Sintez Global and Talentum invest in FPGA

Due to constantly growing market demand for FPGA processing power Sintez Global and Talentum team during last few months successfully established qualified FPGA development team that is responsible for development and customization of FPGA hardware, installation of it in company data centers and development related applications, bitstreams together with memory configuration files to meet market requirements.

Sintez Global aims to provide Talentum coin holders affordable high-performance equipment that can be easily customized to suit their needs. Such giants as Microsoft Azure, Amazon Web Services, Huawei Cloud, Alibaba Cloud already provide FPGA-based accelerated cloud servers. But many startups and well-established businesses have a lot of difficulties in order to open, customize and managing their accounts due to lack of customization options and difficulties with switching payment methods, billing preferences for existing subscription invoices and many more.

All these issues Sintez Global will solve using their existing capacities as well as additional investments in FPGAs. We already have a list of clients who were waiting for us to launch this service and we are happy to deliver it on time.

Our company has a competitive advantage associated with better pricing, more powerful FPGA performance, flexible customization options and easy cryptocurrency payment policies for used resources with a simple and clear interface.

On 1st of December 2018, we launching first 100 FPGA boosted servers equipped according to client pre-ordered requirements, and by end of the year, we are planning to install additional 400 modified FPGAs that will be ready for lease or used for our mining operations.

What is FPGA?
  • The FPGA is Field Programmable Gate Array that can be used to solve any problem which is computable.

FPGAs are semiconductor devices which contain programmable logic blocks and interconnection circuits. It can be programmed or reprogrammed to the required functionality after manufacturing. This feature of FPGA makes it unique from ASIC. Application Specific Integrated Circuits (ASIC) are custom manufactured for specific design task. In past FPGAs are used to develop low speed, complex and volume design, but today FPGA easily pushes the performance barrier.

FPGAs – are those magic pieces of electronic combine what we love the most in GPUs and ASICs. The only issue here is that as for today, those boards are limited for most of the potential users.

Main features:

• FPGAs can be used to accelerate your applications from 10x to 100x when compared with servers that use CPUs or GPUs alone.

FPGAs are reprogrammable, so you get the flexibility to update and optimize your hardware acceleration without having to redesign any hardware

• User can configure FPGA with its own unique architecture

Less nonrecurring engineering costs (NRE)

Fast design period (time to market)

• FPGA can process multiple parallel tasks (Concurrent Processing)

What are the most common uses for FPGA today?
Genomics Research

The amount and complexity of biological data that must be processed by genomics researchers continues to grow and regularly reaches the petabyte range. Researchers and clinicians must process these data sets very quickly to meet the needs for physicians and their patients. Our FPGA are an ideal solution for these time sensitive use cases.

Precision medicine can be implemented through gene sequencing and analysis as well as rapid analysis of massive biological and medical data. Many fields, such as pharmaceutical development and molecular breeding, also require the processing of massive data. These fields require hardware acceleration to resolve performance bottleneck problems for biological computation. FPGAs meet such requirements due to their outstanding programmable hardware computing performance.

Main advantages are:
  • • High Throughput – improved massive data processing performance
  • • Low Latency – custom hardware circuits for speeding up genetic algorithms and shortening latency

 

Financial Analytics

The financial services industry has increasing needs for HPC capabilities for a wide range of applications including risk modeling and analysis, transaction analysis for security, high frequency trading and more. Financial services organizations can use FPGA to improve the accuracy of risk modeling and analysis that helps dramatically improve their decision making processes.

The financial industry has strict requirements for computing capabilities and real-time performance based on ultra-low latency and high throughput for services, such as pricing tree model-based financial computing, high-frequency trading, fund or securities trading algorithms, financial risk analysis and decision-making, and transaction security assurance. Using programmable hardware acceleration, FPGAs offer an optimal hardware acceleration solution for various scenarios. In certain scenarios, FPGA performance has been improved by hundreds of times when compared with the performance of stand-alone software.

Main advantages are:
  • • High Performance – improved computing performance and analysis accuracy
  • • Low Latency – custom hardware circuits for ultra-low latency

 

Real Time Image and Video Processing

High performance broadcast-quality video applications, such as image processing, video analytics, and video transcoding and compression have real time analysis requirements. Our FPGAs are an ideal solution to meet the requirements of these applications without compromising quality.

Main advantages are:
  • • High Performance – flexible combination of highly parallel computing and RAM resources of FPGAs for video and graphics processing
  • • Low Latency – quick access of external memory in UHD video and Internet-based live program scenarios

 

Deep Learning, Big Data Search and Analytics

The volume, variety, and velocity of data analysis and search requirements in many big data applications have risen to the point where customers are looking for hardware acceleration to keep up. For those applications, customers can leverage the increased performance of our FPGAs to meet their big data analytics and search requirements.

Multi-layer neural networks in machine learning require a large number of computing resources. The training process involves handling massive data, while the inference process requires ultra-low latency. In addition, machine learning algorithms are being continuously optimized. FPGAs meet the preceding requirements due to their high parallel computing, programmable hardware, low power consumption, and low latency. FPGAs dynamically provide the optimal hardware circuit design for different machine learning algorithms, meeting strict requirements for massive computing and ultra-low latency. Therefore, FPGAs meet hardware requirements for machine learning.

Main advantages are:
  • • Flexible – architecture adjustment based on computing models
  • • Cost effective – high-performance, low power consumption solution at low costs

Using this FPGA-enabled hardware architecture, trained neural networks run quickly and with lower latency. We can parallelize pre-trained deep neural networks (DNN) across FPGAs to scale out your service. The DNNs can be pre-trained, as a deep featurizer for transfer learning, or fine-tuned with updated weights.

 

Security

FPGA are useful for many security applications including anti-tampering, information assurance, and trusted relationship management solutions.

Data centers

FPGAs are used in servers for such things as high performance computing, data mining, load balancing, gateways, switches, routers, hardware security modules, computer hardware emulation.

Other FPGA Applications are:

Aerospace and Defense, Medical Electronics, Prototyping, Audio, Automotive, Broadcast, Consumer Electronics, Distributed Monetary Systems, High Performance Computing, Industrial, Medical, Scientific Instruments, Wired and Wireless Communications, Digital signal processing, Bioinformatics, Software-defined radio, Random logic, Prototyping, Medical imaging, Computer hardware emulation, integrating multiple SPLDs, Voice recognition, Cryptography, filtering and communication encoding and many more.

Example of  Machine Learning Acceleration

FPGAs accelerating artificial neural networks for machine learning applications, and accelerate a critical data center workload.

Machine learning applications are rapidly expanding across a growing number of end markets, resident at the edge, in the cloud, or a hybrid of both. FPGA provides the development stacks and hardware platform for deploying advanced and efficient neural networks, algorithms and applications.

Example of  Video Structure Analysis

FPGAs can process realtime video according to defined algorithms.

Example of  Intelligent Cloud Facial Recognition

Example of  Video Processing

Video applications, such as image recognition, image searching, video transcoding, real-time rendering, Internet-based live programs. FPGAs offer cost-effective video solutions, which are ideal for video scenarios.

Example of  VR processing

AR/VR, require high real-time computing performance, which cannot be provided by common ECSs.

Internet of Things (IoT) Example

With FPGA parallel execution we could, for example, read the temperature and humidity sensors continuously. We can’t send data to Sigfox continuously (due to the message limits) but we could compare the sensors against a certain threshold and send an alert.

Because we don’t spend time looping and waiting out the 10-second delay (and avoid executing so many idle instructions), FPGAs tend to be more power efficient for IoT. We process sensor data only when there are updates.

And because FPGAs inherently support parallel programming (instead of single-threaded programming in microcontrollers), we solve the sensor concurrency issue too.

Cryptocurrency Mining

Cryptocurrencies are volatile and unstable. The world of cryptos can be compared to a stormy ocean – if you want to surf it you gotta be ready to maneuver a lot.

FPGAs provide performance close to ASICs, but are flexible and reconfigurable over time to implement new logic.

Back when CPU or GPU mining was profitable for most popular coins, you always had the option to quickly adapt to the crypto market. As soon as a coin would fall, you could instantly launch different mining software and start mining a more profitable coin.

Nowadays with ASICs storming the mining pools for most coins, there is only one strategy – choose a coin, buy an ASIC and hope that it pays off in time. GPU mining is not an option anymore as amount of coins you can mine is limited and profit is unsatisfactory.

The issue with ASICs is that they offer zero flexibility when it comes to which coin you can mine. An ASIC is hard-wired to mine one algorithm only. If for some reason the algorithm becomes unpopular or unprofitable to mine your ASIC becomes an expensive doorstop.

FPGA is a piece of hardware that is very similar to an ASIC with one exception. An ASIC is a chip that has been hardwired to perform one type of calculation (for example, to mine Equihash). FPGA is a chip that can be reprogrammed at will to perform any kind of operations. In the field of mining, you could reconfigure your FPGA from Cryptonight to Lyra2z mining in a split second for example.

FPGA advantages in cryptocurrency mining:

  • • FPGA boards perform 3x to 100x times better than a GPU while having less power draw. Depending on the algorithm, FPGA might or might not slightly fall behind ASICs. 
  • • FPGAs have complete flexibility when it comes to mineable algorithms – no soft forks can affect mining operations as long as we update our FPGA bitstream.
  • • Tremendous power efficiency compared to GPUs and ASICs

Our team current plan is to release around two algorithms per month for multiple coins, until all major algorithms will be covered. Our team already developed bitstreams for some of the top 100 mineable coins and keep on working on new, more efficient and profitable bitstreams.

With many gigabytes of external memory, hundreds of different bitstream configuration files can be stored. In this fashion, during real world cryptocurrency mining, an FPGA has the power to reconfigure itself in a fraction of a second based on circumstances that occur during the mining operations.

Some algorithms like Timetravel10, X11Evo, X16R and X16S have hash function sequences that change every few minutes during mining. An FPGA can effectively mine those algorithms by using the rapid reconfiguration function. Since this type of rapid reconfiguration is not available to ASIC chips, it is quite possible that high end FPGA’s could never be defeated by ASIC’s on this class of algorithms.

FPGA will let us mine almost any algorithm at any time. Unlike an ASIC that requires to invest into mining one single algorithm, FPGA will become the all-in-one universal mining solution

Our current FPGA payback period around 4 months. But the most beautiful part of it that any fork will not affect our income and we are still multipurpose.

With our FPGA cloud service payback can be even less.

Currently our system supports 5000 FPGA with profit around $26/day for each, which can compete with top ASIC miners available on the market:

Most profitable ASIC miner as on 21 November 2018 – https://www.asicminervalue.com

Best ASIC payback period as on 21 November 2018 – https://cryptominer.deals/seller/miner-bros?sort=payback&direction=asc&page=1

Profitability curve of one of top Bitmain* miners with is not released yet.

*Bitmain – cryptocurrency mining company and mining equipment manufacturer which is going to IPO on the Hong Kong Stock Exchange between the fourth quarter of 2018 and the first quarter of 2019. According to investment analysts, they expect to raise anywhere from $3 billion to $18 billion, at their market capitalization of $40 to $50 billion, thereby becoming one of the largest initial public offering in the IT market history, beating Facebook with its $16 billion.

Bitmain earned $701 million in net profit in 2017, while various estimates show that the annual income for the same period ranged from $1 billion to $4 billion. A gross income claimed for the first half of 2018 exceeded the one received for the whole previous year and comprised $743 million, despite a significant fall in the crypto market.

We truly believe that this investments in new developers team and FPGA hardware development will provide us stable earnings in a rapidly changing and expanding blockchain technology market as well as improve security for our own developments.

Why FPGAs?

The answer is quite simple, this type of equipment is the most flexible and have longer period of usage compare to ASICs. Other points are:

  • • Current market have a very high demand on video and image processing power
  • • FPGAs are very flexible for cryptocurrency mining operations
  • • FPGAs have less power consumption compare to the most of ASICs
  • • 1 FPGA hashing power with same power consumption can be equal to 30 top-end GPUs
  • • Most of the new cryptocurrency like Talentum have ASIC resistant hashing algorithms but current GPUs hashing power is not enough to meet market demand and make mining operations profitable with current rate of cryptocurrencies.
  • • Faster and low latency processing power for data center operations

One of the reasons why we are investing in FPGAs is that there is data explosion – the amount of data by 2020 will be around 44ZB while in 2013 it was only 4.4 ZB.

Due to constantly increasing demand and shortage of FPGA equipment on the market our company see a big opportunity to join this type of hardware development, hosting, leasing and usage.

Our final testings were completed sucessfully and by the end on November our infrastructure will be completely updated and ready to host our modified FPGA.

Testing units using new bitstreams for Talentum mining operations