AMD wants Xilinx: the battle with Nvidia and Intel for data centers and supercomputers

Source: Tweakers.net added 30th Oct 2020

And again there is a mega takeover in the tech industry. AMD wants to acquire Xilinx for no less than 40 billion dollars, converted to some 32 billion euros. Xilinx won’t sound like a household name to the average consumer, and some tweakers may not be fully familiar with the company either. Why is AMD willing to pay such a large amount for a company? What development is the basis of the acquisition?

AMD is not known as a company that regularly makes acquisitions, but of course for that one major acquisition in 2006, when the chip company acquired video card manufacturer ATi for $ 5.4 billion. At the time, that was a large amount for a tech takeover. This takeover not only enabled AMD to release its own video cards, but was especially important to be able to produce APUs, processors with an integrated GPU. The ATi takeover led to a substantial debt burden, which was exacerbated in the years that followed by large losses and fierce competition with Intel. Incidentally, an antitrust settlement with the same Intel, for an amount of 1, 25 billion dollars, AMD a lot of air 2009. Financially, there were in any case few opportunities for major acquisitions for AMD.

Under the current CEO, Lisa Su, who in 2014 took office, the situation has since improved considerably. Under her, AMD has started to focus more on what she believes the company is good at: designing CPUs for desktops and servers in particular. That resulted in the introduction of the Ryzen and the EPYC generation respectively in 2017, which put AMD back on the map. Long-term debt has been significantly reduced and the share price has risen dramatically since 2016. That in turn opened up opportunities for acquisitions and AMD is funding the acquisition with shares.

At Tweakers, we are of course mainly familiar with the Ryzen processors for desktops and laptops, but AMD has now done good business with EPYC. Where Intel had for years in its grip on the server market with the Xeon processors, AMD now has a formidable alternative offering with EPYC. Thanks to their high memory bandwidth, large amount of cores and relatively low price, they are particularly popular for high performance computing , or hpc, such as artificial intelligence and big data.

AMD EPYC code names Gen. Year Name Cores max. 1st 2017 Naples 32 x Zen 1 2nd 2019 Rome 64 x Zen 2 3rd 2020 Milan 64 x Zen 3 4th 2022 Genoa ? x Zen 4 In the current Top 500 of supercomputers are eleven supercomputers with AMD. That’s not much, but in 2018 only two systems were AMD-based. Planning and building supercomputers takes time and we’ve seen a lot of announcements for supercomputers with AMD hardware in recent times. There were, for example, the Hawk supercomputer of the University of Stuttgart, the Flemish Hortense, the exascale supercomputer El Capitan of the Lawrence Livermore National Laboratory and recently the European LUMI, which will be installed in Finland.

From Zilog to Xilinx AMD’s plans to expand Xilinx All to do with the expected further growth of the hpc market, as can be seen from the words of Lisa Su, during the announcement of the acquisition plans. Xilinx is an American company founded in 1984 by former Zilog employees who believed there should be inexpensive chips that customers could program with hardware, in order to optimize them for certain tasks. This field-programmable gate array, or FPGAs, had to be distinguished from generic chips. In the 1990s, the market for FPGAs grew significantly, especially for telecommunications and networks.

Xilinx Virtex UltraScale + -fpga with high-bandwidth memory

FPGAs were also increasingly used for industrial, embedded , automotive and broadcast applications. In recent years, FPGAs have gained popularity for many computing tasks, such as cloud providers Amazon and Microsoft. Amazon, for example, offers Amazon EC2 F1 instances with Xilinx FPGAs to allow developers to use certain computing tasks in research, networking, security, video editing, big data, and analytics. Microsoft has now equipped every Azure server with FPGAs as an accelerator. The chips can not only complete a specific task faster, but also more energy-efficient. In addition, flexibility is an advantage: the same chips can be set in such a way that they can fulfill different functions.

The rise of adaptive computing According to Lisa Su, the market for high performance computing is growing to to be able to handle the increasing calculations considerably and FPGAs would play a greater role in this. GrandView’s research supports its expectation: the estimated total market for FPGAs was approximately $ 9 billion in 2019 and that value would be from 2020 to 2027 are increasing by 9.7 percent annually. Su speaks of ‘adaptive computing’, where hardware is adapted to the chip level to the computing tasks for which it is used. This is one of the ways the computing market is trying to look beyond Moore’s Law.

Su refers to ACAP, or Adaptive Compute Acceleration Platform. Xilinx shares its Versal range under this name. These are chips produced on 7nm by TSMC that include ARM Cortex cores, programmable network-on-a chips, dsp engines, AI engines, PCIe interfaces and CCIX interconnects. Samsung and Ericsson, for example, are going to use ACAP for 5G networks.

The chips work in combination with Xilinx ‘software platform Vitis, which should make programming the chips easy. By combining the various components on a compact chip, Xilinx now has the necessary experience with packaging, in particular with the vertical stacking of different components or chips, knowledge that AMD can benefit from. In the future, AMD could expand upcoming EPYC processors with programmable components that users can optimize for their computing work. per cent. The company derives nearly half of its revenue from the aerospace, defense and industrial systems. Sun 29 percent of revenue comes from the network market and 19 percent comes from automotive and the broadcast sector. According to AMD, the merger with Xilinx would create a company with a turnover of 11, $ 6 billion. Xilinx reported in its recent annual figures a turnover of 3, 19 billion dollars, AMD had sales of 6.7 billion dollars last year but expects its annual sales this year by 41 percent will increase. AMD had such 2019 such a 11. 000 employees, Xilinx has about 5000. Together, according to AMD, the companies receive 13. 000 engineers and the R&D investments together amount to 2.7 billion dollars.

The big competitor of Xilinx is Altera, which also produces FPGAs and therefore according to AMD 2016 serves percent of the market, while Xilinx’s market share 54 would be percent. Altera was acquired by Intel in 2015 for an amount of 16, 7 Billion dollar, still Intel’s largest acquisition. Whether that acquisition has been so successful so far is a matter of debate. The quarterly revenue of Intel’s Programmable Solutions Group, which includes Altera, decreased by 25 Percent from last year to 411 million dollars. In any case, Intel is also getting competition from AMD in this area.

Nvidia and Intel And then there’s Nvidia, which acquired Mellanox last year for $ 6.9 billion. Mellanox makes adapters, switches and chips for InfiniBand and Ethernet network equipment, which are used for supercomputers and data centers. Xilinx, which was also reportedly in the running for an acquisition of Mellanox, subsequently bought Solarflare, which also develops high-bandwidth, low-latency network interconnects. This means that AMD also gets its hands on this technology.

The plans for a takeover that hit like a bomb this year were those of Nvidia, which Arm wants to employ for 40 billion dollars. We know poor as the designer of economical chips for smartphones, but the socs are increasingly used for more powerful devices and also for supercomputers. In fact, the most powerful supercomputer at the moment is Fujitsu’s Fugaku supercomputer with 7.3 million ARM cores. Nvidia therefore intends to acquire Arm mainly because of the growth opportunities in the market for artificial intelligence and other hpc applications.

If the acquisition plans of AMD and Nvidia get approval from the competition authorities, chips for data centers and supercomputers reshuffled. Especially Intel can prepare itself for fierce competition. As we all use more applications that involve complex computing, the battle for the chip market will be more and more fought out in ever-growing computing centers and computing clusters. AMD wants to significantly strengthen its position and broaden its portfolio considerably. It will be a year before that happens: the takeover could be 2021 at the earliest.

You can read more information about FPGAs in this background article: Programming at transistor level – FPGAs on the rise, also for the hobbyist

Read the full article at Tweakers.net

brands: Amazon  AMD  ATI  Fujitsu  Intel  Microsoft  MILAN  NVIDIA  Samsung  SUN  XILINX  
media: Tweakers.net  
keywords: 5G  Amazon  Cloud  Memory  Ryzen  Samsung  Server  Software  Sound  Zen 3  

Related posts


Notice: Undefined variable: all_related in /var/www/vhosts/rondea.com/httpdocs/wp-content/themes/rondea-2-0/single-article.php on line 88

Notice: Undefined variable: all_related in /var/www/vhosts/rondea.com/httpdocs/wp-content/themes/rondea-2-0/single-article.php on line 88

Related Products



Notice: Undefined variable: all_related in /var/www/vhosts/rondea.com/httpdocs/wp-content/themes/rondea-2-0/single-article.php on line 91

Warning: Invalid argument supplied for foreach() in /var/www/vhosts/rondea.com/httpdocs/wp-content/themes/rondea-2-0/single-article.php on line 91