Today’s multi-core processors
promise to deliver significant performance benefits to the
enterprise. As the number of cores per processor increase,
application performance is limited by the
processor’s memory bandwidth. Within large clusters
running many jobs, the Operating System (OS) at times places too many jobs on the same core resulting in memory bandwidth
over-subscription.
In most workload management software solutions,
the OS assigns jobs to cores or CPUs. In heterogeneous cluster environments,
not all OS's do a proper job of job assignment, nor are all
applications optimized or updated to take advantage of
multi-core processors. This results in degraded application
performance, forcing users to turn off the additional cores in
their quad-core machines and ultimately, under-utilizing expensive
hardware resources.
Join HP, Platform Computing, and Intel for a
1-hour live webinar and learn how to:
- Get maximum return for your IT
investments
- Provide resource flexibility for your
users
- Increase efficiency and output from your
IT resources
- Get increased visibility and effective
usage of multi-core high performance systems
- Create greater system density, allowing
you to maximize the productivity of available floor space
requiring less power and generate less heat per core
In addition, you will:
- Determine whether your applications are
ready for multi-core
- Find out what steps you can take to be
ready for multi-core
- Learn more about HP's Multi-core
Optimization Program
Attend this webinar and receive a FREE
whitepaper:
“Scheduling to Overcome the Multi-Core
Memory Bandwidth Bottleneck”
Who Should Attend:
-
Managers and Directors for: IS, IT, CAD, Design
-
|
|
Speakers: |
David Korf,
HP Marketing for Scalable Computing Infrastructure,
HP
|
|
Jim Prelack,
Worldwide Channel Marketing Manager, Intel
|
Ashar Baig,
Sr. Product Marketing Manager,
Platform Computing
|
|
|
|
