AMD
is due to give the first public demonstration of their
second-generation Opteron X-Series server processor, code-named
“Berlin”, at the Red Hat Summit, San Francisco, US on Wednesday. The demo will consist of X2100 Series Opteron that will run a Linux environment, which is based on the Fedora Project.
Berlin is what AMD calls an APU, an accelerated processing unit. This is basically a mash-up of CPU and GPU cores, which work together in a heterogeneous system architecture (HSA) using a shared memory. The compute and graphics cores in the HSA-enabled X2100 operate together by means of what is called, heterogeneous unified memory architecture or (hUMA) and heterogeneous queueing (hQ). This enable threads to run simultaneously and independently on all of the CPU and GPU cores. It shares the same memory space so that the CPU does not have to become occupied with feeding the GPU.
AMD’s first HSA APU was called Kaveri, which was released back in
January. AMD CTO Joe Macri said at a briefing that accompanied that
release, in the HSA world, it was time to get rid of terms like CPU and
GPU. “We’re going to need to create a new term, because HSA has really
changed what can be done on the graphics part of the die.” His
suggestion was to call them both compute cores.
Berlin is a server-level part and as such it is not designed for the video, gaming and other consumer-level user models at which Kaveri is intended to excel; The Kaveri chip is aimed at notebook and desktop PC’s.
AMD’s server unit general manager Suresh Gopalakrishnan said, “As servers adapt to new and evolving workloads…it’s critical that the software ecosystem support the requirements of these new workloads. “We are actively engaged with a broad set of partners in the data center software community who are bringing to market the software infrastructure to seamlessly enable x86 APU based servers.”
As always, if you would like to leave a sensible comment, then please do so in the comments section below.
Berlin is what AMD calls an APU, an accelerated processing unit. This is basically a mash-up of CPU and GPU cores, which work together in a heterogeneous system architecture (HSA) using a shared memory. The compute and graphics cores in the HSA-enabled X2100 operate together by means of what is called, heterogeneous unified memory architecture or (hUMA) and heterogeneous queueing (hQ). This enable threads to run simultaneously and independently on all of the CPU and GPU cores. It shares the same memory space so that the CPU does not have to become occupied with feeding the GPU.
Berlin is a server-level part and as such it is not designed for the video, gaming and other consumer-level user models at which Kaveri is intended to excel; The Kaveri chip is aimed at notebook and desktop PC’s.
AMD’s server unit general manager Suresh Gopalakrishnan said, “As servers adapt to new and evolving workloads…it’s critical that the software ecosystem support the requirements of these new workloads. “We are actively engaged with a broad set of partners in the data center software community who are bringing to market the software infrastructure to seamlessly enable x86 APU based servers.”
As always, if you would like to leave a sensible comment, then please do so in the comments section below.
No comments:
Post a Comment