x Welcome to the jungle. The free lunch is so over

Transcription

1 Herb Sutter

2 Put a computer on every desk, in every home, in every pocket. The free lunch is so over Put a parallel supercomputer on every desk, in every home, in every pocket. Welcome to the jungle x Put a heterogeneous supercomputer on every desk, in every home, in every pocket.

3

4

5 Processors Memory

6 Processors Xbox 360 & mainstream computer AMD GPU AMD 80x86 Fusion APU Other GPU Athlon Phenom II Memory

7 Processors Xbox 360 & mainstream computer AMD GPU AMD 80x86 Athlon Fusion APU Phenom II Memory Other GPU Microsoft Azure Cloud Computing Cloud + GPU

8 Processors Memory

9 Processors (GP)GPU Cloud IaaS/HaaS ISO ISO C++0x Multicore CPU Memory

10

11 Processors (GP)GPU Cloud IaaS/HaaS ISO C++0x C++ PPL Multicore CPU Memory

12 Processors? DirectCompute (GP)GPU Cloud IaaS/HaaS ISO C++0x C++ PPL Multicore CPU Memory

13 Processors C++ AMP DirectCompute (GP)GPU Accelerated Massive Parallelism Cloud IaaS/HaaS ISO C++0x C++ PPL Multicore CPU Memory

14 Convert this (serial loop nest) void MatrixMult( float* C, const vector<float>& A, const vector<float>& B, int M, int N, int W ) { for (int y = 0; y < M; y++) for (int x = 0; x < N; x++) { float sum = 0; for(int i = 0; i < W; i++) sum += A[y*W + i] * B[i*N + x]; C[y*N + x] = sum; } }

15 Convert this (serial loop nest) void MatrixMult( to this float* (parallel C, const loop, vector<float>& CPU GPU) A, const vector<float>& B, int M, int N, int W ) { void MatrixMult( float* C, const vector<float>& A, const vector<float>& B, for (int y = 0; y < M; y++) int M, int N, int W ) for (int { x = 0; x < N; x++) { float sum array_view<const = 0; float,2> a(m,w,a), b(w,n,b); for(int i array_view<writeonly<float>,2> = 0; i < W; i++) c(m,n,c); sum += parallel_for_each( A[y*W + i] * B[i*N c.grid, + x]; [=](index<2> idx) restrict(direct3d) { C[y*N + x] float = sum; = 0; } for(int i = 0; i < a.x; i++) } sum += a(idx.y, i) * b(i, idx.x); c[idx] = sum; } ); }

16

17 Architecture Maturity & Programmer Accessibility Poor Excellent EVOLUTION OF HETEROGENEOUS COMPUTING Standards Drivers Era Architected Era Fusion System Architecture GPU Peer Processor Proprietary Drivers Era Graphics & Proprietary Driver-based APIs Adventurous programmers Exploit early programmable shader cores in the GPU Make your program look like graphics to the GPU CUDA, Brook+, etc OpenCL, DirectCompute Driver-based APIs Expert programmers C and C++ subsets Compute centric APIs, data types Multiple address spaces with explicit data movement Specialized work queue based structures Kernel mode dispatch Mainstream programmers Full C++ GPU as a co-processor Unified coherent address space Task parallel runtimes Nested Data Parallel programs User mode dispatch Pre-emption and context switching The Programmer s Guide to the APU Galaxy June 2011

18 Processors Memory

19 Single-core to multi-core ISO C++0x? PPL Parallel Patterns Library (VS2010)

20 ISO C++0x forall( x, y ) forall( z; w; v ) forall( k, l, m, n )...? Single-core to multi-core PPL Parallel Patterns Library (VS2010)

21 ISO C++0x λ parallel_for_each( items.begin(), items.end(), [=]( Item e ) { your code here } ); Single-core to multi-core PPL Parallel Patterns Library (VS2010)

22 1 language feature for multicore and STL, functors, callbacks, events,...

23 ? ISO C++0x Multi-core to hetero-core C++ AMP Accelerated Massive Parallelism

24 Multi-core to hetero-core ISO C++0x restrict parallel_for_each( items.grid, [=](index<2> i) restrict(direct3d) { your code here } ); C++ AMP Accelerated Massive Parallelism

25 1 language feature for heterogeneous cores

26 Processors Memory

27 Problem: Some cores don t support the entire C++ language. Solution: General restriction qualifiers enable expressing language subsets within the language. Direct3d math functions in the box. Example double sin( double ); // 1a: general code double sin( double ) restrict(direct3d); // 1b: specific code double cos( double ) restrict(direct3d); // 2: same code for either parallel_for_each( c.grid, [=](index<2> idx) restrict(direct3d) { sin( data.angle ); // ok, chooses overload based on context cos( data.angle ); // ok });

28 Initially supported restriction qualifiers: restrict(cpu): The implicit default. restrict(direct3d): Can execute on any DX11 device via DirectCompute. Restrictions follow limitations of DX11 device model (e.g., no function pointers, virtual calls, goto). Potential future directions: restrict(pure): Declare and enforce a function has no side effects. Great to be able to state declaratively for parallelism. General facility for language subsets, not just about compute targets.

29 Problem: Memory may be flat, nonuniform, incoherent, and/or disjoint. Solution: Portable view that works like an N-dimensional iterator range. Future-proof: No explicit.copy()/.sync(). As needed by each actual device. Example void MatrixMult( float* C, const vector<float>& A, const vector<float>& B, int M, int N, int W ) { array_view<const float,2> a(m,w,a), b(w,n,b); // 2D view over C array array_view<writeonly<float>,2> c(m,n,c); // 2D view over C++ std::vector } parallel_for_each( c.grid, [=](index<2> idx) restrict(direct3d) { } );

30 TM

31 Bring CPU debugging experience to the GPU

32 Bring CPU debugging experience to the GPU

33

34

35 TM

36 Cloud GPU # cores, not counting SIMD Cloud OoO GPU InO CPU OoO CPU

37 Cloud GPU # cores, not counting SIMD Cloud OoO Welcome to the jungle GPU The free lunch is so over InO CPU OoO CPU

38 Processors Memory

39 C++ PPL: 9:45am C++ AMP: 2:00pm, Room 406 Herb Sutter