Download Block Diagram-Parallel Processing-Assignments and more Exercises Parallel Computing and Programming in PDF only on Docsity!
Question # 01:
Com- puter
Proce- ssor Speed (GHz)
Country Interconnec t Family
Application Area
Vendor Proce- ssor Family
Operating System Family
Total Number Of Cores
Cores Per Socket
Power 5 Power 8C
3.83 UK Custom Interconnec t
Research IBM IBM MPP
Unix 8192 8
Power 2C
4.7 Germany Infiniband Weather & climate research
IBM IBM
MPP
Unix 8064 2
Xenon E 4C
3 US Infiniband Research SGI SGI Cluster
Linux 111104 4
PowerX Cell 8i 9C
3.2 US Infiniband Not Specified
IBM IBM
Cluster
Linux 122400 9
Xenon X
3.33 US Infiniband Academic Dell Cluster Linux 22656 6
Xenon X 6C
3.06 France Infiniband Industry Hewlett
Packar d
Cluster Linux 24192 6
Xenon X56xx
3.47 US Infiniband Research SGI SGI Cluster
Linux 27648 6
Power 8C
3.86 US Gigabit Ethernet
Defense IBM IBM MPP
Linux 2560 8
Intel/Xe non EM64T
3.6 US Infiniband Research Dell Dell Cluster
Linux 9024 1
Power 5 P
4.7 Italy Infiniband Not Specified
IBM IBM
MPP
Linux 5376 2
Question # 02:
1 )-Pipeline Architecture:
-24 stage pipeline -A15 cortex pipeline
3 )- VLIW Architecture:
Architectural Features in VLIW Processors
- VLIW processors rely on the compiler to identify a packet of instructions.
- Simply issues the packet because no dependence checking in hardware.
- Scheduling instructions is the responsibility of compiler so that their dependences are satisfied.
- Architectural support helps compiler exploit more ILP.
Question# 03: UMA Intel new core i
Features
. Extra performance on demand
. Four-way multitasking
. L3 cache
MPP
Cisco ASR 1000 Series Aggregation Service Router
Features
- First revolutionary Cisco Quantum Flow Processor (QFP)
- Lights-out remote-management facility
- These routers reuse the investment made in network I/O by using Cisco shared adaptors
- Flexible and faster services delivery based on the Cisco QFP
CLUSTER SYSTEM
Features
. Ability to reduce the frequency of system failure
. Ability to provide end user access to a service for a high
percentage of scheduled time while attempting to reduce unscheduled outages
. Ability to add resources and computers while attempting to
improve performance
