Tag Archives: Graphics memory

US Targets Chinese Military With New Chip Export Ban!

Leave a reply

The US government just imposed sweeping chip and chipmaking export restrictions that target Chinese military capabilities!

 

US Targets Chinese Military With New Chip Export Ban!

On Friday, October 7, 2022, the US government announced new regulations on computer chips and chipmaking technologies designed to target Chinese military capabilities.

The stated goal of these new export restrictions is to “block the People’s Liberation Army and China’s domestic surveillance apparatus from gaining access to advanced computing capabilities that require the use of advanced semiconductors.”

According to US officials, these rules target not only the sale and export of chips, but also the tools and software that might help the Chinese military in any way, including aiding the development of weapons of mass destruction.

Unlike earlier rules that target specific Chinese companies (like HUAWEI), the new rules have far greater reach, covering everything from chips made by AMD and NVIDIA, to complex and expensive hardware and software used to design and manufacture semiconductor chips.

These rules, some of which go into effect immediately, build on existing restrictions that were earlier applied to top chip toolmaking companies in the US, like KLA Corporation, Lam Research Corporation, and Applied Material Incorporated.

It’s not just advanced chips used for AI computing and national security or military applications that are being targeted.

The new rules are more generic, and forbid US companies from selling technologies to indigenous Chinese companies that will enable the production of:

  • DRAM chips at 18 nm or below,
  • NAND flash chips at 128 layers or above, and
  • logic chips at 14 nm or below.

Even foreign companies operating China are somewhat affected. US companies will need to obtain a license to export more advanced equipment to them.

But in a concession to South Korea, the US government will spare SK Hynix and Samsung from these measures. US companies will be able to continue supplying their production facilities in China.

Many of the rules aim to block foreign companies from selling advanced chips or chipmaking technologies and tools to China. However, the US government will need to “lean” on those countries to introduce similar measures.

Most advanced chips are manufactured in South Korea and Taiwan. If they continue to export to China, they will allow China to bypass US restrictions.

Read more : Biden To Hit China With More Chip Restrictions!

 

Chinese Government Criticised US Chip Export Ban

On Saturday, the Chinese government criticised the new US chip export ban, calling it a violation of International economic and trade rules that will “isolate and backfire” on the US.

Out of the need to maintain its sci-tech hegemony, the U.S. abuses export control measures to maliciously block and suppress Chinese companies.

It will not only damage the legitimate rights and interests of Chinese companies, but also affect American companies’ interest.

– Mao Ning, Chinese Foreign Ministry spokesperson

Despite Mao’s assertions that US actions will not stop China’s progress, the new wide-ranging chip and chipmaking export ban will undoubtedly be detrimental to the Chinese semiconductor industry, and set back its attempts at indigenous production by many years.

On the other hand, this export ban will spur China to redouble its efforts to develop its own chipmaking capabilities, and possibly “strike back” by restricting exports of rare earths to US companies.

Read more : Did China Make 7nm Chips In Spite Of US Sanctions?!

 

US Chip Export Rules Affecting Chinese Companies Summarised

Here is a summary of the new restrictions that take effect on Friday (October 7, 2022):

  • Tools that are capable of producing logic chips made using fin field-effect transistors (FinFET) are blocked from sale to China.
  • Tools capable of fabricating NAND flash storage chips with 128-layer technology or greater, and DRAM based on 18-nanometer half-pitch or less technology are blocked from sale to China.
  • Servicing and maintenance of restricted tools are also banned, which would prevent keeping advanced equipment in good enough shape to keep producing quality chips at high volume.
  • US citizens currently servicing or supporting tools on the restricted list must halt their activities by Wednesday, October 12.
  • Export of items that China can use to make its own chip manufacturing tools, such as a photolithography light source and other specialised components are also blocked from sale to China.

The US Commerce Department also enacted these additional measures on Friday:

  • 31 Chinese entities were added to the Unverified List, which consists of companies that the US government believes could divert technology that they purchase to restricted entities.
  • The Commerce Department expanded the scope of controls for the 28 Chinese firms already on the US Entity List, including presumptively denying any licenses because of the risk they will divert technology to the Chinese military.

Then two weeks later, these restrictions will take effect on Friday, October 21, 2022:

  • Using a new foreign direct product rule, the U.S. will block any chips that used in advanced computing and artificial intelligence applications.
  • The foreign direct product rule can block chips made by non-US companies — including Chinese chip designers — if they use American technology or software.
  • TSMC may be forced by this new rule to halt production on advanced AI or supercomputer chips designed by Chinese firms that are fabricated in Taiwan, unless it gets an exemption.
  • A new foreign direct product rule will apply to components and chips destined for supercomputers in China.

 

Please Support My Work!

Support my work through a bank transfer /  PayPal / credit card!

Name : Adrian Wong
Bank Transfer : CIMB 7064555917 (Swift Code : CIBBMYKL)
Credit Card / Paypal : https://paypal.me/techarp

Dr. Adrian Wong has been writing about tech and science since 1997, even publishing a book with Prentice Hall called Breaking Through The BIOS Barrier (ISBN 978-0131455368) while in medical school.

He continues to devote countless hours every day writing about tech, medicine and science, in his pursuit of facts in a post-truth world.

 

Recommended Reading

Go Back To > Business | Computer | Tech ARP

 

Support Tech ARP!

Please support us by visiting our sponsors, participating in the Tech ARP Forums, or donating to our fund. Thank you!

VGA Share Memory Size from The Tech ARP BIOS Guide!

Leave a reply

VGA Share Memory Size

Common Options for UMA : 1MB, 4MB, 8MB, 16MB, 32MB, 64MB, 128MB

Common Options for DVMT : 1MB, 8MB

 

VGA Share Memory Size : A Quick Review

The VGA Share Memory Size BIOS feature controls the amount of system memory that is allocated to the integrated graphics processor when the system boots up.

However, its effect depends on whether your motherboard supports the older Unified Memory Architecture (UMA) or the newer Dynamic Video Memory Technology (DVMT).

If you have a motherboard that supports UMA, the memory size you select determines the maximum amount of system memory that is allocated to the graphics processor. Once allocated, it can only be used as graphics memory. It is no longer accessible to the operating system or applications.

Therefore, it is recommended that you select the absolute minimum amount of system memory that the graphics processor requires for your monitor. You can calculate it by multiplying the resolution and colour depth that you are using. Of course, if you intend to play 3D games, you will need to allocate more memory.

If you have a motherboard that supports DVMT, the memory size you select determines the maximum amount of system memory that is pre-allocated to the graphics processor. Once allocated, it can only be used as graphics memory. It is no longer accessible to the operating system or applications.

However, unlike in a UMA system, this memory is only allocated for use during the boot process or with MS-DOS or legacy operating systems. Additional system memory is allocated only after the graphics driver is loaded. It is recommended that you set it to 8MB as this allows for high-resolution splash screens as well as higher resolutions in MS-DOS applications and games.

 

VGA Share Memory Size : The Full Details

Some motherboard chipsets come with an integrated graphics processor. To reduce costs, it usually makes use of UMA (Unified Memory Architecture) or DVMT (Dynamic Video Memory Technology) for its memory requirements.

Both technologies allow the integrated graphics processor to requisition some system memory for use as graphics memory. This reduces cost by obviating the need for dedicated graphics memory. Of course, it has some disadvantages :

  • Allocating system memory to the graphics processor reduces the amount of system memory available for the operating system and programs to use.
  • Sharing system memory with the graphics processor saturates the memory bus and reduces the amount of memory bandwidth for both the processor and the graphics processor.

Therefore, integrated graphics processors are usually unsuitable for high-demand 3D applications and games. They are best used for basic 2D graphics and video functions.

The VGA Share Memory Size BIOS feature controls the amount of system memory that is allocated to the integrated graphics processor when the system boots up.

However, its effect depends on whether your motherboard supports the older Unified Memory Architecture (UMA) or the newer Dynamic Video Memory Technology (DVMT).

If you have a motherboard that supports UMA, the memory size you select determines the maximum amount of system memory that is allocated to the graphics processor. Once allocated, it can only be used as graphics memory. It is no longer accessible to the operating system or applications.

Therefore, it is recommended that you select the absolute minimum amount of system memory that the graphics processor requires for your monitor. You can calculate it by multiplying the resolution and colour depth that you are using.

For example, if you use a resolution of 1600 x 1200 and a colour depth of 32-bits, the amount of graphics memory you require will be 1600 x 1200 x 32-bits = 61,440,000 bits or 7.68 MB.

After doubling that to allow for double buffering, the minimum amount of graphics memory you need would be 15.36 MB. You should set this BIOS feature to 16MB in this example.

Of course, if you intend to play 3D games, you will need to allocate more memory. But please remember that once allocated as graphics memory, it is no longer available to the operating system or applications. You need to balance the performance of your 3D games with that of your operating system and applications.

If you have a motherboard that supports DVMT, the memory size you select determines the maximum amount of system memory that is pre-allocated to the graphics processor. Once allocated, it can only be used as graphics memory. It is no longer accessible to the operating system or applications.

However, unlike in a UMA system, this memory is only allocated for use during the boot process or with MS-DOS or legacy operating systems. Additional system memory is allocated only after the graphics driver is loaded. Therefore, the amount of system memory that can be selected is small – only a choice of 1MB or 8MB.

It is recommended that you set it to 8MB as this allows for high-resolution splash screens as well as higher resolutions in MS-DOS applications and games.

 

Recommended Reading

Go Back To > Tech ARP BIOS Guide | Computer | Home

 

Support Tech ARP!

If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!


Samsung Aquabolt – World’s Fastest HBM2 Memory Revealed!

Leave a reply

2018-01-11 – Samsung Electronics today announced that it has started mass production of the Samsung Aquabolt – its 2nd-generation 8 GB High Bandwidth Memory-2 (HBM2) with the fastest data transmission speed on the market today. This is the industry’s first HBM2 to deliver a 2.4 Gbps data transfer speed per pin.

 

Samsung Aquabolt – World’s Fastest HBM2 Memory

Samsung’s new 8GB HBM2 delivers the highest level of DRAM performance, featuring a 2.4Gbps pin speed at 1.2V, which translates into a performance upgrade of nearly 50% per each package, compared to the 1st-generation 8GB HBM2 package with its 1.6Gbps pin speed at 1.2V and 2.0Gbps at 1.35V.

With these improvements, a single Samsung 8GB HBM2 package will offer a 307 GB/s data bandwidth – 9.6X faster than an 8 Gb GDDR5 chip, which provides a 32 GB/s data bandwidth. Using four of the new HBM2 packages in a system will enable a 1.2 TB/s bandwidth. This improves overall system performance by as much as 50%, compared to a system that uses the first-generation 1.6 Gbps HBM2 memory.

[adrotate group=”1″]

 

How Samsung Created Aquabolt

To achieve Aquabolt’s unprecedented performance, Samsung has applied new technologies related to TSV design and thermal control.

A single 8GB HBM2 package consists of eight 8Gb HBM2 dies, which are vertically interconnected using over 5,000 TSVs (Through Silicon Via’s) per die. While using so many TSVs can cause collateral clock skew, Samsung succeeded in minimizing the skew to a very modest level and significantly enhancing chip performance in the process.

In addition, Samsung increased the number of thermal bumps between the HBM2 dies, which enables stronger thermal control in each package. Also, the new HBM2 includes an additional protective layer at the bottom, which increases the package’s overall physical strength.

Go Back To > News | Home

 

Support Tech ARP!

If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!


AMD Vega Memory Architecture Q&A With Jeffrey Cheng

5 Replies

At the AMD Computex 2017 Press Conference, AMD President & CEO Dr. Lisa Su announced that AMD will launch the Radeon Vega Frontier Edition on 27 June 2017, and the Radeon RX Vega graphics cards at the end of July 2017. We figured this is a great time to revisit the new AMD Vega memory architecture.

Now, who better to tell us all about it than AMD Senior Fellow Jeffrey Cheng, who built the AMD Vega memory architecture? Check out this exclusive Q&A session from the AMD Tech Summit in Sonoma!

Updated @ 2017-06-11 : We clarified the difference between the AMD Vega’s 64-bit flat address space, and the 512 TB addressable memory. We also added new key points, and time stamps for the key points.

Originally posted @ 2017-02-04

Don’t forget to also check out the following AMD Vega-related articles :

 

The AMD Vega Memory Architecture

Jeffrey Cheng is an AMD Senior Fellow in the area of memory architecture. The AMD Vega memory architecture refers to how the AMD Vega GPU manages memory utilisation and handles large datasets. It does not deal with the AMD Vega memory hardware design, which includes the High Bandwidth Cache and HBM2 technology.

 

AMD Vega Memory Architecture Q&A Summary

Here are the key takeaway points from the Q&A session with Jeffrey Cheng :

  • Large amounts of DRAM can be used to handle big datasets, but this is not the best solution because DRAM is costly and consumes lots of power (see 2:54).
  • AMD chose to design a heterogenous memory architecture to support various memory technologies like HBM2 and even non-volatile memory (e.g. Radeon Solid State Graphics) (see 4:40 and 8:13).[adrotate group=”2″]
  • At any given moment, the amount of data processed by the GPU is limited, so it doesn’t make sense to store a large dataset in DRAM. It would be better to cache the data required by the GPU on very fast memory (e.g. HBM2), and intelligently move them according to the GPU’s requirements (see 5:40).
  • The AMD Vega’s heterogenous memory architecture allows for easy integration of future memory technologies like storage-class memory (flash memory that can be accessed in bytes, instead of blocks) (see 8:13).
  • The AMD Vega has a 64-bit flat address space for its shaders (see 12:08, 12:36 and 18:21), but like NVIDIA, AMD is (very likely) limiting the addressable memory to 49-bits, giving it 512 TB of addressable memory.
  • AMD Vega has full access to the CPU’s 48-bit address space, with additional bits beyond that used to handle its own internal memory, storage and registers (see 12:16). This ties back to the High Bandwidth Cache Controller and heterogenous memory architecture, which allows the use of different memory and storage types.

  • Game developers currently try to manage data and memory usage, often extremely conservatively to support graphics cards with limited amounts of graphics memory (see 16:29).
  • With the introduction of AMD Vega, AMD wants game developers to leave data and memory management to the GPU. Its High Bandwidth Cache Controller and heterogenous memory system will automatically handle it for them (see 17:19).
  • The memory architectural advantages of AMD Vega will initially have little impact on gaming performance (due to the current conservative approach of game developers). This will change when developers hand over data and memory management to the GPU. (see 24:42).[adrotate group=”2″]
  • The improved memory architecture in AMD Vega will mainly benefit AI applications (e.g. deep machine learning) with their large datasets (see 24:52).

Don’t forget to also check out the following AMD Vega-related articles :

Go Back To > Computer Hardware + Systems | Home

 

Support Tech ARP!

If you like our work, you can help support our work by visiting our sponsors, participating in the Tech ARP Forums, or even donating to our fund. Any help you can render is greatly appreciated!