AZORES
Power Member
http://www.youtube.com/watch?v=YZnb.../Siggraph-2008-Rome-Reborn-Cell-B-E-Demo.html
Siggraph 2008 Demo of Rome Reborn model v1.1 running on the IBM QS21 Cell/B.E. Blades using iRT ray tracer
Cumpr´z
EDIT:
IBM QS21 Cell/B.E. Blades
For many years, organizations have relied on performance gains from increasing clock speeds of “traditional” microprocessor architectures. However, in recent years this approach has been challenged by the physical limitations of semiconductors and by traditional processor architecture implementations. Issues with power consumption, heat dissipation and memory latencies have led to diminishing returns on performance. High performance computing (HPC) cluster applications such as digital content creation, digital video surveillance, electronic design automation, image and signal processing, financial algorithms, scientific research, and seismic processing, may need a fundamentally new technology and approach to the system-level architecture to achieve the desired level of performance.
Cell/B.E. multi-core innovation
Enter the IBM BladeCenter® QS21 based on the multi-core Cell Broadband Engine (Cell/B.E.) processor. When used to complement systems based on traditional processors, the QS21 can yield application results faster and with more fidelity. This can enable organizations to get information faster to facilitate important business decisions.
Product features
Hardware summary
ESPECIFICAÇÕES:
Form factor
Single-wide blade server for BladeCenter
Processors
3.2 GHz Cell/B.E. Processors
Number of processors
Two standard, each with one PPE core and eight SPE cores
L2 cache
512 KB per Cell/B.E. Processor, plus 256 KB of local store memory for each SPE
Memory
2 GB (1 GB per processor)
Networking
Dual Gigabit EthernetI/O upgrade
Serial Attached SCSI (SAS) daughter card connected via PCI-X (CFFv)
Optional connectivity
Dual-port InfiniBand 4x HCA connected via PCI-Express (SFF)
Warranty
1-year
Siggraph 2008 Demo of Rome Reborn model v1.1 running on the IBM QS21 Cell/B.E. Blades using iRT ray tracer
Cumpr´z
EDIT:
IBM QS21 Cell/B.E. Blades
- Second generation blade system based on Cell Broadband Engine™ Processors
- Potential 10x or greater performance acceleration of target workloads
- High-performance blade solution for graphics-intensive and high performance computing (HPC) applications
For many years, organizations have relied on performance gains from increasing clock speeds of “traditional” microprocessor architectures. However, in recent years this approach has been challenged by the physical limitations of semiconductors and by traditional processor architecture implementations. Issues with power consumption, heat dissipation and memory latencies have led to diminishing returns on performance. High performance computing (HPC) cluster applications such as digital content creation, digital video surveillance, electronic design automation, image and signal processing, financial algorithms, scientific research, and seismic processing, may need a fundamentally new technology and approach to the system-level architecture to achieve the desired level of performance.
Cell/B.E. multi-core innovation
Enter the IBM BladeCenter® QS21 based on the multi-core Cell Broadband Engine (Cell/B.E.) processor. When used to complement systems based on traditional processors, the QS21 can yield application results faster and with more fidelity. This can enable organizations to get information faster to facilitate important business decisions.
Product features
- Cell/B.E. processors offer extreme performance to accelerate compute-intensive tasks such as image processing, signal processing and graphics rendering
- Single-wide blade form factor provides extraordinary computing density—up to 6.4 TFLOPS in a single BladeCenter H chassis
- Build a complex multi-platform environment in a highly efficient architecture
- Features two high-performance Cell/B.E. processors
- 2 GB standard processor memory
- Dual Gigabit Ethernet
- Optional Dual-port 4x InfiniBand HCA connected through PCI-Express
- Optional Serial Attached SCSI daughter cards connected through PCI-X
- Optional I/O buffer memory DIMMs (up to 2 GB, 4x512 MB)
ESPECIFICAÇÕES:
Form factor
Processors
Number of processors
L2 cache
Memory
Networking
Optional connectivity
Warranty
Última edição: