Partitioned Global Address Space Languages Kathy Yelick Lawrence Berkeley National Laboratory and UC Berkeley Joint work with The Titanium Group: S. Graham, P. Hilfinger, P. Colella, D. Bonachea, K. Datta, E. Givelberg, A. Kamil, N. Mai, A. Solar, J. Su, T. Wen The Berkeley UPC Group: C. Bell, D. Bonachea, W. Chen, J. Duell, P. Hargrove, P. Husbands, C. Iancu, R. Nishtala, M. Welcome PGAS Languages 1 Kathy Yelick
The 3 P’s of Parallel Computing • Productivity • Global address space supports construction of complex shared data structures • High level constructs (e. g. , multidimensional arrays) simplify programming • Performance • PGAS Languages are Faster than two-sided MPI • Some surprising hints on performance tuning • Portability • These languages are nearly ubiquitous Kathy Yelick, 2
Partitioned Global Address Space Global address space • Global address space: any thread/process may directly read/write data allocated by another • Partitioned: data is designated as local (near) or global (possibly far); programmer controls layout x: 1 y: x: 5 y: l: l: g: g: p 0 p 1 x: 7 y: 0 By default: • Object heaps are shared • Program stacks are private pn • 3 Current languages: UPC, CAF, and Titanium • Emphasis in this talk on UPC & Titanium (based on Java) Kathy Yelick, 3
PGAS Language Overview • Many common concepts, although specifics differ • Consistent with base language • Both private and shared data • int x[10]; and shared int y[10]; • Support for distributed data structures • Distributed arrays; local and global pointers/references • One-sided shared-memory communication • Simple assignment statements: x[i] = y[i]; or t = *p; • Bulk operations: memcpy in UPC, array ops in Titanium and CAF • Synchronization • Global barriers, locks, memory fences • Collective Communication, IO libraries, etc. Kathy Yelick, 4
Example: Titanium Arrays • Ti Arrays created using Domains; indexed using Points: double [3 d] grid. A = new double [[0, 0, 0]: [10, 10]]; • Eliminates some loop bound errors using foreach (p in grid. A. domain()) grid. A[p] = grid. A[p]*c + grid. B[p]; • Rich domain calculus allow for slicing, subarray, transpose and other operations without data copies • Array copy operations automatically work on intersection data[neighbor. Pos]. copy(mydata); intersection (copied area) “restrict”-ed (non-ghost) cells ghost cells mydata[neighor. Pos] Kathy Yelick, 5
Productivity: Line Count Comparison • Comparison of NAS Parallel Benchmarks • UPC version has modest programming effort relative to C • Titanium even more compact, especially for MG, which uses multi-d arrays • Caveat: Titanium FT has user-defined Complex type and cross-language support used to call FFTW for serial 1 D FFTs UPC results from Tarek El-Gazhawi et al; CAF from Chamberlain et al; Titanium joint with Kaushik Datta & Dan Bonachea Kathy Yelick, 6
Case Study 1: Block-Structured AMR • Adaptive Mesh Refinement (AMR) is challenging • Irregular data accesses and control from boundaries • Mixed global/local view is useful Titanium AMR benchmarks available AMR Titanium work by Tong Wen and Philip Colella Kathy Yelick, 7
AMR in Titanium C++/Fortran/MPI AMR Titanium AMR • Chombo package from LBNL • Bulk-synchronous comm: • Entirely in Titanium • Finer-grained communication • Pack boundary data between procs • No explicit pack/unpack code • Automated in runtime system Code Size in Lines C++/Fortran/MPI Titanium AMR data Structures 35000 2000 AMR operations 6500 1200 Elliptic PDE solver 4200* 1500 * Somewhat more functionality in PDE part of Chombo code 10 X reduction in lines of code! Elliptic PDE solver running time (secs) PDE Solver Time (secs) Serial Parallel (28 procs) C++/Fortran/MPI Titanium 57 53 113 Comparable running time 126 Work by Tong Wen and Philip Colella; Communication optimizations joint with Jimmy Su Kathy Yelick, 8
Immersed Boundary Simulation in Titanium • Modeling elastic structures in an incompressible fluid. • Blood flow in the heart, blood clotting, inner ear, embryo growth, and many more • Complicated parallelization • Particle/Mesh method • “Particles” connected into materials Code Size in Lines Fortran 8000 Joint work with Ed Givelberg, Armando Solar-Lezama Titanium 4000 Kathy Yelick, 9
The 3 P’s of Parallel Computing • Productivity • Global address space supports complex shared structures • High level constructs simplify programming • Performance • PGAS Languages are Faster than two-sided MPI • Better match to most HPC networks • Some surprising hints on performance tuning • Send early and often is sometimes best • Portability • These languages are nearly ubiquitous Kathy Yelick, 10
PGAS Languages: High Performance Strategy for acceptance of a new language • Make it run faster than anything else Keys to high performance • Parallelism: • Scaling the number of processors • Maximize single node performance • Generate friendly code or use tuned libraries (BLAS, FFTW, etc. ) • Avoid (unnecessary) communication cost • Latency, bandwidth, overhead • Berkeley UPC and Titanium use GASNet communication layer • Avoid unnecessary delays due to dependencies • Load balance; Pipeline algorithmic dependencies Kathy Yelick, 11
One-Sided vs Two-Sided one-sided put message address data payload two-sided message id host CPU network interface data payload memory • A one-sided put/get message can be handled directly by a network interface with RDMA support • Avoid interrupting the CPU or storing data from CPU (preposts) • A two-sided messages needs to be matched with a receive to identify memory address to put data • Offloaded to Network Interface in networks like Quadrics • Need to download match tables to interface (from host) Kathy Yelick, 12
(up is good) Performance Advantage of One-Sided Communication: GASNet vs MPI • Opteron/Infini. Band (Jacquard at NERSC): • GASNet’s vapi-conduit and OSU MPI 0. 9. 5 MVAPICH • Half power point (N ½ ) differs by one order of magnitude Joint work with Paul Hargrove and Dan Bonachea Kathy Yelick, 13
(down is good) GASNet: Portability and High-Performance GASNet better for latency across machines Joint work with UPC Group; GASNet design by Dan Bonachea Kathy Yelick, 14
(up is good) GASNet: Portability and High-Performance GASNet at least as high (comparable) for large messages Joint work with UPC Group; GASNet design by Dan Bonachea Kathy Yelick, 15
(up is good) GASNet: Portability and High-Performance GASNet excels at mid-range sizes: important for overlap Joint work with UPC Group; GASNet design by Dan Bonachea Kathy Yelick, 16
Case Study 2: NAS FT • Performance of Exchange (Alltoall) is critical • 1 D FFTs in each dimension, 3 phases • Transpose after first 2 for locality • Bisection bandwidth-limited • Problem as #procs grows • Three approaches: • Exchange: • wait for 2 nd dim FFTs to finish, send 1 message per processor pair • Slab: • wait for chunk of rows destined for 1 proc, send when ready • Pencil: • send each row as it completes Joint work with Chris Bell, Rajesh Nishtala, Dan Bonachea Kathy Yelick, 17
Overlapping Communication • Goal: make use of “all the wires all the time” • Schedule communication to avoid network backup • Trade-off: overhead vs. overlap • Exchange has fewest messages, less message overhead • Slabs and pencils have more overlap; pencils the most • Example: Class D problem on 256 Processors Exchange (all data at once) 512 Kbytes Slabs (contiguous rows that go to 1 processor) 64 Kbytes Pencils (single row) 16 Kbytes Joint work with Chris Bell, Rajesh Nishtala, Dan Bonachea Kathy Yelick, 18
NAS FT Variants Performance Summary. 5 Tflops • Slab is always best for MPI; small message cost too high • Pencil is always best for UPC; more overlap Joint work with Chris Bell, Rajesh Nishtala, Dan Bonachea Kathy Yelick, 19
Case Study 2: LU Factorization • Direct methods have complicated dependencies • Especially with pivoting (unpredictable communication) • Especially for sparse matrices (dependence graph with holes) • LU Factorization in UPC • Use overlap ideas and multithreading to mask latency • Multithreaded: UPC threads + user threads + threaded BLAS • Panel factorization: Including pivoting • Update to a block of U • Trailing submatrix updates • Status: • Dense LU done: HPL-compliant • Sparse version underway Joint work with Parry Husbands Kathy Yelick, 20
UPC HPL Performance • MPI HPL numbers from HPCC database • Large scaling: • 2. 2 TFlops on 512 p, • 4. 4 TFlops on 1024 p (Thunder) • Comparison to Sca. LAPACK on an Altix, a 2 x 4 process grid • Sca. LAPACK (block size 64) 25. 25 GFlop/s (tried several block sizes) • UPC LU (block size 256) - 33. 60 GFlop/s, (block size 64) - 26. 47 GFlop/s • n = 32000 on a 4 x 4 process grid • Sca. LAPACK - 43. 34 GFlop/s (block size = 64) • UPC - 70. 26 Gflop/s (block size = 200) Kathy Yelick, 21 Joint work with Parry Husbands
The 3 P’s of Parallel Computing • Productivity • Global address space supports complex shared structures • High level constructs simplify programming • Performance • PGAS Languages are Faster than two-sided MPI • Some surprising hints on performance tuning • Portability • These languages are nearly ubiquitous • Source-to-source translators are key • Combined with portable communication layer • Specialized compilers are useful in some cases Kathy Yelick, 22
Portability of Titanium and UPC • Titanium and the Berkeley UPC translator use a similar model • Source-to-source translator (generate ISO C) • Runtime layer implements global pointers, etc • Common communication layer (GASNet) Also used by gcc/upc • Both run on most PCs, SMPs, clusters & supercomputers • Support Operating Systems: • Linux, Free. BSD, Tru 64, AIX, IRIX, HPUX, Solaris, Cygwin, Mac. OSX, Unicos, Super. UX • UPC translator somewhat less portable: we provide a http-based compile server • Supported CPUs: • x 86, Itanium, Alpha, Sparc, Power. PC, PA-RISC, Opteron • GASNet communication: • Myrinet GM, Quadrics Elan, Mellanox Infiniband VAPI, IBM LAPI, Cray X 1, SGI Altix, Cray/SGI SHMEM, and (for portability) MPI and UDP • Specific supercomputer platforms: • HP Alpha. Server, Cray X 1, IBM SP, NEC SX-6, Cluster X (Big Mac), SGI Altix 3000 • Underway: Cray XT 3, BG/L (both run over MPI) • Can be mixed with MPI, C/C++, Fortran Joint work with Titanium and UPC groups Kathy Yelick, 23
Portability of PGAS Languages Other compilers also exist for PGAS Languages • UPC • • Gcc/UPC by Intrepid: runs on GASNet HP UPC for Alpha. Servers, clusters, … MTU UPC uses HP compiler on MPI (source to source) Cray UPC • Co-Array Fortran: • Cray CAF Compiler: X 1, X 1 E • Rice CAF Compiler (on ARMCI or GASNet), John Mellor-Crummey • • Source to source Processors: Pentium, Itanium 2, Alpha, MIPS Networks: Myrinet, Quadrics, Altix, Origin, Ethernet OS: Linux 32 Red. Hat, IRIS, Tru 64 NB: source-to-source requires cooperation by backend compilers Kathy Yelick, 24
Summary • PGAS languages offer performance advantages • Good match to RDMA support in networks • Smaller messages may be faster: • make better use of network: postpone bisection bandwidth pain • can also prevent cache thrashing for packing • PGAS languages offer productivity advantage • Order of magnitude in line counts for grid-based code in Titanium • Push decisions about packing/not into runtime for portability (advantage of language with translator vs. library approach) • Source-to-source translation • The way to ubiquity • Complement highly tuned machine-specific compilers Kathy Yelick, 25
End of Slides PGAS Languages 26 Kathy Yelick
Productizing BUPC • Recent Berkeley UPC release • Support full 1. 2 language spec • Supports collectives (tuning ongoing); memory model compliance • Supports UPC I/O (naïve reference implementation) • Large effort in quality assurance and robustness • Test suite: 600+ tests run nightly on 20+ platform configs • Tests correct compilation & execution of UPC and GASNet • >30, 000 UPC compilations and >20, 000 UPC test runs per night • Online reporting of results & hookup with bug database • Test suite infrastructure extended to support any UPC compiler • now running nightly with GCC/UPC + UPCR • also support HP-UPC, Cray UPC, … • Online bug reporting database • Over >1100 reports since Jan 03 • > 90% fixed (excl. enhancement requests) Kathy Yelick, 27
Benchmarking • Next few UPC and MPI application benchmarks use the following systems • • Myrinet: Myrinet 2000 PCI 64 B, P 4 -Xeon 2. 2 GHz Infini. Band: IB Mellanox Cougar 4 X HCA, Opteron 2. 2 GHz Elan 3: Quadrics Qs. Net 1, Alpha 1 GHz Elan 4: Quadrics Qs. Net 2, Itanium 2 1. 4 GHz Kathy Yelick, 30
PGAS Languages: Key to High Performance One way to gain acceptance of a new language • Make it run faster than anything else Keys to high performance • Parallelism: • Scaling the number of processors • Maximize single node performance • Generate friendly code or use tuned libraries (BLAS, FFTW, etc. ) • Avoid (unnecessary) communication cost • Latency, bandwidth, overhead • Avoid unnecessary delays due to dependencies • Load balance • Pipeline algorithmic dependencies Kathy Yelick, 31
Hardware Latency • Network latency is not expected to improve significantly • Overlapping communication automatically (Chen) • Overlapping manually in the UPC applications (Husbands, Welcome, Bell, Nishtala) • Language support for overlap (Bonachea) Kathy Yelick, 32
Effective Latency Communication wait time from other factors • Algorithmic dependencies • Use finer-grained parallelism, pipeline tasks (Husbands) • Communication bandwidth bottleneck • Message time is: Latency + 1/Bandwidth * Size • Too much aggregation hurts: wait for bandwidth term • De-aggregation optimization: automatic (Iancu); • Bisection bandwidth bottlenecks • Spread communication throughout the computation (Bell) Kathy Yelick, 33
Fine-grained UPC vs. Bulk-Synch MPI • How to waste money on supercomputers • Pack all communication into single message (spend memory bandwidth) • Save all communication until the last one is ready (add effective latency) • Send all at once (spend bisection bandwidth) • Or, to use what you have efficiently: • Avoid long wait times: send early and often • Use “all the wires, all the time” • This requires having low overhead! Kathy Yelick, 34
What You Won’t Hear Much About • Compiler/runtime/gasnet bug fixes, performance tuning, testing, … • >13, 000 e-mail messages regarding cvs checkins • Nightly regression testing • 25 platforms, 3 compilers (head, opt-branch, gcc-upc), • Bug reporting • 1177 bug reports, 1027 fixed • Release scheduled for later this summer • Beta is available • Process significantly streamlined Kathy Yelick, 35
Take-Home Messages • Titanium offers tremendous gains in productivity • High level domain-specific array abstractions • Titanium is being used for real applications • Not just toy problems • Titanium and UPC are both highly portable • Run on essentially any machine • Rigorously tested and supported • PGAS Languages are Faster than two-sided MPI • Better match to most HPC networks • Berkeley UPC and Titanium benchmarks • Designed from scratch with one-side PGAS model • Focus on 2 scalability challenges: AMR and Sparse LU Kathy Yelick, 36
Titanium Background • Based on Java, a cleaner C++ • Classes, automatic memory management, etc. • Compiled to C and then machine code, no JVM • Same parallelism model at UPC and CAF • SPMD parallelism • Dynamic Java threads are not supported • Optimizing compiler • Analyzes global synchronization • Optimizes pointers, communication, memory Kathy Yelick, 37
Do these Features Yield Productivity? Joint work with Kaushik Datta, Dan Bonachea Kathy Yelick, 38
GASNet/X 1 Performance single word put single word get • GASNet/X 1 improves small message performance over shmem and MPI • Leverages global pointers on X 1 • Highlights advantage of languages vs. library approach Joint work with Christian Bell, Wei Chen and Dan Bonachea Kathy Yelick, 39
High Level Optimizations in Titanium • Irregular communication can be expensive • “Best” strategy differs by data size/distribution and machine parameters • E. g. , packing, sending bounding boxes, fine-grained are • Use of runtime optimizations • Inspector-executor Speedup relative to MPI code (Aztec library) • Performance on Sparse Mat. Vec Mult • Results: best strategy differs within the machine on a single matrix (~ 50% better) Average and maximum speedup of the Titanium version relative to the Aztec version on 1 to 16 processors Joint work with Jimmy Su Kathy Yelick, 40
Source to Source Strategy • Source-to-source translation strategy • Tremendous portability advantage • Still can perform significant optimizations • Relies on high quality back-end compilers and some coaxing in code generation • Use of “restrict” pointers in C • Understand Multi-D array indexing (Intel/Itanium issue) • Support for pragmas like IVDEP • Robust vectorizators (X 1, SSE, NEC, …) 48 x • On machines with integrated shared memory hardware need access to shared memory operations Joint work with Jimmy Su Kathy Yelick, 41
Скачать презентацию Partitioned Global Address Space Languages Kathy Yelick Lawrence
d20a2a4e2f7e792e483be9961ab15fb1.ppt