SlideShare a Scribd company logo
GPGPU on ARM
Tom Gall, Gil Pitney, 30th Oct 2013
connect.linaro.org
• This session will discuss the current state of the art of GPGPU
technologies on ARM SoC systems.
• What standards are there? Where are the drivers? What is the
best hardware? What are the problem areas? What open source
software has been accelerated? What opportunities for taking
advantage of GPGPU computing on ARM exist?
• There will also be some time for interactive discussion about
current and potential plans.
Session Description
connect.linaro.org
• At previous Connect (Dublin) discussed GPGPU, and what Linaro should
do in that area.
• Recognized that:
GPUs are much more energy efficient for highly parallel computing problems than
using just the main processor, even with NEON SIMD.
Big performance gains (> 20X) using the GPU for certain open source workloads.
• Decided:
• to get hardware and start gaining some experience with GPGPU drivers;
• it would be good to accelerate some OSS projects with GPGPU.
• So, we formed a GPGPU sub-team as part of the GWG.
GWG: New GPGPU Sub-group
connect.linaro.org
• Graphics pipeline operations migrate from the CPU to GPU:
– ’80’s: CPU + framebuffers with blitters.
– ’90’s: GPUs output one pixel per clock cycle, add pipeline operations:
• texture mapping, z-buffering, rasterization;
• transform and lighting (still fixed hardware) operations.
• Fixed functions become programmable kernels, adding multiple pipelines:
– 2000’s: GPUs become programmable, multicore:
• Programmable kernels (shaders) added to the per-pixel and vertex processing stages.
• More floating point precision, more GPU memory
• Unified shaders (eg: GeForce 8 “Streaming Multiprocessor”) handle vertex, pixel and geometry computations.
• GPU: initially a single core, fixed function hardware pipeline implementation
designed solely to accelerate graphics becomes…
• GPGPU: set of parallel and highly programmable cores (shaders) which can be
leveraged for more general purpose computation.
GPU to GPGPU: Hardware Evolution
connect.linaro.org
• Software co-evolves with GPU hardware.
– ’90’s: Graphics Languages: OpenGL (led GPU hardware features), DirectX.
– 2000: Shading Languages: OpenGL/GLSL, Direct3D/HLSL, Cg, others...
• Shading Languages: Need to understand graphics pipeline ops.
• But, shaders basically perform matrix and vector operations, ideal
for many non-graphics, scientific applications.
• Enter higher level parallel computing “stream processing”
languages:
– 2007: CUDA “Compute Unified Device Architecture” (NVIDIA)
– 2008: OpenCL “Open Compute Language” (Apple/Khronos)
GPU to GPGPU: Software Evolution
connect.linaro.org
• Ideal GPGPU applications have large data sets, high parallelism, and minimal
dependency between data elements:
– Advanced imaging: (Face recognition, cloth simulation, object detection).
– Cryptography searches: (Bitcoin, distributed.net).
– Video decode and post-processing: (iDCT, motion compensation, VLD, IQ, edge enhancement)
– Bioinformatics: (BLAST (protein and genome sequence comparisons), Folding@Home).
– Genetic Algorithms, Weather prediction, Particle Physics simulations
– MapReduce algorithm.
• The “13 Dwarfs of GP/GPU computing” [2] and benchmarks [1] evaluate parallel
programming models and hardware architectures for algorithms sharing similar
computation and inter-processor communication patterns:
- Computationally limited: Dense linear algebra, N-Body methods, cryptography.
- Memory bandwidth limited: Sparse linear algebra, structured grid.
- Memory latency limited: Graph traversal, dynamic programming, unstructured grid, FFT.
GPGPU Use Cases
connect.linaro.org
• Hardware:
– HSA (Heterogeneous System Architecture) foundation creates hardware and software standards.
Implementations coming…. (AMD/Kaveri, PS4: 4Q13?). Hardware must be certified HSA compliant.
– HSA requires one fully coherent shared memory heap, with unified addressing, between CPUs and
Compute Devices (i.e. a pointer can be passed between CPU and GPU directly). An HSA-specified
MMU (hMMU) allows the compute devices to share the page table mappings with host CPU(s).
• Software:
– OpenCL (Open Compute Language) is an open standard by the Khronos Group. One is free to
implement the standard and pay for certification. A large and growing number of problem domain
libraries and apps have been built on OpenCL.
– CUDA (Compute Unified Device Architecture) by NVidia is limited to NVidia hardware. There is no
standardization body and no certification. Historically CUDA was not able to be licensed but on June
18th, 2013 nVidia announced it was going to start licensing its GPU IP.
– Renderscript, part of Android, by Google. Renderscript is not a standard. There is no certification.
There is no implementation outside of Android.
– HSA tools: Compiler to HSAIL, HSAIL Finalizer to accelerator code, and HSA host runtime software.
Standards / Programming Platforms
connect.linaro.org
• Many OSS libraries support CUDA and/or OpenCL.
– Linear Algebra Libraries: ViennaCL, MAGMA, clMath(clBLAS), VexCL.
– Vision: OpenCV (Computer Vision) some modules written to CUDA/OpenCL.
– Database: sqLite on CUDA (20x-70x speedup for SELECT queries).
– Finance: QuantLib (faster option pricing for HFT, Monte Carlo simulations, and
company credit risk).
– Scientific Domains: OpenMM (Molecular Modeling), OpenBR (Biometric
Recognition).
• Language bindings ease programmability, proliferate GPU usage:
– WebCL: JavaScript binding to OpenCL, and Khronos standard. Demos have
shown (50x-100x) performance improvements over pure javascript web apps.
– Python: PyOpenCL, Java: Aparapi, Jogamp/JOCL.
Open Source Software accelerated with GPGPU
connect.linaro.org
• ZiiLabs:
– H/W: ZMS-40: Quad ARM Cortex A9 with 96 StemCell cores.
– S/W: OpenCL 1.1 (desktop profile) drivers from Creative for custom devices.
• Vivante:
– H/W: GC600+ series licensed to Marvell ARMADA, Freescale i.MX6, etc.
– S/W: OpenCL 1.1 EP drivers that work on FreeScale processors (SabreBoard).
• Qualcomm:
– H/W: Adreno 300+ series, in Snapdragon SoC’s.
– S/W: OpenCL 1.2 EP Android drivers in Adreno SDK (not shipped on device).
• ARM Mali:
– H/W: Samsung Exynos 4/5 series, many others. (eg: Arndale, Chromebook).
– S/W: Drivers for Exynos only seem to work on the Arndale-board when the LCD is also ordered; Chromebook Mali
drivers just made available (Oct 2013).
• Imagination Technologies:
– H/W: TI OMAP, Exynos 4/5 series with PowerVR SGX GPUs
– S/W: OpenCL 1.1 EP: TI requires NDA, Samsung ODROID-XU Android drivers avail (Linux drivers require NDA).
» Source: http://streamcomputing.eu/blog/2013-09-05/mobile-processor-opencl-drivers-q3-2013/
OpenCL Drivers on Various Hardware
connect.linaro.org
• Challenges
– GPGPU drivers are mostly closed, and only slowly becoming available.
– Google invested in Renderscript vs OpenCL on Android.
– HSA compliant hardware can extend the GPGPU into new problem domains by
increasing performance for memory latency constrained algorithms, but new hardware
designs take a long time to implement.
– Programming models like Renderscript/OpenCL still not so simple.
• Opportunities
– Performance: Select a set of key algorithm libraries to benchmark and accelerate.
– OpenCL: Invest in a CPU OSS version of OpenCL, and update to latest OpenCL spec.
– Simplify Programmability: Explore directive based models: eg: OpenACC/OpenMP 4.0 ?
Challenges and Opportunities
connect.linaro.org
• Address availability of OpenCL driver needs through the use of Clover (an
open source CPU only driver) and advance the use of non GPU devices such as
TI DSPs which will also be supported by Clover. Update Clover to OCL v1.2.
• Accelerate libjpeg-turbo : utilizing an existing OpenCL port, validate it works on
an ARM OpenCL implementation, and compare to a NEON optimized version.
• Accelerate sqlite : Sqlite is a database found in mobile and embedded solutions
for both Android and Linux. A past project accelerated it using CUDA. We
propose to accelerate the database using OpenCL.
• Accelerate OpenCV: OpenCV is an open source computer vision and machine
learning software library built to provide a common infrastructure for computer
vision applications and to accelerate the use of machine perception in the
commercial products. Accelerate using OpenCL.
GWG Roadmap: GPGPU Next Steps
More about Linaro: http://www.linaro.org/about/
More about Linaro engineering: http://www.linaro.org/engineering/
How to join: http://www.linaro.org/about/how-to-join
Linaro members: www.linaro.org/members
connect.linaro.org

More Related Content

What's hot

Hadoop + GPU
Hadoop + GPUHadoop + GPU
Hadoop + GPU
Vladimir Starostenkov
 
PostgreSQL with OpenCL
PostgreSQL with OpenCLPostgreSQL with OpenCL
PostgreSQL with OpenCL
Muhaza Liebenlito
 
PL-4047, Big Data Workload Analysis Using SWAT and Ipython Notebooks, by Moni...
PL-4047, Big Data Workload Analysis Using SWAT and Ipython Notebooks, by Moni...PL-4047, Big Data Workload Analysis Using SWAT and Ipython Notebooks, by Moni...
PL-4047, Big Data Workload Analysis Using SWAT and Ipython Notebooks, by Moni...
AMD Developer Central
 
PT-4053, Advanced OpenCL - Debugging and Profiling Using AMD CodeXL, by Uri S...
PT-4053, Advanced OpenCL - Debugging and Profiling Using AMD CodeXL, by Uri S...PT-4053, Advanced OpenCL - Debugging and Profiling Using AMD CodeXL, by Uri S...
PT-4053, Advanced OpenCL - Debugging and Profiling Using AMD CodeXL, by Uri S...
AMD Developer Central
 
The Small Batch (and other) solutions in Mantle API, by Guennadi Riguer, Mant...
The Small Batch (and other) solutions in Mantle API, by Guennadi Riguer, Mant...The Small Batch (and other) solutions in Mantle API, by Guennadi Riguer, Mant...
The Small Batch (and other) solutions in Mantle API, by Guennadi Riguer, Mant...
AMD Developer Central
 
PT-4058, Measuring and Optimizing Performance of Cluster and Private Cloud Ap...
PT-4058, Measuring and Optimizing Performance of Cluster and Private Cloud Ap...PT-4058, Measuring and Optimizing Performance of Cluster and Private Cloud Ap...
PT-4058, Measuring and Optimizing Performance of Cluster and Private Cloud Ap...
AMD Developer Central
 
PL-4048, Adapting languages for parallel processing on GPUs, by Neil Henning
PL-4048, Adapting languages for parallel processing on GPUs, by Neil HenningPL-4048, Adapting languages for parallel processing on GPUs, by Neil Henning
PL-4048, Adapting languages for parallel processing on GPUs, by Neil Henning
AMD Developer Central
 
GS-4150, Bullet 3 OpenCL Rigid Body Simulation, by Erwin Coumans
GS-4150, Bullet 3 OpenCL Rigid Body Simulation, by Erwin CoumansGS-4150, Bullet 3 OpenCL Rigid Body Simulation, by Erwin Coumans
GS-4150, Bullet 3 OpenCL Rigid Body Simulation, by Erwin Coumans
AMD Developer Central
 
PT-4142, Porting and Optimizing OpenMP applications to APU using CAPS tools, ...
PT-4142, Porting and Optimizing OpenMP applications to APU using CAPS tools, ...PT-4142, Porting and Optimizing OpenMP applications to APU using CAPS tools, ...
PT-4142, Porting and Optimizing OpenMP applications to APU using CAPS tools, ...
AMD Developer Central
 
PT-4102, Simulation, Compilation and Debugging of OpenCL on the AMD Southern ...
PT-4102, Simulation, Compilation and Debugging of OpenCL on the AMD Southern ...PT-4102, Simulation, Compilation and Debugging of OpenCL on the AMD Southern ...
PT-4102, Simulation, Compilation and Debugging of OpenCL on the AMD Southern ...
AMD Developer Central
 
Keynote (Phil Rogers) - The Programmers Guide to Reaching for the Cloud - by ...
Keynote (Phil Rogers) - The Programmers Guide to Reaching for the Cloud - by ...Keynote (Phil Rogers) - The Programmers Guide to Reaching for the Cloud - by ...
Keynote (Phil Rogers) - The Programmers Guide to Reaching for the Cloud - by ...
AMD Developer Central
 
CC-4001, Aparapi and HSA: Easing the developer path to APU/GPU accelerated Ja...
CC-4001, Aparapi and HSA: Easing the developer path to APU/GPU accelerated Ja...CC-4001, Aparapi and HSA: Easing the developer path to APU/GPU accelerated Ja...
CC-4001, Aparapi and HSA: Easing the developer path to APU/GPU accelerated Ja...
AMD Developer Central
 
HC-4020, Enhancing OpenCL performance in AfterShot Pro with HSA, by Michael W...
HC-4020, Enhancing OpenCL performance in AfterShot Pro with HSA, by Michael W...HC-4020, Enhancing OpenCL performance in AfterShot Pro with HSA, by Michael W...
HC-4020, Enhancing OpenCL performance in AfterShot Pro with HSA, by Michael W...
AMD Developer Central
 
HSA-4123, HSA Memory Model, by Ben Gaster
HSA-4123, HSA Memory Model, by Ben GasterHSA-4123, HSA Memory Model, by Ben Gaster
HSA-4123, HSA Memory Model, by Ben Gaster
AMD Developer Central
 
Keynote (Johan Andersson) - Mantle for Developers - by Johan Andersson, Techn...
Keynote (Johan Andersson) - Mantle for Developers - by Johan Andersson, Techn...Keynote (Johan Andersson) - Mantle for Developers - by Johan Andersson, Techn...
Keynote (Johan Andersson) - Mantle for Developers - by Johan Andersson, Techn...
AMD Developer Central
 
HC-4021, Efficient scheduling of OpenMP and OpenCL™ workloads on Accelerated ...
HC-4021, Efficient scheduling of OpenMP and OpenCL™ workloads on Accelerated ...HC-4021, Efficient scheduling of OpenMP and OpenCL™ workloads on Accelerated ...
HC-4021, Efficient scheduling of OpenMP and OpenCL™ workloads on Accelerated ...
AMD Developer Central
 
HC-4019, "Exploiting Coarse-grained Parallelism in B+ Tree Searches on an APU...
HC-4019, "Exploiting Coarse-grained Parallelism in B+ Tree Searches on an APU...HC-4019, "Exploiting Coarse-grained Parallelism in B+ Tree Searches on an APU...
HC-4019, "Exploiting Coarse-grained Parallelism in B+ Tree Searches on an APU...
AMD Developer Central
 
GPU Ecosystem
GPU EcosystemGPU Ecosystem
GPU Ecosystem
Ofer Rosenberg
 
MM-4105, Realtime 4K HDR Decoding with GPU ACES, by Gary Demos
MM-4105, Realtime 4K HDR Decoding with GPU ACES, by Gary DemosMM-4105, Realtime 4K HDR Decoding with GPU ACES, by Gary Demos
MM-4105, Realtime 4K HDR Decoding with GPU ACES, by Gary Demos
AMD Developer Central
 
Leverage the Speed of OpenCL™ with AMD Math Libraries
Leverage the Speed of OpenCL™ with AMD Math LibrariesLeverage the Speed of OpenCL™ with AMD Math Libraries
Leverage the Speed of OpenCL™ with AMD Math Libraries
AMD Developer Central
 

What's hot (20)

Hadoop + GPU
Hadoop + GPUHadoop + GPU
Hadoop + GPU
 
PostgreSQL with OpenCL
PostgreSQL with OpenCLPostgreSQL with OpenCL
PostgreSQL with OpenCL
 
PL-4047, Big Data Workload Analysis Using SWAT and Ipython Notebooks, by Moni...
PL-4047, Big Data Workload Analysis Using SWAT and Ipython Notebooks, by Moni...PL-4047, Big Data Workload Analysis Using SWAT and Ipython Notebooks, by Moni...
PL-4047, Big Data Workload Analysis Using SWAT and Ipython Notebooks, by Moni...
 
PT-4053, Advanced OpenCL - Debugging and Profiling Using AMD CodeXL, by Uri S...
PT-4053, Advanced OpenCL - Debugging and Profiling Using AMD CodeXL, by Uri S...PT-4053, Advanced OpenCL - Debugging and Profiling Using AMD CodeXL, by Uri S...
PT-4053, Advanced OpenCL - Debugging and Profiling Using AMD CodeXL, by Uri S...
 
The Small Batch (and other) solutions in Mantle API, by Guennadi Riguer, Mant...
The Small Batch (and other) solutions in Mantle API, by Guennadi Riguer, Mant...The Small Batch (and other) solutions in Mantle API, by Guennadi Riguer, Mant...
The Small Batch (and other) solutions in Mantle API, by Guennadi Riguer, Mant...
 
PT-4058, Measuring and Optimizing Performance of Cluster and Private Cloud Ap...
PT-4058, Measuring and Optimizing Performance of Cluster and Private Cloud Ap...PT-4058, Measuring and Optimizing Performance of Cluster and Private Cloud Ap...
PT-4058, Measuring and Optimizing Performance of Cluster and Private Cloud Ap...
 
PL-4048, Adapting languages for parallel processing on GPUs, by Neil Henning
PL-4048, Adapting languages for parallel processing on GPUs, by Neil HenningPL-4048, Adapting languages for parallel processing on GPUs, by Neil Henning
PL-4048, Adapting languages for parallel processing on GPUs, by Neil Henning
 
GS-4150, Bullet 3 OpenCL Rigid Body Simulation, by Erwin Coumans
GS-4150, Bullet 3 OpenCL Rigid Body Simulation, by Erwin CoumansGS-4150, Bullet 3 OpenCL Rigid Body Simulation, by Erwin Coumans
GS-4150, Bullet 3 OpenCL Rigid Body Simulation, by Erwin Coumans
 
PT-4142, Porting and Optimizing OpenMP applications to APU using CAPS tools, ...
PT-4142, Porting and Optimizing OpenMP applications to APU using CAPS tools, ...PT-4142, Porting and Optimizing OpenMP applications to APU using CAPS tools, ...
PT-4142, Porting and Optimizing OpenMP applications to APU using CAPS tools, ...
 
PT-4102, Simulation, Compilation and Debugging of OpenCL on the AMD Southern ...
PT-4102, Simulation, Compilation and Debugging of OpenCL on the AMD Southern ...PT-4102, Simulation, Compilation and Debugging of OpenCL on the AMD Southern ...
PT-4102, Simulation, Compilation and Debugging of OpenCL on the AMD Southern ...
 
Keynote (Phil Rogers) - The Programmers Guide to Reaching for the Cloud - by ...
Keynote (Phil Rogers) - The Programmers Guide to Reaching for the Cloud - by ...Keynote (Phil Rogers) - The Programmers Guide to Reaching for the Cloud - by ...
Keynote (Phil Rogers) - The Programmers Guide to Reaching for the Cloud - by ...
 
CC-4001, Aparapi and HSA: Easing the developer path to APU/GPU accelerated Ja...
CC-4001, Aparapi and HSA: Easing the developer path to APU/GPU accelerated Ja...CC-4001, Aparapi and HSA: Easing the developer path to APU/GPU accelerated Ja...
CC-4001, Aparapi and HSA: Easing the developer path to APU/GPU accelerated Ja...
 
HC-4020, Enhancing OpenCL performance in AfterShot Pro with HSA, by Michael W...
HC-4020, Enhancing OpenCL performance in AfterShot Pro with HSA, by Michael W...HC-4020, Enhancing OpenCL performance in AfterShot Pro with HSA, by Michael W...
HC-4020, Enhancing OpenCL performance in AfterShot Pro with HSA, by Michael W...
 
HSA-4123, HSA Memory Model, by Ben Gaster
HSA-4123, HSA Memory Model, by Ben GasterHSA-4123, HSA Memory Model, by Ben Gaster
HSA-4123, HSA Memory Model, by Ben Gaster
 
Keynote (Johan Andersson) - Mantle for Developers - by Johan Andersson, Techn...
Keynote (Johan Andersson) - Mantle for Developers - by Johan Andersson, Techn...Keynote (Johan Andersson) - Mantle for Developers - by Johan Andersson, Techn...
Keynote (Johan Andersson) - Mantle for Developers - by Johan Andersson, Techn...
 
HC-4021, Efficient scheduling of OpenMP and OpenCL™ workloads on Accelerated ...
HC-4021, Efficient scheduling of OpenMP and OpenCL™ workloads on Accelerated ...HC-4021, Efficient scheduling of OpenMP and OpenCL™ workloads on Accelerated ...
HC-4021, Efficient scheduling of OpenMP and OpenCL™ workloads on Accelerated ...
 
HC-4019, "Exploiting Coarse-grained Parallelism in B+ Tree Searches on an APU...
HC-4019, "Exploiting Coarse-grained Parallelism in B+ Tree Searches on an APU...HC-4019, "Exploiting Coarse-grained Parallelism in B+ Tree Searches on an APU...
HC-4019, "Exploiting Coarse-grained Parallelism in B+ Tree Searches on an APU...
 
GPU Ecosystem
GPU EcosystemGPU Ecosystem
GPU Ecosystem
 
MM-4105, Realtime 4K HDR Decoding with GPU ACES, by Gary Demos
MM-4105, Realtime 4K HDR Decoding with GPU ACES, by Gary DemosMM-4105, Realtime 4K HDR Decoding with GPU ACES, by Gary Demos
MM-4105, Realtime 4K HDR Decoding with GPU ACES, by Gary Demos
 
Leverage the Speed of OpenCL™ with AMD Math Libraries
Leverage the Speed of OpenCL™ with AMD Math LibrariesLeverage the Speed of OpenCL™ with AMD Math Libraries
Leverage the Speed of OpenCL™ with AMD Math Libraries
 

Viewers also liked

Optimized Rendering Techniques for Mobile VR
Optimized Rendering Techniques  for Mobile VROptimized Rendering Techniques  for Mobile VR
Optimized Rendering Techniques for Mobile VR
DevGAMM Conference
 
How to Build & Use OpenCL on Android Studio
How to Build & Use OpenCL on Android StudioHow to Build & Use OpenCL on Android Studio
Newbie’s guide to_the_gpgpu_universe
Newbie’s guide to_the_gpgpu_universeNewbie’s guide to_the_gpgpu_universe
Newbie’s guide to_the_gpgpu_universe
Ofer Rosenberg
 
PT-4057, Automated CUDA-to-OpenCL™ Translation with CU2CL: What's Next?, by W...
PT-4057, Automated CUDA-to-OpenCL™ Translation with CU2CL: What's Next?, by W...PT-4057, Automated CUDA-to-OpenCL™ Translation with CU2CL: What's Next?, by W...
PT-4057, Automated CUDA-to-OpenCL™ Translation with CU2CL: What's Next?, by W...
AMD Developer Central
 
CSTalks - GPGPU - 19 Jan
CSTalks  -  GPGPU - 19 JanCSTalks  -  GPGPU - 19 Jan
CSTalks - GPGPU - 19 Jan
cstalks
 
General Programming on the GPU - Confoo
General Programming on the GPU - ConfooGeneral Programming on the GPU - Confoo
General Programming on the GPU - Confoo
SirKetchup
 
Cliff sugerman
Cliff sugermanCliff sugerman
Cliff sugerman
clifford sugerman
 
[Harvard CS264] 06 - CUDA Ninja Tricks: GPU Scripting, Meta-programming & Aut...
[Harvard CS264] 06 - CUDA Ninja Tricks: GPU Scripting, Meta-programming & Aut...[Harvard CS264] 06 - CUDA Ninja Tricks: GPU Scripting, Meta-programming & Aut...
[Harvard CS264] 06 - CUDA Ninja Tricks: GPU Scripting, Meta-programming & Aut...
npinto
 
Advances in the Solution of Navier-Stokes Eqs. in GPGPU Hardware. Modelling F...
Advances in the Solution of Navier-Stokes Eqs. in GPGPU Hardware. Modelling F...Advances in the Solution of Navier-Stokes Eqs. in GPGPU Hardware. Modelling F...
Advances in the Solution of Navier-Stokes Eqs. in GPGPU Hardware. Modelling F...
Storti Mario
 
Gpgpu
GpgpuGpgpu
Gpgpu
Su Yan-Jen
 
Open CL For Haifa Linux Club
Open CL For Haifa Linux ClubOpen CL For Haifa Linux Club
Open CL For Haifa Linux ClubOfer Rosenberg
 
GPU Technology Conference 2014 Keynote
GPU Technology Conference 2014 KeynoteGPU Technology Conference 2014 Keynote
GPU Technology Conference 2014 Keynote
NVIDIA
 
Intro to GPGPU with CUDA (DevLink)
Intro to GPGPU with CUDA (DevLink)Intro to GPGPU with CUDA (DevLink)
Intro to GPGPU with CUDA (DevLink)
Rob Gillen
 
E-Learning: Introduction to GPGPU
E-Learning: Introduction to GPGPUE-Learning: Introduction to GPGPU
E-Learning: Introduction to GPGPU
Nur Ahmadi
 
Nvidia cuda tutorial_no_nda_apr08
Nvidia cuda tutorial_no_nda_apr08Nvidia cuda tutorial_no_nda_apr08
Nvidia cuda tutorial_no_nda_apr08Angela Mendoza M.
 
GPUDirect RDMA and Green Multi-GPU Architectures
GPUDirect RDMA and Green Multi-GPU ArchitecturesGPUDirect RDMA and Green Multi-GPU Architectures
GPUDirect RDMA and Green Multi-GPU Architectures
inside-BigData.com
 
Introduction to gpu architecture
Introduction to gpu architectureIntroduction to gpu architecture
Introduction to gpu architecture
CHIHTE LU
 
CS 354 GPU Architecture
CS 354 GPU ArchitectureCS 354 GPU Architecture
CS 354 GPU Architecture
Mark Kilgard
 
Introduction to OpenCL, 2010
Introduction to OpenCL, 2010Introduction to OpenCL, 2010
Introduction to OpenCL, 2010
Tomasz Bednarz
 

Viewers also liked (20)

Optimized Rendering Techniques for Mobile VR
Optimized Rendering Techniques  for Mobile VROptimized Rendering Techniques  for Mobile VR
Optimized Rendering Techniques for Mobile VR
 
How to Build & Use OpenCL on Android Studio
How to Build & Use OpenCL on Android StudioHow to Build & Use OpenCL on Android Studio
How to Build & Use OpenCL on Android Studio
 
Gpgpu intro
Gpgpu introGpgpu intro
Gpgpu intro
 
Newbie’s guide to_the_gpgpu_universe
Newbie’s guide to_the_gpgpu_universeNewbie’s guide to_the_gpgpu_universe
Newbie’s guide to_the_gpgpu_universe
 
PT-4057, Automated CUDA-to-OpenCL™ Translation with CU2CL: What's Next?, by W...
PT-4057, Automated CUDA-to-OpenCL™ Translation with CU2CL: What's Next?, by W...PT-4057, Automated CUDA-to-OpenCL™ Translation with CU2CL: What's Next?, by W...
PT-4057, Automated CUDA-to-OpenCL™ Translation with CU2CL: What's Next?, by W...
 
CSTalks - GPGPU - 19 Jan
CSTalks  -  GPGPU - 19 JanCSTalks  -  GPGPU - 19 Jan
CSTalks - GPGPU - 19 Jan
 
General Programming on the GPU - Confoo
General Programming on the GPU - ConfooGeneral Programming on the GPU - Confoo
General Programming on the GPU - Confoo
 
Cliff sugerman
Cliff sugermanCliff sugerman
Cliff sugerman
 
[Harvard CS264] 06 - CUDA Ninja Tricks: GPU Scripting, Meta-programming & Aut...
[Harvard CS264] 06 - CUDA Ninja Tricks: GPU Scripting, Meta-programming & Aut...[Harvard CS264] 06 - CUDA Ninja Tricks: GPU Scripting, Meta-programming & Aut...
[Harvard CS264] 06 - CUDA Ninja Tricks: GPU Scripting, Meta-programming & Aut...
 
Advances in the Solution of Navier-Stokes Eqs. in GPGPU Hardware. Modelling F...
Advances in the Solution of Navier-Stokes Eqs. in GPGPU Hardware. Modelling F...Advances in the Solution of Navier-Stokes Eqs. in GPGPU Hardware. Modelling F...
Advances in the Solution of Navier-Stokes Eqs. in GPGPU Hardware. Modelling F...
 
Gpgpu
GpgpuGpgpu
Gpgpu
 
Open CL For Haifa Linux Club
Open CL For Haifa Linux ClubOpen CL For Haifa Linux Club
Open CL For Haifa Linux Club
 
GPU Technology Conference 2014 Keynote
GPU Technology Conference 2014 KeynoteGPU Technology Conference 2014 Keynote
GPU Technology Conference 2014 Keynote
 
Intro to GPGPU with CUDA (DevLink)
Intro to GPGPU with CUDA (DevLink)Intro to GPGPU with CUDA (DevLink)
Intro to GPGPU with CUDA (DevLink)
 
E-Learning: Introduction to GPGPU
E-Learning: Introduction to GPGPUE-Learning: Introduction to GPGPU
E-Learning: Introduction to GPGPU
 
Nvidia cuda tutorial_no_nda_apr08
Nvidia cuda tutorial_no_nda_apr08Nvidia cuda tutorial_no_nda_apr08
Nvidia cuda tutorial_no_nda_apr08
 
GPUDirect RDMA and Green Multi-GPU Architectures
GPUDirect RDMA and Green Multi-GPU ArchitecturesGPUDirect RDMA and Green Multi-GPU Architectures
GPUDirect RDMA and Green Multi-GPU Architectures
 
Introduction to gpu architecture
Introduction to gpu architectureIntroduction to gpu architecture
Introduction to gpu architecture
 
CS 354 GPU Architecture
CS 354 GPU ArchitectureCS 354 GPU Architecture
CS 354 GPU Architecture
 
Introduction to OpenCL, 2010
Introduction to OpenCL, 2010Introduction to OpenCL, 2010
Introduction to OpenCL, 2010
 

Similar to LCU13: GPGPU on ARM Experience Report

The GPGPU Continuum
The GPGPU ContinuumThe GPGPU Continuum
The GPGPU Continuum
Ofer Rosenberg
 
"Making Computer Vision Software Run Fast on Your Embedded Platform," a Prese...
"Making Computer Vision Software Run Fast on Your Embedded Platform," a Prese..."Making Computer Vision Software Run Fast on Your Embedded Platform," a Prese...
"Making Computer Vision Software Run Fast on Your Embedded Platform," a Prese...
Edge AI and Vision Alliance
 
Hands on OpenCL
Hands on OpenCLHands on OpenCL
Hands on OpenCL
Vladimir Starostenkov
 
Deep Learning on ARM Platforms - SFO17-509
Deep Learning on ARM Platforms - SFO17-509Deep Learning on ARM Platforms - SFO17-509
Deep Learning on ARM Platforms - SFO17-509
Linaro
 
Making the most out of Heterogeneous Chips with CPU, GPU and FPGA
Making the most out of Heterogeneous Chips with CPU, GPU and FPGAMaking the most out of Heterogeneous Chips with CPU, GPU and FPGA
Making the most out of Heterogeneous Chips with CPU, GPU and FPGA
Facultad de Informática UCM
 
Debugging Numerical Simulations on Accelerated Architectures - TotalView fo...
 Debugging Numerical Simulations on Accelerated Architectures  - TotalView fo... Debugging Numerical Simulations on Accelerated Architectures  - TotalView fo...
Debugging Numerical Simulations on Accelerated Architectures - TotalView fo...
Rogue Wave Software
 
Utilizing AMD GPUs: Tuning, programming models, and roadmap
Utilizing AMD GPUs: Tuning, programming models, and roadmapUtilizing AMD GPUs: Tuning, programming models, and roadmap
Utilizing AMD GPUs: Tuning, programming models, and roadmap
George Markomanolis
 
"The Vision API Maze: Options and Trade-offs," a Presentation from the Khrono...
"The Vision API Maze: Options and Trade-offs," a Presentation from the Khrono..."The Vision API Maze: Options and Trade-offs," a Presentation from the Khrono...
"The Vision API Maze: Options and Trade-offs," a Presentation from the Khrono...
Edge AI and Vision Alliance
 
TRACK F: OpenCL for ALTERA FPGAs, Accelerating performance and design product...
TRACK F: OpenCL for ALTERA FPGAs, Accelerating performance and design product...TRACK F: OpenCL for ALTERA FPGAs, Accelerating performance and design product...
TRACK F: OpenCL for ALTERA FPGAs, Accelerating performance and design product...chiportal
 
AMD It's Time to ROC
AMD It's Time to ROCAMD It's Time to ROC
AMD It's Time to ROC
inside-BigData.com
 
Implementing AI: High Performace Architectures
Implementing AI: High Performace ArchitecturesImplementing AI: High Performace Architectures
Implementing AI: High Performace Architectures
KTN
 
Development of Signal Processing Algorithms using OpenCL for FPGA based Archi...
Development of Signal Processing Algorithms using OpenCL for FPGA based Archi...Development of Signal Processing Algorithms using OpenCL for FPGA based Archi...
Development of Signal Processing Algorithms using OpenCL for FPGA based Archi...Pradeep Singh
 
Массовый параллелизм для гетерогенных вычислений на C++ для беспилотных автом...
Массовый параллелизм для гетерогенных вычислений на C++ для беспилотных автом...Массовый параллелизм для гетерогенных вычислений на C++ для беспилотных автом...
Массовый параллелизм для гетерогенных вычислений на C++ для беспилотных автом...
CEE-SEC(R)
 
OpenACC Monthly Highlights September 2020
OpenACC Monthly Highlights September 2020OpenACC Monthly Highlights September 2020
OpenACC Monthly Highlights September 2020
OpenACC
 
OpenCL & the Future of Desktop High Performance Computing in CAD
OpenCL & the Future of Desktop High Performance Computing in CADOpenCL & the Future of Desktop High Performance Computing in CAD
OpenCL & the Future of Desktop High Performance Computing in CAD
Design World
 
AI Hardware Landscape 2021
AI Hardware Landscape 2021AI Hardware Landscape 2021
AI Hardware Landscape 2021
Grigory Sapunov
 
Backend.AI Technical Introduction (19.09 / 2019 Autumn)
Backend.AI Technical Introduction (19.09 / 2019 Autumn)Backend.AI Technical Introduction (19.09 / 2019 Autumn)
Backend.AI Technical Introduction (19.09 / 2019 Autumn)
Lablup Inc.
 
XPDDS17: Keynote: Shared Coprocessor Framework on ARM - Oleksandr Andrushchen...
XPDDS17: Keynote: Shared Coprocessor Framework on ARM - Oleksandr Andrushchen...XPDDS17: Keynote: Shared Coprocessor Framework on ARM - Oleksandr Andrushchen...
XPDDS17: Keynote: Shared Coprocessor Framework on ARM - Oleksandr Andrushchen...
The Linux Foundation
 
Exploring the Programming Models for the LUMI Supercomputer
Exploring the Programming Models for the LUMI Supercomputer Exploring the Programming Models for the LUMI Supercomputer
Exploring the Programming Models for the LUMI Supercomputer
George Markomanolis
 
"New Standards for Embedded Vision and Neural Networks," a Presentation from ...
"New Standards for Embedded Vision and Neural Networks," a Presentation from ..."New Standards for Embedded Vision and Neural Networks," a Presentation from ...
"New Standards for Embedded Vision and Neural Networks," a Presentation from ...
Edge AI and Vision Alliance
 

Similar to LCU13: GPGPU on ARM Experience Report (20)

The GPGPU Continuum
The GPGPU ContinuumThe GPGPU Continuum
The GPGPU Continuum
 
"Making Computer Vision Software Run Fast on Your Embedded Platform," a Prese...
"Making Computer Vision Software Run Fast on Your Embedded Platform," a Prese..."Making Computer Vision Software Run Fast on Your Embedded Platform," a Prese...
"Making Computer Vision Software Run Fast on Your Embedded Platform," a Prese...
 
Hands on OpenCL
Hands on OpenCLHands on OpenCL
Hands on OpenCL
 
Deep Learning on ARM Platforms - SFO17-509
Deep Learning on ARM Platforms - SFO17-509Deep Learning on ARM Platforms - SFO17-509
Deep Learning on ARM Platforms - SFO17-509
 
Making the most out of Heterogeneous Chips with CPU, GPU and FPGA
Making the most out of Heterogeneous Chips with CPU, GPU and FPGAMaking the most out of Heterogeneous Chips with CPU, GPU and FPGA
Making the most out of Heterogeneous Chips with CPU, GPU and FPGA
 
Debugging Numerical Simulations on Accelerated Architectures - TotalView fo...
 Debugging Numerical Simulations on Accelerated Architectures  - TotalView fo... Debugging Numerical Simulations on Accelerated Architectures  - TotalView fo...
Debugging Numerical Simulations on Accelerated Architectures - TotalView fo...
 
Utilizing AMD GPUs: Tuning, programming models, and roadmap
Utilizing AMD GPUs: Tuning, programming models, and roadmapUtilizing AMD GPUs: Tuning, programming models, and roadmap
Utilizing AMD GPUs: Tuning, programming models, and roadmap
 
"The Vision API Maze: Options and Trade-offs," a Presentation from the Khrono...
"The Vision API Maze: Options and Trade-offs," a Presentation from the Khrono..."The Vision API Maze: Options and Trade-offs," a Presentation from the Khrono...
"The Vision API Maze: Options and Trade-offs," a Presentation from the Khrono...
 
TRACK F: OpenCL for ALTERA FPGAs, Accelerating performance and design product...
TRACK F: OpenCL for ALTERA FPGAs, Accelerating performance and design product...TRACK F: OpenCL for ALTERA FPGAs, Accelerating performance and design product...
TRACK F: OpenCL for ALTERA FPGAs, Accelerating performance and design product...
 
AMD It's Time to ROC
AMD It's Time to ROCAMD It's Time to ROC
AMD It's Time to ROC
 
Implementing AI: High Performace Architectures
Implementing AI: High Performace ArchitecturesImplementing AI: High Performace Architectures
Implementing AI: High Performace Architectures
 
Development of Signal Processing Algorithms using OpenCL for FPGA based Archi...
Development of Signal Processing Algorithms using OpenCL for FPGA based Archi...Development of Signal Processing Algorithms using OpenCL for FPGA based Archi...
Development of Signal Processing Algorithms using OpenCL for FPGA based Archi...
 
Массовый параллелизм для гетерогенных вычислений на C++ для беспилотных автом...
Массовый параллелизм для гетерогенных вычислений на C++ для беспилотных автом...Массовый параллелизм для гетерогенных вычислений на C++ для беспилотных автом...
Массовый параллелизм для гетерогенных вычислений на C++ для беспилотных автом...
 
OpenACC Monthly Highlights September 2020
OpenACC Monthly Highlights September 2020OpenACC Monthly Highlights September 2020
OpenACC Monthly Highlights September 2020
 
OpenCL & the Future of Desktop High Performance Computing in CAD
OpenCL & the Future of Desktop High Performance Computing in CADOpenCL & the Future of Desktop High Performance Computing in CAD
OpenCL & the Future of Desktop High Performance Computing in CAD
 
AI Hardware Landscape 2021
AI Hardware Landscape 2021AI Hardware Landscape 2021
AI Hardware Landscape 2021
 
Backend.AI Technical Introduction (19.09 / 2019 Autumn)
Backend.AI Technical Introduction (19.09 / 2019 Autumn)Backend.AI Technical Introduction (19.09 / 2019 Autumn)
Backend.AI Technical Introduction (19.09 / 2019 Autumn)
 
XPDDS17: Keynote: Shared Coprocessor Framework on ARM - Oleksandr Andrushchen...
XPDDS17: Keynote: Shared Coprocessor Framework on ARM - Oleksandr Andrushchen...XPDDS17: Keynote: Shared Coprocessor Framework on ARM - Oleksandr Andrushchen...
XPDDS17: Keynote: Shared Coprocessor Framework on ARM - Oleksandr Andrushchen...
 
Exploring the Programming Models for the LUMI Supercomputer
Exploring the Programming Models for the LUMI Supercomputer Exploring the Programming Models for the LUMI Supercomputer
Exploring the Programming Models for the LUMI Supercomputer
 
"New Standards for Embedded Vision and Neural Networks," a Presentation from ...
"New Standards for Embedded Vision and Neural Networks," a Presentation from ..."New Standards for Embedded Vision and Neural Networks," a Presentation from ...
"New Standards for Embedded Vision and Neural Networks," a Presentation from ...
 

More from Linaro

Deep Learning Neural Network Acceleration at the Edge - Andrea Gallo
Deep Learning Neural Network Acceleration at the Edge - Andrea GalloDeep Learning Neural Network Acceleration at the Edge - Andrea Gallo
Deep Learning Neural Network Acceleration at the Edge - Andrea Gallo
Linaro
 
Arm Architecture HPC Workshop Santa Clara 2018 - Kanta Vekaria
Arm Architecture HPC Workshop Santa Clara 2018 - Kanta VekariaArm Architecture HPC Workshop Santa Clara 2018 - Kanta Vekaria
Arm Architecture HPC Workshop Santa Clara 2018 - Kanta Vekaria
Linaro
 
Huawei’s requirements for the ARM based HPC solution readiness - Joshua Mora
Huawei’s requirements for the ARM based HPC solution readiness - Joshua MoraHuawei’s requirements for the ARM based HPC solution readiness - Joshua Mora
Huawei’s requirements for the ARM based HPC solution readiness - Joshua Mora
Linaro
 
Bud17 113: distribution ci using qemu and open qa
Bud17 113: distribution ci using qemu and open qaBud17 113: distribution ci using qemu and open qa
Bud17 113: distribution ci using qemu and open qa
Linaro
 
OpenHPC Automation with Ansible - Renato Golin - Linaro Arm HPC Workshop 2018
OpenHPC Automation with Ansible - Renato Golin - Linaro Arm HPC Workshop 2018OpenHPC Automation with Ansible - Renato Golin - Linaro Arm HPC Workshop 2018
OpenHPC Automation with Ansible - Renato Golin - Linaro Arm HPC Workshop 2018
Linaro
 
HPC network stack on ARM - Linaro HPC Workshop 2018
HPC network stack on ARM - Linaro HPC Workshop 2018HPC network stack on ARM - Linaro HPC Workshop 2018
HPC network stack on ARM - Linaro HPC Workshop 2018
Linaro
 
It just keeps getting better - SUSE enablement for Arm - Linaro HPC Workshop ...
It just keeps getting better - SUSE enablement for Arm - Linaro HPC Workshop ...It just keeps getting better - SUSE enablement for Arm - Linaro HPC Workshop ...
It just keeps getting better - SUSE enablement for Arm - Linaro HPC Workshop ...
Linaro
 
Intelligent Interconnect Architecture to Enable Next Generation HPC - Linaro ...
Intelligent Interconnect Architecture to Enable Next Generation HPC - Linaro ...Intelligent Interconnect Architecture to Enable Next Generation HPC - Linaro ...
Intelligent Interconnect Architecture to Enable Next Generation HPC - Linaro ...
Linaro
 
Yutaka Ishikawa - Post-K and Arm HPC Ecosystem - Linaro Arm HPC Workshop Sant...
Yutaka Ishikawa - Post-K and Arm HPC Ecosystem - Linaro Arm HPC Workshop Sant...Yutaka Ishikawa - Post-K and Arm HPC Ecosystem - Linaro Arm HPC Workshop Sant...
Yutaka Ishikawa - Post-K and Arm HPC Ecosystem - Linaro Arm HPC Workshop Sant...
Linaro
 
Andrew J Younge - Vanguard Astra - Petascale Arm Platform for U.S. DOE/ASC Su...
Andrew J Younge - Vanguard Astra - Petascale Arm Platform for U.S. DOE/ASC Su...Andrew J Younge - Vanguard Astra - Petascale Arm Platform for U.S. DOE/ASC Su...
Andrew J Younge - Vanguard Astra - Petascale Arm Platform for U.S. DOE/ASC Su...
Linaro
 
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainlineHKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
Linaro
 
HKG18-100K1 - George Grey: Opening Keynote
HKG18-100K1 - George Grey: Opening KeynoteHKG18-100K1 - George Grey: Opening Keynote
HKG18-100K1 - George Grey: Opening Keynote
Linaro
 
HKG18-318 - OpenAMP Workshop
HKG18-318 - OpenAMP WorkshopHKG18-318 - OpenAMP Workshop
HKG18-318 - OpenAMP Workshop
Linaro
 
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainlineHKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
Linaro
 
HKG18-315 - Why the ecosystem is a wonderful thing, warts and all
HKG18-315 - Why the ecosystem is a wonderful thing, warts and allHKG18-315 - Why the ecosystem is a wonderful thing, warts and all
HKG18-315 - Why the ecosystem is a wonderful thing, warts and all
Linaro
 
HKG18- 115 - Partitioning ARM Systems with the Jailhouse Hypervisor
HKG18- 115 - Partitioning ARM Systems with the Jailhouse HypervisorHKG18- 115 - Partitioning ARM Systems with the Jailhouse Hypervisor
HKG18- 115 - Partitioning ARM Systems with the Jailhouse Hypervisor
Linaro
 
HKG18-TR08 - Upstreaming SVE in QEMU
HKG18-TR08 - Upstreaming SVE in QEMUHKG18-TR08 - Upstreaming SVE in QEMU
HKG18-TR08 - Upstreaming SVE in QEMU
Linaro
 
HKG18-113- Secure Data Path work with i.MX8M
HKG18-113- Secure Data Path work with i.MX8MHKG18-113- Secure Data Path work with i.MX8M
HKG18-113- Secure Data Path work with i.MX8M
Linaro
 
HKG18-120 - Devicetree Schema Documentation and Validation
HKG18-120 - Devicetree Schema Documentation and Validation HKG18-120 - Devicetree Schema Documentation and Validation
HKG18-120 - Devicetree Schema Documentation and Validation
Linaro
 
HKG18-223 - Trusted FirmwareM: Trusted boot
HKG18-223 - Trusted FirmwareM: Trusted bootHKG18-223 - Trusted FirmwareM: Trusted boot
HKG18-223 - Trusted FirmwareM: Trusted boot
Linaro
 

More from Linaro (20)

Deep Learning Neural Network Acceleration at the Edge - Andrea Gallo
Deep Learning Neural Network Acceleration at the Edge - Andrea GalloDeep Learning Neural Network Acceleration at the Edge - Andrea Gallo
Deep Learning Neural Network Acceleration at the Edge - Andrea Gallo
 
Arm Architecture HPC Workshop Santa Clara 2018 - Kanta Vekaria
Arm Architecture HPC Workshop Santa Clara 2018 - Kanta VekariaArm Architecture HPC Workshop Santa Clara 2018 - Kanta Vekaria
Arm Architecture HPC Workshop Santa Clara 2018 - Kanta Vekaria
 
Huawei’s requirements for the ARM based HPC solution readiness - Joshua Mora
Huawei’s requirements for the ARM based HPC solution readiness - Joshua MoraHuawei’s requirements for the ARM based HPC solution readiness - Joshua Mora
Huawei’s requirements for the ARM based HPC solution readiness - Joshua Mora
 
Bud17 113: distribution ci using qemu and open qa
Bud17 113: distribution ci using qemu and open qaBud17 113: distribution ci using qemu and open qa
Bud17 113: distribution ci using qemu and open qa
 
OpenHPC Automation with Ansible - Renato Golin - Linaro Arm HPC Workshop 2018
OpenHPC Automation with Ansible - Renato Golin - Linaro Arm HPC Workshop 2018OpenHPC Automation with Ansible - Renato Golin - Linaro Arm HPC Workshop 2018
OpenHPC Automation with Ansible - Renato Golin - Linaro Arm HPC Workshop 2018
 
HPC network stack on ARM - Linaro HPC Workshop 2018
HPC network stack on ARM - Linaro HPC Workshop 2018HPC network stack on ARM - Linaro HPC Workshop 2018
HPC network stack on ARM - Linaro HPC Workshop 2018
 
It just keeps getting better - SUSE enablement for Arm - Linaro HPC Workshop ...
It just keeps getting better - SUSE enablement for Arm - Linaro HPC Workshop ...It just keeps getting better - SUSE enablement for Arm - Linaro HPC Workshop ...
It just keeps getting better - SUSE enablement for Arm - Linaro HPC Workshop ...
 
Intelligent Interconnect Architecture to Enable Next Generation HPC - Linaro ...
Intelligent Interconnect Architecture to Enable Next Generation HPC - Linaro ...Intelligent Interconnect Architecture to Enable Next Generation HPC - Linaro ...
Intelligent Interconnect Architecture to Enable Next Generation HPC - Linaro ...
 
Yutaka Ishikawa - Post-K and Arm HPC Ecosystem - Linaro Arm HPC Workshop Sant...
Yutaka Ishikawa - Post-K and Arm HPC Ecosystem - Linaro Arm HPC Workshop Sant...Yutaka Ishikawa - Post-K and Arm HPC Ecosystem - Linaro Arm HPC Workshop Sant...
Yutaka Ishikawa - Post-K and Arm HPC Ecosystem - Linaro Arm HPC Workshop Sant...
 
Andrew J Younge - Vanguard Astra - Petascale Arm Platform for U.S. DOE/ASC Su...
Andrew J Younge - Vanguard Astra - Petascale Arm Platform for U.S. DOE/ASC Su...Andrew J Younge - Vanguard Astra - Petascale Arm Platform for U.S. DOE/ASC Su...
Andrew J Younge - Vanguard Astra - Petascale Arm Platform for U.S. DOE/ASC Su...
 
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainlineHKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
 
HKG18-100K1 - George Grey: Opening Keynote
HKG18-100K1 - George Grey: Opening KeynoteHKG18-100K1 - George Grey: Opening Keynote
HKG18-100K1 - George Grey: Opening Keynote
 
HKG18-318 - OpenAMP Workshop
HKG18-318 - OpenAMP WorkshopHKG18-318 - OpenAMP Workshop
HKG18-318 - OpenAMP Workshop
 
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainlineHKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
HKG18-501 - EAS on Common Kernel 4.14 and getting (much) closer to mainline
 
HKG18-315 - Why the ecosystem is a wonderful thing, warts and all
HKG18-315 - Why the ecosystem is a wonderful thing, warts and allHKG18-315 - Why the ecosystem is a wonderful thing, warts and all
HKG18-315 - Why the ecosystem is a wonderful thing, warts and all
 
HKG18- 115 - Partitioning ARM Systems with the Jailhouse Hypervisor
HKG18- 115 - Partitioning ARM Systems with the Jailhouse HypervisorHKG18- 115 - Partitioning ARM Systems with the Jailhouse Hypervisor
HKG18- 115 - Partitioning ARM Systems with the Jailhouse Hypervisor
 
HKG18-TR08 - Upstreaming SVE in QEMU
HKG18-TR08 - Upstreaming SVE in QEMUHKG18-TR08 - Upstreaming SVE in QEMU
HKG18-TR08 - Upstreaming SVE in QEMU
 
HKG18-113- Secure Data Path work with i.MX8M
HKG18-113- Secure Data Path work with i.MX8MHKG18-113- Secure Data Path work with i.MX8M
HKG18-113- Secure Data Path work with i.MX8M
 
HKG18-120 - Devicetree Schema Documentation and Validation
HKG18-120 - Devicetree Schema Documentation and Validation HKG18-120 - Devicetree Schema Documentation and Validation
HKG18-120 - Devicetree Schema Documentation and Validation
 
HKG18-223 - Trusted FirmwareM: Trusted boot
HKG18-223 - Trusted FirmwareM: Trusted bootHKG18-223 - Trusted FirmwareM: Trusted boot
HKG18-223 - Trusted FirmwareM: Trusted boot
 

Recently uploaded

GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
James Anderson
 
Knowledge engineering: from people to machines and back
Knowledge engineering: from people to machines and backKnowledge engineering: from people to machines and back
Knowledge engineering: from people to machines and back
Elena Simperl
 
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdfSmart TV Buyer Insights Survey 2024 by 91mobiles.pdf
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf
91mobiles
 
Generating a custom Ruby SDK for your web service or Rails API using Smithy
Generating a custom Ruby SDK for your web service or Rails API using SmithyGenerating a custom Ruby SDK for your web service or Rails API using Smithy
Generating a custom Ruby SDK for your web service or Rails API using Smithy
g2nightmarescribd
 
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Product School
 
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
Tobias Schneck
 
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdfFIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
FIDO Alliance
 
State of ICS and IoT Cyber Threat Landscape Report 2024 preview
State of ICS and IoT Cyber Threat Landscape Report 2024 previewState of ICS and IoT Cyber Threat Landscape Report 2024 preview
State of ICS and IoT Cyber Threat Landscape Report 2024 preview
Prayukth K V
 
De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...
De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...
De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...
Product School
 
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered QualitySoftware Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Inflectra
 
GraphRAG is All You need? LLM & Knowledge Graph
GraphRAG is All You need? LLM & Knowledge GraphGraphRAG is All You need? LLM & Knowledge Graph
GraphRAG is All You need? LLM & Knowledge Graph
Guy Korland
 
How world-class product teams are winning in the AI era by CEO and Founder, P...
How world-class product teams are winning in the AI era by CEO and Founder, P...How world-class product teams are winning in the AI era by CEO and Founder, P...
How world-class product teams are winning in the AI era by CEO and Founder, P...
Product School
 
Designing Great Products: The Power of Design and Leadership by Chief Designe...
Designing Great Products: The Power of Design and Leadership by Chief Designe...Designing Great Products: The Power of Design and Leadership by Chief Designe...
Designing Great Products: The Power of Design and Leadership by Chief Designe...
Product School
 
Slack (or Teams) Automation for Bonterra Impact Management (fka Social Soluti...
Slack (or Teams) Automation for Bonterra Impact Management (fka Social Soluti...Slack (or Teams) Automation for Bonterra Impact Management (fka Social Soluti...
Slack (or Teams) Automation for Bonterra Impact Management (fka Social Soluti...
Jeffrey Haguewood
 
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
Sri Ambati
 
The Art of the Pitch: WordPress Relationships and Sales
The Art of the Pitch: WordPress Relationships and SalesThe Art of the Pitch: WordPress Relationships and Sales
The Art of the Pitch: WordPress Relationships and Sales
Laura Byrne
 
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
Product School
 
Bits & Pixels using AI for Good.........
Bits & Pixels using AI for Good.........Bits & Pixels using AI for Good.........
Bits & Pixels using AI for Good.........
Alison B. Lowndes
 
Neuro-symbolic is not enough, we need neuro-*semantic*
Neuro-symbolic is not enough, we need neuro-*semantic*Neuro-symbolic is not enough, we need neuro-*semantic*
Neuro-symbolic is not enough, we need neuro-*semantic*
Frank van Harmelen
 
Connector Corner: Automate dynamic content and events by pushing a button
Connector Corner: Automate dynamic content and events by pushing a buttonConnector Corner: Automate dynamic content and events by pushing a button
Connector Corner: Automate dynamic content and events by pushing a button
DianaGray10
 

Recently uploaded (20)

GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
GDG Cloud Southlake #33: Boule & Rebala: Effective AppSec in SDLC using Deplo...
 
Knowledge engineering: from people to machines and back
Knowledge engineering: from people to machines and backKnowledge engineering: from people to machines and back
Knowledge engineering: from people to machines and back
 
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdfSmart TV Buyer Insights Survey 2024 by 91mobiles.pdf
Smart TV Buyer Insights Survey 2024 by 91mobiles.pdf
 
Generating a custom Ruby SDK for your web service or Rails API using Smithy
Generating a custom Ruby SDK for your web service or Rails API using SmithyGenerating a custom Ruby SDK for your web service or Rails API using Smithy
Generating a custom Ruby SDK for your web service or Rails API using Smithy
 
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
Unsubscribed: Combat Subscription Fatigue With a Membership Mentality by Head...
 
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
Kubernetes & AI - Beauty and the Beast !?! @KCD Istanbul 2024
 
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdfFIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
FIDO Alliance Osaka Seminar: The WebAuthn API and Discoverable Credentials.pdf
 
State of ICS and IoT Cyber Threat Landscape Report 2024 preview
State of ICS and IoT Cyber Threat Landscape Report 2024 previewState of ICS and IoT Cyber Threat Landscape Report 2024 preview
State of ICS and IoT Cyber Threat Landscape Report 2024 preview
 
De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...
De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...
De-mystifying Zero to One: Design Informed Techniques for Greenfield Innovati...
 
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered QualitySoftware Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
Software Delivery At the Speed of AI: Inflectra Invests In AI-Powered Quality
 
GraphRAG is All You need? LLM & Knowledge Graph
GraphRAG is All You need? LLM & Knowledge GraphGraphRAG is All You need? LLM & Knowledge Graph
GraphRAG is All You need? LLM & Knowledge Graph
 
How world-class product teams are winning in the AI era by CEO and Founder, P...
How world-class product teams are winning in the AI era by CEO and Founder, P...How world-class product teams are winning in the AI era by CEO and Founder, P...
How world-class product teams are winning in the AI era by CEO and Founder, P...
 
Designing Great Products: The Power of Design and Leadership by Chief Designe...
Designing Great Products: The Power of Design and Leadership by Chief Designe...Designing Great Products: The Power of Design and Leadership by Chief Designe...
Designing Great Products: The Power of Design and Leadership by Chief Designe...
 
Slack (or Teams) Automation for Bonterra Impact Management (fka Social Soluti...
Slack (or Teams) Automation for Bonterra Impact Management (fka Social Soluti...Slack (or Teams) Automation for Bonterra Impact Management (fka Social Soluti...
Slack (or Teams) Automation for Bonterra Impact Management (fka Social Soluti...
 
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
GenAISummit 2024 May 28 Sri Ambati Keynote: AGI Belongs to The Community in O...
 
The Art of the Pitch: WordPress Relationships and Sales
The Art of the Pitch: WordPress Relationships and SalesThe Art of the Pitch: WordPress Relationships and Sales
The Art of the Pitch: WordPress Relationships and Sales
 
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
From Daily Decisions to Bottom Line: Connecting Product Work to Revenue by VP...
 
Bits & Pixels using AI for Good.........
Bits & Pixels using AI for Good.........Bits & Pixels using AI for Good.........
Bits & Pixels using AI for Good.........
 
Neuro-symbolic is not enough, we need neuro-*semantic*
Neuro-symbolic is not enough, we need neuro-*semantic*Neuro-symbolic is not enough, we need neuro-*semantic*
Neuro-symbolic is not enough, we need neuro-*semantic*
 
Connector Corner: Automate dynamic content and events by pushing a button
Connector Corner: Automate dynamic content and events by pushing a buttonConnector Corner: Automate dynamic content and events by pushing a button
Connector Corner: Automate dynamic content and events by pushing a button
 

LCU13: GPGPU on ARM Experience Report

  • 1. GPGPU on ARM Tom Gall, Gil Pitney, 30th Oct 2013
  • 2. connect.linaro.org • This session will discuss the current state of the art of GPGPU technologies on ARM SoC systems. • What standards are there? Where are the drivers? What is the best hardware? What are the problem areas? What open source software has been accelerated? What opportunities for taking advantage of GPGPU computing on ARM exist? • There will also be some time for interactive discussion about current and potential plans. Session Description
  • 3. connect.linaro.org • At previous Connect (Dublin) discussed GPGPU, and what Linaro should do in that area. • Recognized that: GPUs are much more energy efficient for highly parallel computing problems than using just the main processor, even with NEON SIMD. Big performance gains (> 20X) using the GPU for certain open source workloads. • Decided: • to get hardware and start gaining some experience with GPGPU drivers; • it would be good to accelerate some OSS projects with GPGPU. • So, we formed a GPGPU sub-team as part of the GWG. GWG: New GPGPU Sub-group
  • 4. connect.linaro.org • Graphics pipeline operations migrate from the CPU to GPU: – ’80’s: CPU + framebuffers with blitters. – ’90’s: GPUs output one pixel per clock cycle, add pipeline operations: • texture mapping, z-buffering, rasterization; • transform and lighting (still fixed hardware) operations. • Fixed functions become programmable kernels, adding multiple pipelines: – 2000’s: GPUs become programmable, multicore: • Programmable kernels (shaders) added to the per-pixel and vertex processing stages. • More floating point precision, more GPU memory • Unified shaders (eg: GeForce 8 “Streaming Multiprocessor”) handle vertex, pixel and geometry computations. • GPU: initially a single core, fixed function hardware pipeline implementation designed solely to accelerate graphics becomes… • GPGPU: set of parallel and highly programmable cores (shaders) which can be leveraged for more general purpose computation. GPU to GPGPU: Hardware Evolution
  • 5. connect.linaro.org • Software co-evolves with GPU hardware. – ’90’s: Graphics Languages: OpenGL (led GPU hardware features), DirectX. – 2000: Shading Languages: OpenGL/GLSL, Direct3D/HLSL, Cg, others... • Shading Languages: Need to understand graphics pipeline ops. • But, shaders basically perform matrix and vector operations, ideal for many non-graphics, scientific applications. • Enter higher level parallel computing “stream processing” languages: – 2007: CUDA “Compute Unified Device Architecture” (NVIDIA) – 2008: OpenCL “Open Compute Language” (Apple/Khronos) GPU to GPGPU: Software Evolution
  • 6. connect.linaro.org • Ideal GPGPU applications have large data sets, high parallelism, and minimal dependency between data elements: – Advanced imaging: (Face recognition, cloth simulation, object detection). – Cryptography searches: (Bitcoin, distributed.net). – Video decode and post-processing: (iDCT, motion compensation, VLD, IQ, edge enhancement) – Bioinformatics: (BLAST (protein and genome sequence comparisons), Folding@Home). – Genetic Algorithms, Weather prediction, Particle Physics simulations – MapReduce algorithm. • The “13 Dwarfs of GP/GPU computing” [2] and benchmarks [1] evaluate parallel programming models and hardware architectures for algorithms sharing similar computation and inter-processor communication patterns: - Computationally limited: Dense linear algebra, N-Body methods, cryptography. - Memory bandwidth limited: Sparse linear algebra, structured grid. - Memory latency limited: Graph traversal, dynamic programming, unstructured grid, FFT. GPGPU Use Cases
  • 7. connect.linaro.org • Hardware: – HSA (Heterogeneous System Architecture) foundation creates hardware and software standards. Implementations coming…. (AMD/Kaveri, PS4: 4Q13?). Hardware must be certified HSA compliant. – HSA requires one fully coherent shared memory heap, with unified addressing, between CPUs and Compute Devices (i.e. a pointer can be passed between CPU and GPU directly). An HSA-specified MMU (hMMU) allows the compute devices to share the page table mappings with host CPU(s). • Software: – OpenCL (Open Compute Language) is an open standard by the Khronos Group. One is free to implement the standard and pay for certification. A large and growing number of problem domain libraries and apps have been built on OpenCL. – CUDA (Compute Unified Device Architecture) by NVidia is limited to NVidia hardware. There is no standardization body and no certification. Historically CUDA was not able to be licensed but on June 18th, 2013 nVidia announced it was going to start licensing its GPU IP. – Renderscript, part of Android, by Google. Renderscript is not a standard. There is no certification. There is no implementation outside of Android. – HSA tools: Compiler to HSAIL, HSAIL Finalizer to accelerator code, and HSA host runtime software. Standards / Programming Platforms
  • 8. connect.linaro.org • Many OSS libraries support CUDA and/or OpenCL. – Linear Algebra Libraries: ViennaCL, MAGMA, clMath(clBLAS), VexCL. – Vision: OpenCV (Computer Vision) some modules written to CUDA/OpenCL. – Database: sqLite on CUDA (20x-70x speedup for SELECT queries). – Finance: QuantLib (faster option pricing for HFT, Monte Carlo simulations, and company credit risk). – Scientific Domains: OpenMM (Molecular Modeling), OpenBR (Biometric Recognition). • Language bindings ease programmability, proliferate GPU usage: – WebCL: JavaScript binding to OpenCL, and Khronos standard. Demos have shown (50x-100x) performance improvements over pure javascript web apps. – Python: PyOpenCL, Java: Aparapi, Jogamp/JOCL. Open Source Software accelerated with GPGPU
  • 9. connect.linaro.org • ZiiLabs: – H/W: ZMS-40: Quad ARM Cortex A9 with 96 StemCell cores. – S/W: OpenCL 1.1 (desktop profile) drivers from Creative for custom devices. • Vivante: – H/W: GC600+ series licensed to Marvell ARMADA, Freescale i.MX6, etc. – S/W: OpenCL 1.1 EP drivers that work on FreeScale processors (SabreBoard). • Qualcomm: – H/W: Adreno 300+ series, in Snapdragon SoC’s. – S/W: OpenCL 1.2 EP Android drivers in Adreno SDK (not shipped on device). • ARM Mali: – H/W: Samsung Exynos 4/5 series, many others. (eg: Arndale, Chromebook). – S/W: Drivers for Exynos only seem to work on the Arndale-board when the LCD is also ordered; Chromebook Mali drivers just made available (Oct 2013). • Imagination Technologies: – H/W: TI OMAP, Exynos 4/5 series with PowerVR SGX GPUs – S/W: OpenCL 1.1 EP: TI requires NDA, Samsung ODROID-XU Android drivers avail (Linux drivers require NDA). » Source: http://streamcomputing.eu/blog/2013-09-05/mobile-processor-opencl-drivers-q3-2013/ OpenCL Drivers on Various Hardware
  • 10. connect.linaro.org • Challenges – GPGPU drivers are mostly closed, and only slowly becoming available. – Google invested in Renderscript vs OpenCL on Android. – HSA compliant hardware can extend the GPGPU into new problem domains by increasing performance for memory latency constrained algorithms, but new hardware designs take a long time to implement. – Programming models like Renderscript/OpenCL still not so simple. • Opportunities – Performance: Select a set of key algorithm libraries to benchmark and accelerate. – OpenCL: Invest in a CPU OSS version of OpenCL, and update to latest OpenCL spec. – Simplify Programmability: Explore directive based models: eg: OpenACC/OpenMP 4.0 ? Challenges and Opportunities
  • 11. connect.linaro.org • Address availability of OpenCL driver needs through the use of Clover (an open source CPU only driver) and advance the use of non GPU devices such as TI DSPs which will also be supported by Clover. Update Clover to OCL v1.2. • Accelerate libjpeg-turbo : utilizing an existing OpenCL port, validate it works on an ARM OpenCL implementation, and compare to a NEON optimized version. • Accelerate sqlite : Sqlite is a database found in mobile and embedded solutions for both Android and Linux. A past project accelerated it using CUDA. We propose to accelerate the database using OpenCL. • Accelerate OpenCV: OpenCV is an open source computer vision and machine learning software library built to provide a common infrastructure for computer vision applications and to accelerate the use of machine perception in the commercial products. Accelerate using OpenCL. GWG Roadmap: GPGPU Next Steps
  • 12. More about Linaro: http://www.linaro.org/about/ More about Linaro engineering: http://www.linaro.org/engineering/ How to join: http://www.linaro.org/about/how-to-join Linaro members: www.linaro.org/members connect.linaro.org