In this presentation, Doug Miles from Nvidia describes the new features and performance gains in the PGI 2014 release.
“The use of accelerators in high performance computing is now mainstream,” said Douglas Miles, director of PGI Software at Nvidia. “With PGI 2014, we are taking another big step toward our goal of providing platform-independent, multi-core and accelerator programming tools that deliver outstanding performance on multiple platforms without the need for extensive, device-specific tuning.”
Watch the video presentation: http://insidehpc.com/2014/02/14/slidecast-new-pgi-2014-release-adds-openacc-2-0-features-x64-performance-gains/
2. PGI 2014 Multi-core x64 Highlights
180%
175%
170%
165%
160%
155%
150%
OpenMP almost 80%
faster than GCC on Xeon
Comprehensive
MPI features
Free PGI for your
MacBook
2
3. Industry-leading Multi-core x86 Performance
AMD Piledriver 2p/32 cores
Intel Sandy Bridge 2p/16 cores (hyperthreading enabled)
GCC 4.8.1
Intel 14.0.1
PGI 14.1
0%
20%
40%
60%
80%
100%
120%
140%
160%
180%
SPECompG_base2012 relative performance as measured by The Portland Group during the week of January 27, 2014. The number of OpenMP threads
was set to match the number of cores on each system. SPEComp® is a registered trademark of the Standard Performance Evaluation Corporation
(SPEC).
200%
3
4. PGI Accelerator 2014 Highlights
NVIDIA Tesla K40
and AMD Radeon
GPUs Support
OpenACC 2.0 Features
and Optimizations
CUDA Fortran and
OpenACC
Debugging
4
5. OpenACC Performance Portability
Speed-up
Average Speed-up Across 16 Benchmarks on Tesla and Radeon
16
14
12
10
8
6
4
2
0
CPU
GPU A
GPU B
Platform
CPU results are one core of an Intel Core i7-3930 CPU @ 3.20GHz (Sandy Bridge).
5
6. OpenACC 2.0 New Features
Procedure Calls in Accelerator Regions
Unstructured Data Lifetimes
Comprehensive Support for GPUDirect
And more …
6
8. Summary PGI 2014 Key Benefits
Multi-core x86 performance 80% higher than GCC on SPEC OMP 2012
OpenACC Fortran/C/C++ on both NVIDIA Tesla and AMD Radeon GPUs
OpenACC 2.0 features and optimizations
CUDA Fortran and OpenACC graphical debugging with Allinea DDT
FreePGI OpenMP 3.1 Fortran/C/Debugger for MacBooks
8