Cray provides a suite of highly advanced systems – for the single user
to large research centers – that meet existing and future computational
challenges on a wide variety of applications every day. Building on
decades of experience in designing the world’s most innovative
supercomputers, Cray technologies enable remarkable breakthroughs by
accelerating performance, improving efficiency and extending the
capabilities of the most demanding applications. And with a solution for
every budget and need, Cray makes it easy to take advantage of high
performance computing (HPC) advancements. Put a Cray on your team and
get the sustained performance, scalability and reliability to solve your
toughest problems.
Cray XK6™
Supercomputer
The Cray XK6 supercomputer
combines Cray's
proven Gemini
interconnect, AMD's
leading multi-core
scalar processors
and NVIDIA's powerful
many-core GPU processors
to create a true, productive hybrid supercomputer.
Cray XE6™ and Cray XE6m™ Supercomputers
The Cray XE6 scalable supercomputer is engineered to meet
the demanding needs of capability-class HPC applications. The Cray XE6m
is optimized to support scalable workloads in the midrange market.
YarcData uRiKA™ Graph Appliance
The YarcData uRiKA graph appliance is a purpose built
solution for Big Data relationship analytics. uRiKA enables enterprises
to discover unknown and hidden relationships in Big Data, perform
real-time analytics on Big Data graph problems, and realize rapid time
to value on Big Data solutions.
The uRiKA graph appliance complements an existing data warehouse or Hadoop cluster.
Cray Sonexion™ 1300 Storage System
The Cray Sonexion 1300 system is an integrated, high
performance storage system that features next-generation modular
technology to maximize the performance and capacity scaling capabilities
of the Lustre file system. Cray also offers custom and third-party
storage and data management solutions.
May 23, 2012
Computer :: Super :: Japan’s K Computer Tops 10 Petaflop/s to Stay Atop TOP500 List
BERKELEY, Calif.; KNOXVILLE, Tenn.; and MANNHEIM, Germany (Nov. 14,
2011)—Japan’s “K Computer” maintained its position atop the newest
edition of the TOP500 List of the world’s most powerful supercomputers,
thanks to a full build-out that makes it four times as powerful as its
nearest competitor. Installed at the RIKEN Advanced Institute for
Computational Science (AICS) in Kobe, Japan, the K Computer it achieved
an impressive 10.51 Petaflop/s on the Linpack benchmark using 705,024
SPARC64 processing cores.
The K Computer is the first supercomputer to achieve a performance level of 10 Petaflop/s, or 10 quadrillion calculations per second. In June 2011, the partially built K computer had taken the No. 1 position with a performance of 8.16 Petaflop/s. Contrary to many other recent very large systems, it does not utilize graphics processors or other accelerators. The K Computer is also one of the most energy efficient systems on the list.
Still in second place is the Chinese Tianhe-1A system with 2.57 Petaflop/s performance. One year ago, the Tianhe-1A system took the top spot, but was dethroned when the next TOP500 list was published six months ago.
In fact, the Top 10 supercomputers on the latest list – the 38th edition of the twice-yearly list – remain unchanged from June 2011. The latest list, the data behind it and the trends it reflects will be the topic of a Birds-of-a-Feather session to be held at 5:30 p.m. Tuesday, Nov. 15, at the SC11 supercomputing conference in Seattle.
“This is the first time since we began publishing the list back in 1993 that the top 10 systems showed no turnover,” said TOP500 editor Erich Strohmaier, who will lead the discussion at SC11.
The largest U.S. system is a Cray XT5 system called Jaguar and installed at the Oak Ridge National Laboratory, with a 1.75 Petaflop/s performance running the standard Linpack benchmark application. Other top U.S. systems include Cielo, a Cray XE6 at Los Alamos National Laboratory (No. 6); Pleiades, an SGI Altix machine at NASA’s Ames Research Center (No.7); Hopper, a Cray XE6 at the National Energy Research Scientific Computing Center (No. 8); and Roadrunner, an IBM system that was the first ever to break the Petaflop/s barrier, at Los Alamos (No. 10). Systems in China, Japan and France round out the Top 10.
Although the top rankings did not change, the newest list does highlight a number of other developments. For example:
The K Computer is the first supercomputer to achieve a performance level of 10 Petaflop/s, or 10 quadrillion calculations per second. In June 2011, the partially built K computer had taken the No. 1 position with a performance of 8.16 Petaflop/s. Contrary to many other recent very large systems, it does not utilize graphics processors or other accelerators. The K Computer is also one of the most energy efficient systems on the list.
Still in second place is the Chinese Tianhe-1A system with 2.57 Petaflop/s performance. One year ago, the Tianhe-1A system took the top spot, but was dethroned when the next TOP500 list was published six months ago.
In fact, the Top 10 supercomputers on the latest list – the 38th edition of the twice-yearly list – remain unchanged from June 2011. The latest list, the data behind it and the trends it reflects will be the topic of a Birds-of-a-Feather session to be held at 5:30 p.m. Tuesday, Nov. 15, at the SC11 supercomputing conference in Seattle.
“This is the first time since we began publishing the list back in 1993 that the top 10 systems showed no turnover,” said TOP500 editor Erich Strohmaier, who will lead the discussion at SC11.
The largest U.S. system is a Cray XT5 system called Jaguar and installed at the Oak Ridge National Laboratory, with a 1.75 Petaflop/s performance running the standard Linpack benchmark application. Other top U.S. systems include Cielo, a Cray XE6 at Los Alamos National Laboratory (No. 6); Pleiades, an SGI Altix machine at NASA’s Ames Research Center (No.7); Hopper, a Cray XE6 at the National Energy Research Scientific Computing Center (No. 8); and Roadrunner, an IBM system that was the first ever to break the Petaflop/s barrier, at Los Alamos (No. 10). Systems in China, Japan and France round out the Top 10.
Although the top rankings did not change, the newest list does highlight a number of other developments. For example:
- China keeps increasing its number of systems to 75 and is now clearly the No. 2 country, as a user of HPC, ahead of Japan, UK, France, and Germany.
- The two Chinese systems at No.2 and No. 4 and the Japanese Tsubame 2.0 system at No. 5 are all using NVIDIA GPUs to accelerate computation.
- Thirty-nine systems use GPUs as accelerators (up from 17 six months ago), 35 of these use NVIDIA chips, two use Cell processors, and two use ATI Radeon.
- Already, 62 percent of the systems use processors with six or more cores.
- Intel continues to provide the processors for the largest share (76.8 percent) of TOP500 systems.
- Thanks to the K computer Fujitsu captured the No. 2 spot in market share by total performance slightly ahead of Cray, but IBM stays well ahead of both.
Other points of interest include:
- A total of 384 systems (76.8 percent) are now using Intel processors. This is slightly down from six months ago (386 systems 77.2 percent).
- They are now followed by the AMD Opteron family with 63 systems (12.6 percent), down from 66.
- The share of IBM Power processors has stabilized for now with 49 systems (9.8 percent), up from 45.
- Gigabit Ethernet is still the most-used internal system interconnect technology (223 systems, down from 230 systems), due to its widespread use at industrial customers, followed by InfiniBand technology with 213 systems, up from 208 systems. However, InfiniBand-based systems account for almost twice as much performance (28.7 Petaflop/s) than Gigabit Ethernet ones (14.2 Petaflop/s).
- IBM kept its lead in systems and has now 223 systems (44.6 percent) compared to HP with 140 systems (28.0 percent). HP had 146 systems (29.2 percent) six months ago, compared to IBM with 218 systems (43.6 percent).
Power consumption of supercomputers
TOP500 now tracks actual power consumption of supercomputers in a consistent fashion.- 29 systems on the list are confirmed to use more than 1 megawatt (MW) of power.
- The No. 1 system, the K computer also reports the highest total power consumption of 12.66 MW. Yet due to its performance, the system is one of the most efficient systems on the list, delivering 830 Mflops/watt.
- The most energy efficient supercomputers are BlueGene/Q with 2,029 Mflops/watt.
- Average power efficiency is 282 Mflops/watt (up from 248 Mflops/watt six months ago and 219 Mflops/watt one year ago).
- Average power consumption of a TOP10 system is 4.56 MW (up from 4.3 MW six months ago) and average power efficiency is 464 Mflops/watt (unchanged).
About the TOP500 List
The TOP500 list is compiled by Hans Meuer of the University of Mannheim, Germany; Erich Strohmaier and Horst Simon of NERSC/Lawrence Berkeley National Laboratory; and Jack Dongarra of the University of Tennessee, Knoxville.Computer :: How Computer Memory Work (5)
Cache and Registers
Caches are designed to alleviate this bottleneck by making the data used most often by the CPU instantly available. This is accomplished by building a small amount of memory, known as primary or level 1 cache, right into the CPU. Level 1 cache is very small, normally ranging between 2 kilobytes (KB) and 64 KB.The secondary or level 2 cache typically resides on a memory card located near the CPU. The level 2 cache has a direct connection to the CPU. A dedicated integrated circuit on the motherboard, the L2 controller, regulates the use of the level 2 cache by the CPU. Depending on the CPU, the size of the level 2 cache ranges from 256 KB to 2 megabytes (MB). In most systems, data needed by the CPU is accessed from the cache approximately 95 percent of the time, greatly reducing the overhead needed when the CPU has to wait for data from the main memory.
Some inexpensive systems dispense with the level 2 cache altogether. Many high performance CPUs now have the level 2 cache actually built into the CPU chip itself. Therefore, the size of the level 2 cache and whether it is onboard (on the CPU) is a major determining factor in the performance of a CPU. For more details on caching, see How Caching Works.
A particular type of RAM, static random access memory (SRAM), is used primarily for cache. SRAM uses multiple transistors, typically four to six, for each memory cell. It has an external gate array known as a bistable multivibrator that switches, or flip-flops, between two states. This means that it does not have to be continually refreshed like DRAM. Each cell will maintain its data as long as it has power. Without the need for constant refreshing, SRAM can operate extremely quickly. But the complexity of each cell make it prohibitively expensive for use as standard RAM.
The SRAM in the cache can be asynchronous or synchronous. Synchronous SRAM is designed to exactly match the speed of the CPU, while asynchronous is not. That little bit of timing makes a difference in performance. Matching the CPU's clock speed is a good thing, so always look for synchronized SRAM. (For more information on the various types of RAM, see How RAM Works.)
The final step in memory is the registers. These are memory cells built right into the CPU that contain specific data needed by the CPU, particularly the arithmetic and logic unit (ALU). An integral part of the CPU itself, they are controlled directly by the compiler that sends information for the CPU to process. See How Microprocessors Work for details on registers.
For a handy printable guide to computer memory, you can print the HowStuffWorks Big List of Computer Memory Terms.
For more information on computer memory and related topics, check out the links on the next page.
Computer :: Super :: Supercomputers will reach 'exascale' speeds within decade
Supercomputers powerful enough to simulate the human brain will be developed within seven years,
a leading supercomputer expert has predicted.
Professor Hans Werner Meuer, co-founder of the Top500 list of supercomputers, said that computers capable of processing a million trillion calculations a second would be available by 2019.
These powerful machines would open new possibilities for simulation in medicine, weather forecasting, and aerospace design, Meuer (pictured left) told a meeting of computer specialists at the House of Lords.
“Moore’s Law says you double performance every 18 months. The speed of the number one performing supercomputer is doubling every 13.2 months,” he said.
Within the next 10 years, all supercomputers in the Top500 list will be operating at least at petaflop speed - a quadrillion floating-point calculations per second.
And the most powerful computers, operating at exascale speeds, a thousand times faster than current machines - would be powerful enough to simulate the operation of the human brain, he said.
Scientists expect to be able to simulate a rat’s brain, consisting of 100 million brain cells, within two years, Meuer revealed.
“The end of the story is a human brain. That is [equivalent to] 1000 rat brains. That is where they need exascale computers,” he said.
The technology is trickling down rapidly into consumer products. Within 13 years, consumer devices will be as powerful as today’s supercomputers, Meuer predicted.
The Cray II ,the most powerful supercomputer in 1986,cost $22m.
“Twenty-five years later, Apple came up with the iPad2. Like it or not, it's very close to the performance of the Cray II” he said.
But Meuer said it was unclear whether supercomputers could continue to increase in speed at the current rate, as Moore’s Law began to reach its natural limit.
“We are reaching the critical end, but when this end will be will be is still open. No one knows,” he said.
Technologies, such as quantum computing, could give supercomputers a new lease of life, with IBM predicting the first workable machines within 15 years, he said.
New types of supercomputers capable of crunching huge quantities of data rather than numbers, may also emerge to meet the demand for data analytics.
The topic is an area for debate among computer scientists, said Meuer: “It is not yet clear whether we will need new infrastructure for that.”
Professor Hans Werner Meuer, co-founder of the Top500 list of supercomputers, said that computers capable of processing a million trillion calculations a second would be available by 2019.
These powerful machines would open new possibilities for simulation in medicine, weather forecasting, and aerospace design, Meuer (pictured left) told a meeting of computer specialists at the House of Lords.
“Moore’s Law says you double performance every 18 months. The speed of the number one performing supercomputer is doubling every 13.2 months,” he said.
Within the next 10 years, all supercomputers in the Top500 list will be operating at least at petaflop speed - a quadrillion floating-point calculations per second.
And the most powerful computers, operating at exascale speeds, a thousand times faster than current machines - would be powerful enough to simulate the operation of the human brain, he said.
Scientists expect to be able to simulate a rat’s brain, consisting of 100 million brain cells, within two years, Meuer revealed.
“The end of the story is a human brain. That is [equivalent to] 1000 rat brains. That is where they need exascale computers,” he said.
Supercomputers double in speed every 13.2 months
These machines will be able forecast weather in areas as small as 10km, and could be used to develop highly accurate simulations of aircraft aerodynamics, or make breakthroughs in medicine, the meeting heard.The technology is trickling down rapidly into consumer products. Within 13 years, consumer devices will be as powerful as today’s supercomputers, Meuer predicted.
The Cray II ,the most powerful supercomputer in 1986,cost $22m.
“Twenty-five years later, Apple came up with the iPad2. Like it or not, it's very close to the performance of the Cray II” he said.
But Meuer said it was unclear whether supercomputers could continue to increase in speed at the current rate, as Moore’s Law began to reach its natural limit.
“We are reaching the critical end, but when this end will be will be is still open. No one knows,” he said.
Technologies, such as quantum computing, could give supercomputers a new lease of life, with IBM predicting the first workable machines within 15 years, he said.
New types of supercomputers capable of crunching huge quantities of data rather than numbers, may also emerge to meet the demand for data analytics.
The topic is an area for debate among computer scientists, said Meuer: “It is not yet clear whether we will need new infrastructure for that.”
Subscribe to:
Posts (Atom)