Intel Phi 7120D SC7120D User Manual

The Intel Xeon 
Phi Coprocessor

Even Higher Efficiency for Parallel Processing

While a majority of applications will continue to achieve 
maximum performance using Intel Xeon processors, 
certain highly-parallel applications will benefit dramatically 
by using Intel Xeon Phi coprocessors. Each coprocessor 
features many more and smaller cores, many more threads, 
and wider vector units. The high degree of parallelism 
compensates for the lower speed of each individual core  
to deliver higher aggregate performance for highly- 
parallel code.

You can use Intel Xeon processors and Intel Xeon Phi 
coprocessors together to optimize performance for almost 
any workload. To take full advantage of Intel Xeon Phi 
coprocessors, an application must scale well to over one-
hundred threads, and either make extensive use of vectors 
or efficiently use more local memory bandwidth than is 
available on an Intel Xeon processor. Learn more at  

http://software.intel.com/en-us/articles/is-intelr-xeon-
phitm-coprocessor-right-for-you

.

A Single Programming Model for All Your Code

A broad ecosystem of programming languages, models,  
and tools support Intel® architecture and all of them can  
be used with both Intel Xeon processors and Intel Xeon Phi 
coprocessors. Applications that run on one processor family 
will run on the other. This uniformity can greatly reduce the 
complexity of software development. Existing applications 
will need to be tuned and recompiled to maximize 
throughput, but your developers won’t need to rethink 
the entire problem or master new tools and programming 
models. Instead, they can reuse existing code and maintain  
a common code base using familiar tools and methods.

Code can be optimized just once for both Intel Xeon 
processors and Intel Xeon Phi coprocessors. The same 
techniques deliver optimal performance for both, so the 
investment you make in parallelizing your code will deliver 
benefits across the full range of computing environments.

A Family of Coprocessors for Diverse Needs

Intel Xeon Phi coprocessors provide up to 61 cores, 244 
threads, and 1.2 teraflops of performance, and they 
come in a variety of configurations to address diverse 
hardware, software, workload, performance, and efficiency 
requirements.

1

 They also come in a variety of form factors, 

including a standard PCIe* x16 form factor (with active, 
passive, or no thermal solution), and a dense form factor 
that offers additional design flexibility (Table 2). 
•  The Intel® Xeon Phi™ Coprocessor 3100 family provides 

outstanding parallel performance. It is an excellent choice 
for compute-bound workloads, such as MonteCarlo, 
black-Scholes, HPL, LifeSc, and many others.  Active and 
passive cooling options provide flexible support for a 
variety of server and workstation systems.

•  The Intel® Xeon Phi™ Coprocessor 5100 family is 

optimized for high-density computing and is well-suited 
for workloads that are memory-bandwidth bound, such 
as STREAM, memory-capacity bound, such as ray-tracing, 
or both, such as reverse time migration (RTM). These 
coprocessors are passively cooled and have the lowest 
thermal design power (TDP) of the Intel Xeon Phi  
product family.

•  The Intel® Xeon Phi™ Coprocessor 7100 family provides 

the most features and the highest performance and 
memory capacity of the Intel Xeon Phi product family. 
This family supports Intel® Turbo boost Technology 1.0, 
which increases core frequencies during peak workloads 
when thermal conditions allow. Passive and no thermal 
solution options enable powerful and innovative 
computing solutions.

2