Implementation of Cumo, a CUDA-aware version of Ruby/Numo

A report about Cumo project at Ruby Association Grant 2017.

Transcript

1. Implementation of Ruby/Cumo, a CUDA-aware version of Ruby/Numo Naotoshi Seo July 07, 2018 Grant 2017 Report https://github.com/sonots/cumo
2. Self Introduction • Naotoshi Seo @sonots • DeNA Co., Ltd. • CRuby committer • Recently working on development of DNN framework at Preferred Networks, Inc (出向) 2
3. Outline 3 • Project Introduction • Cumo Features • Notices (or Difficulties) I met • Feature Proposals to Ruby • Project Achievement
4. Project Introduction 4
5. 5 1SPKFDU
   *OUSPEVDUJPO Project Proposal
6. 6 https://ruby-numo.github.io/narray/ 1SPKFDU
   *OUSPEVDUJPO
7. Why GPU? • GPU is bad at branching • GPU simplifies branch prediction and out-of-order mechanism instead. • GPU is suitable for matrix computation 7 • GPU is fast, and recently essential for Deep Learning • GPU is good at parallel computation • Order of magnitude is like 24 cores with CPU • 3,000 ~ 4,000 cores with GPU 1SPKFDU
   *OUSPEVDUJPO
8. Position of Cumo in Scientific Computing in Ruby 8
9. 4DJFOUJpD
   $PNQVUJOH
   JO
   1ZUIPO 9 NumPy CuPy PyCUDA Chainer TensorFlow MXNet Cython pybind11 DNN Tensor CUDA binding Useful tools for writing bindings C++ C++ Python 1PTJUJPO
   PG
   $VNP
10. 4DJFOUJpD
    $PNQVUJOH
    JO
    3VCZ 10 Numo/NArray Cumo RbCUDA Red-chainer TensorFlow.rb MXNet.rb Rubex /" DNN Tensor CUDA binding Useful tools for writing bindings (or NMatrix) PyCall Binding to Python C++ C++ Ruby 1PTJUJPO
    PG
    $VNP
11. Outline 11 • Project Proposal • Cumo Features • Notices (or Difficulties) I met • Feature Proposals to Ruby • Project Achievement
14. Cumo Features • Highly compatible with Ruby/Numo • Element-wise operations • Reduction operations • Dot operation using cuBLAS • CUDA memory pool • JIT compilation of user-defined functions 14
15. Highly Compatible with Numo • Ruby/Numo users can easily switch into Cumo to leverage power of GPU 15
    pOE
    
    OBNF
     SC
    c
    a
    
    
    YBSHT
    TFE
    J
    a
    
    
    F
    T/VNP$VNPH
    a
    
    
    F
    TOVNPDVNPH JG
    HQV

    
    SFRVJSF
    DVNPOBSSBZ

    
    9VNP
    
    $VNP
 FMTF

    
    SFRVJSF
    OVNPOBSSBZ

    
    9VNP
    
    /VNP
 FOE
 B
    
    9VNP%'MPBU[FSPT 
     
 C
    
    9VNP%'MPBUPOFT 
     
 D
    
    B
    
    C $VNP
    'FBUVSFT
16. Element-wise opeations • Element-wise is like matrix additions • All elements are independent • Easy to perform in parallel 16 1 2 3 4 5 6 2 3 4 5 6 7 + A B 3 5 7 9 11 13 = C       Thread
 IDs $VNP
    'FBUVSFT
17. Reduction opeations • Like sum • sum([1,2,3,4]) #=> 10 • Elements are not independent • Not so easy to perform in parallel 17 $VNP
    'FBUVSFT
18. 18 http://developer.download.nvidia.com/assets/cuda/files/reduction.pdf 3FEVDUJPO
    PQFSBUJPOT
19. 19 3FEVDUJPO
    PQFSBUJPOT
20. Dot product (GEMM) using cuBLAS • Dot is more complicated than reduction • NVIDIA's cuBLAS library supports it as GEMM (GEneral matrix-matrix mulitplication) and fast • However, cuBLAS supports only f-contiguous (column major) although we write CRuby extensions in C (c-contiguous, raw-major) 20 1 2 3 4 5 6 7 8 9 1 4 7 2 5 8 3 6 9 C-contiguous F (Fortran) $VNP
    'FBUVSFT
21. Using cuBLAS with c- contiguous data 21 A = [1, 2, 3, 4, 5, 6] B = [1, 2, 3, 4, 5, 6] C = [9, 12, 15, 19, 26, 33, 29, 40, 51] A = [1, 2, 3, 4, 5, 6] B = [1, 2, 3, 4, 5, 6] C = [9, 12, 15, 19, 26, 33, 29, 40, 51]
    $DPOUJVHPVT
    3PX
    NBKPS
    'DPOUJVHPVT
    $PMVNO
    NBKPS No data copy, changing only attributes (shape) https://www.christophlassner.de/using-blas-from-c-with-row-major-data.html %PU
    QSPEVDU
22. CUDA Memory Pool 22
23. Why We Need Memory Pool 23 $6%"
    .FNPSZ
    1PPM • cudaMalloc / cudaFree makes slow • memory allocation / free themselves are slow • cudaMalloc synchronizes CPU and GPU CPU GPU Free Kernel1 cudaFree synchronize Idle Kernel2 something Idle cudaMalloc synchronize Malloc Launch Launch
24. Memory Pool 24 $6%"
    .FNPSZ
    1PPM • Cache to memory pool • Avoid cudaMalloc / Free as much as possible
25. CUDA Memory Pool 25 512 1024 1536 2048 2560 …. Pop 512 2048 512 1024 1536 2048 2560 …. Push (1) (2) Split next prev use 1. Round up memory size by 512 2. cudaMalloc if no block is available 3. Push to arena intead of cudaFree 4. Pop from arena if a free block is available instead of cudaMalloc Implemented Best-fit with Coalescing (BFC), which is the one used in malloc(3) $6%"
    .FNPSZ
    1PPM
26. 26 • Cumo supports users to write their own CUDA kernel on Ruby • JIT compile using NVRTC (NVIDIA Runtime Compilation), and caches it on file system. JIT compiling user-defined functions (planned) kernel = Cumo::ElementwiseKernel.new(
 'float32 x, float32 y, float32 z',
 'float32 w', # output type
 'w = (x * y) + z;', # CUDA code
 'my_kernel')
 w = kernel.call(x, y, z) $VNP
    'FBUVSFT
27. Performance Comparison with Numo 27
28. Element-wise kernel 4J[F /VNP
    $VNP
    ?   ?  
    ?   ?   ?   a = Xumo::Float32.ones(size) b = Xumo::Float32.ones(size) a + b 40 times faster for size of 10^8 28 Smaller is better UIJT Intel(R) Xeon(R) CPU E5-2686 v4 @ 2.30GHz NVIDIA Volta v100 (AWS p3 xlarge) 1FSGPSNBODF
    $PNQBSJTPO
    XJUI
    /VNP
29. Dot product 29 4J[F /VNP
    $VNP
    ?   ?   ?   ?   ?  
    a = Xumo::Float32.ones(100, size/100) b = Xumo::Float32.ones(size/100, 100) a.dot(b) 2800 times faster for size of 10^8 UIJT ※ Numo without Numo/Linalg Intel(R) Xeon(R) CPU E5-2686 v4 @ 2.30GHz NVIDIA Volta v100 (AWS p3 xlarge) Smaller is better 1FSGPSNBODF
    $PNQBSJTPO
    XJUI
    /VNP
30. red-chainer mnist example 30 • 20 times faster w/o memory pool • 75 times faster w/ memory pool Intel(R) Xeon(R) CPU E5-2686 v4 @ 2.30GHz NVIDIA Volta v100 (AWS p3 xlarge) 1FSGPSNBODF
    $PNQBSJTPO
    XJUI
    /VNP
31. Outline 31 • Project Proposal • Cumo Features • Notices (or Difficulties) I met • Feature Proposals to Ruby • Project Achievement
32. Notices (or Difficulties) I met • GPU unfriendness with GC • Difficulties compiling CUDA kernels • Lack of mkmf features • Incompatibility with Numo is required in reduction kernels for performance • Broadcast operations were slow 32
33. GPU unfriendness with GC • One criteria to perform GC in Ruby is main memory usage (malloc_limit) • GPU memory usage is not taken into account • In the case of CuPy, because Python uses reference counting, we could release GPU memory immediately after the array object is not referenced anymore. 33 def add
 a = Cumo::DFloat.ones(3, 5)
 b = Cumo::DFloat.ones(3, 5)
 a + b end c = add a and b are not immediately freed (16
    VOGSJFOEOFTT
    XJUI
    ($
34. GPU unfriendness with GC (2) • (Partial) Solution • Added NArray#free to release memory to GPU on user-desired timing • Future work? • Something like NSAutoreleasePool to release all (or restricted) objects created inside a scope. 34 def add
 a = Cumo::DFloat.ones(3, 5)
 b = Cumo::DFloat.ones(3, 5)
 c = a + b a.free; b.free c end c = add a and b are immediately freed NSAutoreleasePool *pool = \ [[NSAutoreleasePool alloc] init]; NSObject *obj = \ [[[NSObject alloc] init] autorelease]; .... [pool release]; (16
    VOGSJFOEOFTT
    XJUI
    ($
35. 35 • Need to use nvcc (NVIDIA CUDA Compiler) instead of gcc to compile CUDA kernels. • However, mkmf supports to specify only CC and CXX compilers (no .cu file) • Solution: Made a wrapper ruby script • For files with .cu extensions, use nvcc • For files with .c extensions, use gcc Lack of mkmf features %J⒏DVMUJFT
    DPNQJMJOH
    $6%"
    LFSOFMT
36. 36 • Numo returns a Ruby numeric object for reduction kernels (for cases of 0-dimensional NArray). • In Cumo, needs to copy GPU memory to host memory to create a Ruby nemeric object. • It results in synchronization with CPU. • Solution: Introduced partial incompatibility with Numo to return 0-dimensional NArray. Incompatibility with Numo is required in reduction for performance Numo::Int64.ones(2, 3).sum #=> 6
 Cumo::Int64.ones(2, 3).sum #=> Cumo::Int64#shape=[] 6 Returns a 0-dimensional NArray instead of a Ruby numeric object to avoid CPU and GPU synchronization. *ODPNQBUJCJMJUZ
    XUJI
    /VNP
    JT
    SFRVJSFE
    JO
    3FEVDUJPO
37. 1 2 3 37 • Broadcast Broadcast operations were slow #SPBEDBTU
    PQFSBUJPOT
    XFSF
    TMPX 1 2 3 4 5 6 7 8 9 10 11 12 + 1 2 3 = 2 4 6 5 7 9 8 10 12 11 13 15 1 2 3 1 2 3 4 x 3 1 x 3 4 x 3
38. 38 How Numo Treats Broadcast Example) 1000 x 3 array + 1 x 3 array user loop: loop for 3 narray loop: loop for 1000 int nd = 1; int shape[] = {1000}; for (int i=0; i<nd;++i) { for (int j=0; j<shape[i]; ++j) { (*(nf->func))(&(lp->user)); } } int size = 3; for (int i=0; i<size;++i) { p3[i] = p1[i] + p[2]; } #SPBEDBTU
    PQFSBUJPOT
    XFSF
    TMPX
39. 39 Launches Many CUDA Kernels user loop: loop for 3 narray loop: loop for 1000 int nd = 1; int shape[] = {1000}; for (int i=0; i<nd;++i) { for (int j=0; j<shape[i]; ++j) { (*(nf->func))(&(lp->user)); } } __global__ void my_kernel( int* p3, int* p2, int* p1) { int i = blockIdx.x * blockDim.x + threadIdx.x; p3[i] = p1[i] + p2[i]; } Launches CUDA Kernels 1000 times. • In first implementation of Cumo, modified user loop implementation to CUDA kernels #SPBEDBTU
    PQFSBUJPOT
    XFSF
    TMPX
40. 40 How Slow launching CUDA kernels Type Time(%) Time Calls Avg Min Max Name GPU activities: 99.89% 19.439ms 1000 19.439us 18.880us 21.312us cumo_sfloat_add API calls: 27.23% 330.78ms 13 25.445ms 35.083us 68.418ms cudaDeviceSynchronize 26.34% 319.98ms 1 319.98ms 319.98ms 319.98ms cuCtxCreate 25.32% 307.66ms 1477 208.30us 13.408us 275.62ms cudaMallocManaged 2.58% 18.703ms 1002 18.665us 16.184us 216.70us cudaLaunch nvprof • 18 micro second • Time to take cudaLaunch is almost equivalent with adding two arrays of 500,000 elements. • Also, there is a limit of CUDA queue size, e.g., 1,024. #SPBEDBTU
    PQFSBUJPOT
    XFSF
    TMPX
41. 41 • Finally, stopped using ndloop Solution: Stop using ndloop /EMPPQ
    JT
    TMPX
    GPS
    $VNP #SPBEDBTU
    PQFSBUJPOT
    XFSF
    TMPX a = Cumo::DFloat.ones(1000, 768)
 b = Cumo::DFloat.ones(1000, 1)
 a + b 56 times faster
42. Outline 42 • Project Proposal • Cumo Features • Notices (or Difficulties) I met • Feature Proposals to Ruby • Project Achievement
43. Feature Proposals to Ruby • Inplace math operations • Temporary variable 43
44. Inplace Math Operations • a += b is an abridged notation of a = a + b • Imagine a is a large matrix requiring 1GB. • a += b needs to allocate a new 1GB matrix. • Want to redefine for Cumo::NArray objects. • Current compromise: • Python allows to redefine +=. 44 https://bugs.ruby-lang.org/issues/14701 a.inplace + b 'FBUVSF
    1SPQPTBMT
    UP
    3VCZ
45. Temporary Variable • In python, we can find a variable is a temporary or not by seeing reference counts • In NumPy, • is faster than • because (x + 1) is a temporary variable and new memory is not required to compute (x + 1) + 1 45 https://bugs.ruby-lang.org/issues/14710 y = x + 1 + 1 y = x + 1 y + 1 'FBUVSF
    1SPQPTBMT
    UP
    3VCZ
46. Outline 46 • Project Proposal • Cumo Features • Notices (or Difficulties) I met • Feature Proposals to Ruby • Project Achievement
47. 47 All Achieved
48. Future Works • Support cuDNN for high performance convolutional networks • Support Float16 • Conversion between Numo::NArray and Cumo::NArray • CI server ... 48
49. Supported Functions List 49 4VQQPSUFE
    'VODUJPOT
    -JTU - << atan2 eq floor log10 min_index rms stddev -@ >> atanh erf ge (>=) log1p minimum round store [] | cbrt erfc gemm log2 mulsum seq sum []= ~ ceil exp gt (>) logseq ne sign tan * acos coerce_cast exp10 hypot lt (<) nearly_eq signbit tanh ** acosh conj exp2 im max poly sin trunc / allocate copysign expm1 inspect max_index prod sinc var & asin cos extract ldexp maximum ptp sinh % asinh cosh eye le (<=) mean reciprocal sqrt ^ atan divmod fill log min rint square * 88 methods Int8, Int16, Int32, Int64, Uint8, Uint16, Uint32, Uint64,
 SFloat (float), DFloat (double), SComplex, DComplex mixed
50. Not Yet 50 4VQQPSUFE
    'VODUJPOT
    -JTU abs isnan set_real arg isneginf sort bincount isposinf sort_index clip median cumprod minmax cumsum modf frexp rand imag rand_norm isfinite real isinf set_imag [] count_false []= count_true & eq ^ extract | fill ~ mask all? none? any? store coerce_cast where copy where2 * 20 methods (most of all) IntXX, FloatXX, ComplexXX mixed Bit * 23 methods
51. Acknowledgements • Ruby Association for Grant • Money - GPU machines cost much • Time keeper • Motivation • @mrkn for his mentoring on the grant • @masa16 for answering my questions about Numo • @hatappi and @naitoh for their work of red-chainer • red-data-tools org and Speee, Inc for hosting meetup. • Preferred Networks, Inc and developers (including me) of Chainer/CuPy for reference implementation • And, my wife for giving time to develop 51