Category: News

Nvidia’s reports a whopping $6.43 billion revenue in the Q2 of 2021
Nvidia released its earnings results for the second quarter of the fiscal year of 2022. and according to the released report, the company reported that it made $6.5 billion in revenue, which is slightly more than its guidance at the end of the previous quarter, where NVIDIA had estimated revenue to stand at $6.43 billion during the second quarter.
This growth was the product of NVIDIA’s gaming and data centre segments, and the company had announced that each segment reported a growth of 11% and 16% sequentially. The company’s financial chief, Ms Colette Kress, stated that the company’s revenue growth in gaming came from its o higher of graphics cards and console processors. In terms of growth in data centres, she attributed the company’s growth to its Ampere architecture that witnessed a higher penetration into segments like inference.
By the end of Q2, NVIDIA was at a whopping $2.4 billion in operating income and $2.37 billion in net income, marking a remarkable growth of 275% and 282% in each of the sectors, respectively. The company reported more than double the 112% operating income and 109% net income growths it exhibited at the end of its first quarter.
While reporting the revenue growth the company’s chief financial officer Mss Colette Kress followed the tradition of providing a brief explanation of her company’s performance during the quarter. According to her, NVIDIA’s gaming, data centre and professional visualization segments have a record revenue during the quarter with them standing at $3, $2.4 and $0.5 billion, respectively in Q2.
Specifically, while commenting on the gaming and data centre performance, the CFO outlined:
Gaming revenue was up 85 per cent from a year ago and up 11 per cent sequentially, reflecting higher sales in GeForce GPUs and game console SOCs. We continued to benefit from strong sales of our GeForce RTXTM 30 Series based on the NVIDIA Ampere architecture. We have introduced Low Hash Rate (LHR) GeForce GPUs with limited Ethereum mining capability and increased the supply of Cryptocurrency Mining Processors (CMP) to direct GeForce to gamers and CMP to miners. Over 80 per cent of our Ampere architecture-based GeForce GPU shipments in the quarter were LHR GPUs[EMPHASIS ADDED]. CMP is included in OEM.
Data Center revenue was up 35 per cent from a year ago and up 16 per cent sequentially. The year-on-year growth was led by the ramp of NVIDIA Ampere architecture products into vertical industries and hyperscale customers, including strong growth in inference. Sequentially, growth stemmed from both compute and networking products, led by hyperscale customers.
The Chief Financial Officer also shared some latest information as to what the company has planned for the acquisition of British chip design firm Arm Ltd., which is currently facing significant hurdles across regulatory bodies in major jurisdictions.
We are working through the regulatory process for our pending acquisition of Arm Ltd. Although some Arm licensees have expressed concerns or objected to the transaction, and discussions with regulators are taking longer than initially thought, we are confident in the deal and that regulators should recognize the benefits of the acquisition to Arm, its licensees, and the industry.
source
August 19, 2021
Apple ranks top on TSMC’s customer list for its 3nm node with AMD rumoured to be the second
We already knew that the Taiwan Semiconductor Manufacturing Company’s (TSMC) will be entering into the mass production of its 3nm process nodes. And according to fresh reports, the orders for its next-generation 3nm chip manufacturing process have already started to materialize.
Due to the strong performance of its 7nm process node and intel delaying its manufacturing process, TSMC rose to new heights becoming the world’s largest contract chip manufacturer. It gains more market share as it received orders from big names like Apple.
According to fresh sources, while Apple still retains the place of the biggest customer of TSMC, Intel’s rival Advanced Micro Devices, is soon to become the second-largest customer of the chip manufacturer. This is because AMD witnessed huge success in its chip segments, especially with the EPYC server products.
Recent reports indicate that orders are being flooded to the TSMC’s table and they are coming from three prominent names in the market today namely AMD, Apple and Taiwanese consumer electronics firm MediaTek.
Currently, TSMC is mass-producing semiconductors through its 5nm process node, and it’s expected that most of the bulk of the outputs will head over to Apple for its upcoming annual smartphone refresh.
According to sources, for the company’s 3nm process node technology which will soon be entering mass production Apple, AMD and MediaTek are TSMC’s top three customers.
However, the presence of the company in a higher position in the market isn’t without any hassles. The company is currently dealing with Geopolitical tensions with mainland China and it is still working on establishing its new fab in Arizona, United States.
The current rise and recent drought in Taiwan have also affected the company, but its production remains strong. For more information stay tuned.
source
August 19, 2021
AMD’s Radeon RX 6600 XT GPU can reportedly achieve an overclock up to 2,800 MHz
AMD’s recently launched its latest Radeon RX 6600 XT GPU targeted towards a middle segment of PC gamers. The GPU comes with 2048 streaming processors and can run from 2359 to 2589 MHz on the factory design. However, according to sources, the GPU can reportedly be overclocked to 2,800 MHz on basic air cooling and it remains stable.
The information comes from Igor’s Lab, and an overclocker named Gurdi who has an AMD Radeon RX 6600 XT graphics card has managed to run the silicon at a 2.8 GHz clock speed with full stability under stress testing.
Here’s a breakdown of how it happened, the card used a custom beta software called MorePowerTool (MPT). It is developed by Igor’s Lab forum member “hello”, and the software offers better overclocking of AMD Radeon graphics cards by enabling toggle changes that are not available in the regular AMD Wattman software.
What Guardi has done to manage to achieve the 2.8 GHz overclock, is that he pushed the GPU power limit from 155 to 160 Watts. For the GFX (GPU) core, the current is set anywhere from 127A to 130A and a constant 25A for the whole SoC. The GPUs memory reportedly needs a slight voltage boost where the report notes a DPM 3 value 1400mV.
And after overclocking the GPU, due to its frequency boost the card managed to provide the card with few extra frames. The testing was conducted by Igor’s Lab with 3D Mark synthetic benchmark and on a few games like Cyberpunk 2077, Red Dead Redemption 2, Godfall, and more. However, we expect that this isn’t the limit for the AMD’s Radeon RX 6600 XT’s overclocking capabilities and if you are an owner of the GPU do test it to its limit, if you are willing, and let us know in the comments.
source
August 19, 2021
Netflix starts providing Spatial Audio support on iPhone and iPads
Good news for the iPhone and/or iPad users who are also Netflix fans in parallel. Netflix starts releasing the update for Spatial Audio support to the iPhone and iPad. As you probably know, it applies the directional audio filters to enable a mesmerizing listening experience.
The update already started reaching out to the users having iPhones and/or iPads with iOS 14. And you need to keep in mind that this Spatial Audio feature is exclusive to the AirPods Pro and AirPods Max.
The iOS 15 upgrade will push the experience even further as it can provide the Spatial Audio experience for the non-Dolby Atmos contents as well. Thanks to the new Spatialize Stereo option.
If you still didn’t get the Spatial Audio in the Netflix application on your Apple device, check if you have the latest version of the application or not. If you are on the latest version and still not getting it, you need to keep checking over the coming weeks. That is because the rollout is very slow and has yet to reach various areas across the globe.
Source
August 19, 2021

Baidu’s Kunlun enter’s into the mass production phase of its Kunlun II AI Processor

Back in June, it was reported that Baidu has successfully spun off its semiconductor design business into an independent company and the company is now called the Kunlun Chip Technology Co. Who values it at around $2 billion?

Kunlun Chip is the wholly-owned subsidiary of Chinese high-tech giant Baidu, and the company has now entered into volume production of its Kunlun II processor for AI applications. Their new chip for AI is based on the 2nd generation XPU microarchitecture and is made of a 7 nm process technology.

Three years ago the first generation Kunlun K200 processor was released and this is designed for the cloud, edge, and autonomous vehicles applications. It offers around 256 INT8 TOPS performance, around 64 TOPS INT/FP16 performance, and 16 INT/FP32 TOPS performance while powered at 150 Watts.

The company has now announced that its latest generation of Kunlun processors will have about 2-3 times higher performance than its predecessor. If this is true then the new chip can provide from 512 to 768 INT8 TOPS, 128 – 192 INT/FP16 TOPS, and 32 – 48 INT/FP32 TOPS throughput. It can easily compete against the likes of Nvidia’s A100 in AI computations.

Baidu Kunlun II’s Relative Performance

	Baidu Kunlun	Baidu Kunlun II	Nvidia A100
INT8	256 TOPS	512 ~ 768 TOPS	624/1248* TOPS
INT/FP16	64 TOPS	128 ~ 192 TOPS	312/624* TFLOPS (bfloat16/FP16 tensor)
Tensor Float 32 (TF32)	–	–	156/312* TFLOPS
INT/FP32	16 TOPS	32 ~ 48 TOPS	19.5 TFLOPS
FP64 Tensor Core	–	–	19.5 TFLOPS
FP64	–	–	9.7 TFLOPS

*With sparsity

For those who don’t know, the Kunlun project was initiated by Baidu back in 2011. Initially, its many-small-core XPU microarchitecture using FPGAs were emulated however in 2018 the company built dedicated silicon using one of Samsung Foundry’s 14nm fabrication process.

source

August 19, 2021

US Tech Giants Slow Manufacturing Shift Out Of China
Covid-19 is disrupting plans by Apple, Google, Amazon and others to shift production from China to Vietnam, according to Nikkei Asia. Both Apple and Google have been considering shifting some of their production out of China and into Vietnam. Apple originally planned to have the third-generation AirPods made in Vietnam but two people familiar with Apple’s plans told Nikkei Asia that Apple will have the next-gen wireless Bluetooth earbuds manufactured in China. The sources said that Apple still hopes to move 20% of new AirPods production to Vietnam eventually.
Back in 2019, Apple did move some AirPods and AirPods Pro production to Vietnam. This took place during the U.S.-China trade war when the U.S. was taxing imports from China and vice versa. The same year, it was reported that Google would aggressively move production of U.S. bound hardware out of China and into an old Nokia plant in North Vietnam. But this move, which was supposed to happen in time for the production of the Pixel 4 line, was not done.
Instead, the Pixel 5 was built in China and according to four people familiar with the matter, Google will build the Pixel 6 series in China as well although production of the line was supposed to move to Vietnam early last year. The reason why Google will continue to build the new Pixel units in Shenzhen where the Pixel 5 models were made instead of in Vietnam has to do with the latter country’s limited engineering resources and travel restrictions.
Vietnam has a lot going for it as a place where tech products can be assembled. It is close to China and has a young labor force. When U.S. companies were looking for countries to replace China in order to avoid tariffs, Vietnam was one of the top choices. But the lack of experienced engineers makes building a regional supply chain in the area difficult.
The absence of experienced engineers in Vietnam also makes it difficult to start the production of new products in the country. One supply chain executive whose company works with Apple and Google says, “Vietnam’s engineering workforce is still far from adequate. With all the travel restrictions, it’s only feasible to make products in Vietnam that are already in mass production elsewhere, rather than starting production of upcoming products from scratch in the country.”
China and Vietnam have both tightened up their borders and are issuing passports at a slower rate in an attempt to slow down the spread of the delta variant of COVID-19. This has been confirmed by China’s National Immigration Administration. Moving production to Vietnam was made more difficult by a wave of COVID-19 infections that took place in the country back in May.
Several supply chain managers in China told the Nikkei that their staff is finding it harder to apply for business trips to Vietnam. A senior manager of an Amazon supplier says, “The border control has been tightened in the past few months. We could not easily dispatch our Chinese engineers to support our production projects for Amazon in northern Vietnam, so the company has been bringing in fully vaccinated Taiwanese engineers from China.”
Also Read:
- OPPO sets up specialised camera innovation lab at Hyderabad Research centre
- Google launches the Pixel 5a featuring 5G support
August 19, 2021
Apple plans to resume in-person store classes despite the growing COVID cases in the US
Apple, one of the leading tech giants of the world has reported to the U.S. and European retail staff that it is planning to bring back in-person classes to its stores on Aug. 30. however, fresh reports indicate that these plans might be delayed due to the increasing cases of COVID-19 in the united states. But this might only affect the opening of stores in some places which are reported a surge in cases.
The Cupertino giant started a reservation portal for in-store classes on Monday, which allowed the consumers to sign up for the courses in advance. The company also announced that it would resume its classes in most of the company’s U.S. locations, including New York stores, but not everywhere. The classes — branded Today at Apple — were last held in March 2020, just before the pandemic triggered widespread lockdowns.
This is Cupertino giants latest attempt to get its operations back on track after a year and a half of Covid disruptions. Following the decrease in cases of COVID the company even announced back in June that mask will not be mandated in its stores, however, this didn’t last long. The company had to again reinstate COVID procedures after the delta variant led to a spike in virus cases.
Apple further announced that its corporate employees have to return to their offices until October at the earliest and its also being speculated that this return to office date will be even further extended owing to the increase in the new Delta Variant cases.
The Today at Apple classes, first launched in 2017, include lessons for new iPhone and Mac buyers. This class includes tutorials on how to use their devices, along with more advanced courses related to tools such as the GarageBand music-editing software.
source
August 19, 2021
Apple filed Apple Car patents, see details here…
Apple has amassed billions of dollars with iPhones, MacBooks, Apple Watches, and their many other products. Now the company is determined to get a share of the growing electric car industry, with their next big thing — the Apple Car. Tesla is currently dominating the EV market and will only become increasingly popular, with nearly 500,000 cars delivered in 2020.
Apple in 2021 also sent a team of Apple Car employees to South Korea to meet with LG, the SK group, and others to discuss possible business opportunities related to the Apple Car. Now, Apple has filed a patent which revealed that Apple car owners can authorize other users to drive the car without releasing the keys. Indeed, there have been several Patents relating to the Apple car covering a wide array of concepts and components that have been filed.
“If access is controlled by a physical key or fob, the owner must physically deliver the key/fob to the individual and then must physically retrieve the key/fob after the individual is finished with access,” Apple said.
“If access is controlled by secret data, the owner must communicate the secret data to the individual. If the communication is insecure, a third party can observe the secret data. Additionally, the secret data must be changed after the individual is finished with access,” Apple added further.
Although we are still unaware about other details of the Apple Car and when the car will going to debut. Some sources claim that it can come in 2025 but it’s just the speculations. We have to wait till the official announcement from Apple.
Also Read:
August 19, 2021

AMD’s next-gen Ryzen Threadripper CPUs TDPs leak online

The latest un-intentional or intentional leak by Gigabyte has given away the details about AMD’s next-generation Ryzen Threadripper CPUs based on the Zen 3 core architecture. The CPUs are codenamed Chagall, and from previous leaks, we already had an idea that AMD will be introducing its next-generation Ryzen Threadripper & Threadripper Pro families based on the Zen 3 core architecture under the Chagall codename. It’s expected that the lineup will release the CPUs in November 2021.

According to the latest leaks, courtesy of Gigabyte, we now know that the AMD Ryzen Threadripper 5000 lineup is an ‘AMD Family 19h 08he processors and will have at least eight different SKUs, which will be 5 SKUs for the workstation WRX80 and 3 SKUs for the HEDT TRX40 platform.

It also revealed that the workstation models will feature 64, 32, 24, 16, 12 core and HEDT models will have 64, 32, and 24 core variants. The most interesting thing to note here is that the IOD on the HEDT variants reports a power rating of 80W, however, each workstation chip has a different IOD power rating.

Coming first to the 64 core variant it will feature an IOD power rating of 110W, the 85W will be for the 32 core, 80W for the 24 core, 75W for the 16 core, and 62W for the 12 core variant.

The 64 core workstation has a 35W higher power rating than its HEDT equivalent and this is due to the workstation variant offering 8-channel IMC, deliver up to 128 Gen 4 PCIe lanes vs 64 Gen 4 PCIe lanes on the HEDT models.

We know that AMD will split its Threadripper 5000 family into two segments, the Ryzen and the Pro variants. It aimed at the HEDT and Prosumer market. It will also be featuring the Ryzen Threadripper 5990X which is the flagship with up to 64 cores, 128 threads & 256 MB of L3 cache. The SKU will also offer 64 PCIe Gen 4.0 lanes and support a 4-channel DDR-3200 memory interface.

AMD Zen CPU / APU Roadmap:

Zen Architecture	Zen 1	Zen+	Zen 2	Zen 3	Zen 3+	Zen 4	Zen 5
Process Node	14nm	12nm	7nm	7nm	6nm?	5nm	3nm?
Server	EPYC Naples (1st Gen)	N/A	EPYC Rome (2nd Gen)	EPYC Milan (3rd Gen)	N/A	EPYC Genoa (4th Gen) EPYC Bergamo (5th Gen?)	EPYC Turin (6th Gen)
High-End Desktop	Ryzen Threadripper 1000 (White Haven)	Ryzen Threadripper 2000 (Coflax)	Ryzen Threadripper 3000 (Castle Peak)	Ryzen Threadripper 5000 (Chagal)	N/A	Ryzen Threadripper 6000 (TBA)	TBA
Mainstream Desktop CPUs	Ryzen 1000 (Summit Ridge)	Ryzen 2000 (Pinnacle Ridge)	Ryzen 3000 (Matisse)	Ryzen 5000 (Vermeer)	Ryzen 6000 (Warhol / Cancelled)	Ryzen 7000 (Raphael)	Ryzen 8000 (Granite Ridge)
Mainstream Desktop . Notebook APU	Ryzen 2000 (Raven Ridge)	Ryzen 3000 (Picasso)	Ryzen 4000 (Renoir) Ryzen 5000 (Lucienne)	Ryzen 5000 (Cezanne) Ryzen 6000 (Barcelo)	Ryzen 6000 (Rembrandt)	Ryzen 7000 (Phoenix)	Ryzen 8000 (Strix Point)
Low-Power Mobile	N/A	N/A	Ryzen 5000 (Van Gogh) Ryzen 6000 (Dragon Crest)	TBA	TBA	TBA	TBA

source

August 18, 2021

Rambus announces the completion of the HBM3 memory subsystem topping at 8.4 Gbps of transfer speed
Rambus Incorporated, founded in 1990, is an American technology company that designs, develops and licenses chip interface technologies and architectures that are used in digital electronics products. The company is well known for inventing RDRAM and for its intellectual property-based litigation following the introduction of DDR-SDRAM memory.
Rambus, a renowned chip manufacturer has just announced that it has completed the development of its advanced HBM3 memory subsystem. According to the manufacturer, this memory subsystem can hit transfer speeds of up to 8.4 Gbps.
Currently, in the market, we have the HBM2E which boasts to be the fastest memory option available and offer transfer rates of up to 3.2 Gbps. However, the HBM3 aims to offer more than double the speed of the current market standard offering an insane 8.4 Gbps transfer speed resulting in higher bandwidth.
According to sources, a single HBM2E package can peak out at 460 GB/s bandwidth however an HBM3 will offer up to 1.075 TB/s bandwidth which is 2x the total bandwidth jump.
Rambus had this to say about the technology:
The memory bandwidth requirements of AI/ML training are insatiable with leading-edge training models now surpassing billions of parameters,” said Soo Kyoum Kim, associate vice president, Memory Semiconductors at IDC. “The Rambus HBM3-ready memory subsystem raises the bar for performance enabling state-of-the-art AI/ML and HPC applications.”
Rambus achieves HBM3 operation of up to 8.4 Gbps leveraging over 30 years of high-speed signalling expertise, and a strong history of 2.5D memory system architecture design and enablement. In addition to the fully integrated HBM3-ready memory subsystem, Rambus provides its customers with interposer and package reference designs to speed their products to market.
“With the performance achieved by our HBM3-ready memory subsystem, designers can deliver the bandwidth needed by the most demanding designs,” said Matt Jones, general manager of Interface IP at Rambus. “Our fully-integrated PHY and digital controller solution builds on our broad installed base of HBM2 customer deployments and is backed by a full suite of support services to ensure first-time-right implementations for mission-critical AI/ML designs.”
Let’s look at some benefits of the Rambus HBM3-ready Memory Interface Subsystem:
- Supports up to 8.4 Gbps data rate delivering bandwidth of 1.075 Terabytes per second (TB/s)
- Reduces ASIC design complexity and speeds time to market with fully integrated PHY and digital controller
- Delivers full bandwidth performance across all data traffic scenarios
- Supports HBM3 RAS features
- Includes built-in hardware-level performance activity monitor
- Provides access to Rambus system and SI/PI experts helping ASIC designers to ensure maximum signal and power integrity for devices and systems
- Includes 2.5D package and interposer reference design as part of IP license
- Features LabStation development environment that enables quick system bring-up, characterization, and debugs
- Enables the highest performance in applications including state-of-the-art AI/ML training and high-performance computing (HPC) systems
Many in the industry expect the first generation of HBM3 memory to be very similar to HBM2E, meaning that in terms of capacity it should be made up of 16Gb DRAM Dies for a total of 16 GB (8-hi stack).
However, there are reports that the memory densities of HBM3 will increase once the specifications are finalized by JEDEC. It also reported that products such as the AMD’s Instinct Accelerators based on next-gen CDNA architecture, NVIDIA’s Hopper GPUs, and Intel’s future HPC accelerators based on next-gen Xe-HPC architecture, all possess the HDM3 memory stack.
source
August 18, 2021

Category: News

Baidu Kunlun II’s Relative Performance

Also Read:

AMD Zen CPU / APU Roadmap: