Furthermore, Windows 11 LTSC upgrade will be available for LTSC versions in the second half of 2024.

As Microsoft continues to work on Windows 11 and Windows 12, it’s no surprise that support for Windows 10 is coming to an end, marking the end of an era. So, let’s bid adieu to Windows 10 and welcome new beginnings with Windows 11.

Final Thoughts

Hello #LavegaCrew,

Thank you for being a part of our community!

If you are interested in NFTs and Meta world, don’t forget to check out our LavegaPro website

End of the Road! Windows 10 Support is Coming to an End yazısı ilk önce Lavega Games üzerinde ortaya çıktı.

Nvidia Launches RTX Video Super Resolution

Onur Bahadır — Thu, 02 Mar 2023 11:03:43 +0000

Nvidia recently introduced its RTX Video Super Resolution (RVSR) technology, which enhances the resolution of browser-based videos. And now, at the end of February, Nvidia has released the driver update (531.18 GeForce Game Ready) that activates RVSR.

Nvidia System

RVSR is currently available in the Video > Video image settings > RTX video enhancement section of Nvidia drivers. For supported browsers like Chrome or Edge, RVSR offers four different quality settings ranging from 1-4. Note that the highest quality setting requires more GPU resources.

Nvidia RTX VSR

With support for video resolutions between 360p and 1440p, and refresh rates up to 144Hz, RVSR utilizes the Tensor cores found in RTX cards to provide a significant improvement in the image quality of old and blurry videos.

Nvidia has announced that RTX Video Super Resolution is already available for the RTX 30 and RTX 40 series graphics cards, and that it will support the GeForce RTX 20 series in the future. However, Nvidia has not shared a specific release date yet and mentioned that the technology needs to be restructured. As a result, users may need to wait a little longer for this feature to become available.

For more detailed info, you can check this link.

Final Thoughts

Hello #LavegaCrew,

Thank you for being a part of our community!

If you are interested in NFTs and Meta world, don’t forget to check out our LavegaPro website

Nvidia Launches RTX Video Super Resolution yazısı ilk önce Lavega Games üzerinde ortaya çıktı.

Tencent Ends VR Hardware Plans as Metaverse Competition Falters

Onur Bahadır — Wed, 22 Feb 2023 08:40:09 +0000

As the popularity of the metaverse has risen, many companies have introduced various products that enable users to connect to the social network. However, the declining interest in the metaverse has led to major losses for leading companies, including Facebook. Now, Tencent, a Chinese tech giant, is putting its VR hardware plans on hold.

Tencent VR

For those who don’t know, Tencent established a special XR division for VR-based products last year, which was expected to feature VR hardware, applications, and games. However, things don’t seem to have gone according to plan.

According to a recent report by Routers, Tencent “encountered challenges in quickly generating profits in metaverse competition” and is now discontinuing its long-running VR efforts. In simpler terms, it decided that the investment required for a competitive VR headset would not yield profits. The decision is also said to be due to the lack of games and non-game applications planned for the metaverse.

As we remember, Tencent appeared with a ring-like controller concept. However, this project was also put on hold shortly thereafter. It seems that the Chinese giant will wait a while longer for the metaverse. Sources close to the matter say that the company will not part ways with its VR division.

Final Thoughts

To sum up, the game industry is one of the most important elements of today. Good image quality, being powerful, and extensible are among the most significant features to look for in the games you will play. We can help you in this regard with our sectoral knowledge and extensive connections. For more detailed information on the subject, visit our website. At the same time, you can reach us through our communication channels and review our work through our social media channels (Instagram, Reddit, Twitter, and GitHub).

For futher information about all our services, please visit our Lavega Games and Lavega Project website.

Tencent Ends VR Hardware Plans as Metaverse Competition Falters yazısı ilk önce Lavega Games üzerinde ortaya çıktı.