Advantages of Tecal

X8000: high-availability Hierarchical Management

X8000, as a fist product of, has been successfully applied to data centers of large Internet enterprises for a long time. The X8000 cabinet is divided into three layers of Management module, RMC (Racks Management Controller) Cabinet Management Controller of the entire cabinet, and 4U large nodes (up to 8 XH server nodes can be installed) to the single-server BMC (Baseboard Management Controller) Controller. This communication level design ensures high availability in management. For example, even if an MMC located in the "middle layer" fails, communication between other MMC and RMC will not be affected, while MMC is a bridge connecting server BMC with a single node. From this point of view, the X8000 design ensures the high availability of system management and reduces the number of switches used to connect to the server management port-which can save a lot of ports. In general, it increases the computing density in a limited space. At the same time, the Management port of the server itself can also be connected to the switch, because some users have their own management system and need to use the node's BMC, so there is no conflict with the previous management method.

X8000: effectively reduces energy consumption and facilitates plug-and-play by customers

It not only achieves unprecedented computing density, but also provides unified power supply and cooling, which can save about 10% of the overall power consumption. This is self-evident in the temptation to save energy costs for large-scale data centers. X8000 meets the customer's requirements for product delivery time, greatly facilitating the customer's rapid deployment and installation. In addition, you can customize different numbers and specifications of servers and network devices as needed. The modular design also facilitates the flexible and Fast replacement and deployment of the "plug-and-play. At the same time, supporting the previous generation of products also protects users' investment.

Powered by E5 AVX, the computing performance of Tecal V2 has soared

We also tested the last generation of products based on the intel 5600 series. We found that the floating point computing performance of the next generation of Tecal V2 servers has been greatly improved, it is almost twice that of the previous generation. The theoretical floating point peak value is the maximum number of times that the computer can theoretically complete floating point computing per second. It is mainly determined by the CPU clock speed, theoretical floating point peak = CPU clock speed × Number of floating point operations performed by the CPU in each clock cycle × Number of CPU cores in the system. In this test, the theoretical floating point peak value is 2.4 GHz * 8*16 = 307.2 GFLOPS, and the actual test data is 279.6097 GFLOPS. It is about 91% of the theoretical peak.

JAVA performance has soared

The test environment is Novell's SUSE Linux 11. The processor is a dual 8-core 16 thread clock speed of GHz Turbo to GHz with 20 mb level 3 cache E5-2650 engineering sample. SPECjvm2008 is a benchmark test suite for observing the Running Performance of JRE. Its test cases cover most basic java application scenarios and are a tool for architecture selection and JVM performance evaluation. SPECjvm2008 test case list is measured in ops/m operations per minute. The previous generation of RH2285 is configured with a dual-circuit E5645 processor. The 6-Core 12-thread 2.67 GHz clock speed can be Turbo to GHz. It has 12 Mb Level 3 cache. In JAVA integrated performance testing, the performance of the new generation of E5-2650 engineering samples is about 125% of the performance of the E5645 processor.