Lecture2 cuda spring 2010

1
ECE 498AL
Lecture 2:
The CUDA Programming Model
© David Kirk/NVIDIA and Wen-mei W. Hwu, 2007-2009
ECE 498AL Spring 2010, University of Illinois, Urbana-Champaign

Parallel Programming Basics
• Things we need to consider:
– Control
– Synchronization
– Communication
• Parallel programming languages offer different ways
of dealing with above
2© David Kirk/NVIDIA and Wen-mei W. Hwu, 2007-2009

3
What is (Historical) GPGPU ?
• General Purpose computation using GPU and graphics API in
applications other than 3D graphics
– GPU accelerates critical path of application
• Data parallel algorithms leverage GPU attributes
– Large data arrays, streaming throughput
– Fine-grain SIMD parallelism
– Low-latency floating point (FP) computation
• Applications – see //GPGPU.org
– Game effects (FX) physics, image processing
– Physical modeling, computational engineering, matrix algebra,
convolution, correlation, sorting

4
Previous GPGPU Constraints
• Dealing with graphics API
– Working with the corner cases of the
graphics API
• Addressing modes
– Limited texture size/dimension
• Shader capabilities
– Limited outputs
• Instruction sets
– Lack of Integer & bit ops
• Communication limited
– Between pixels
– Scatter a[i] = p
Input Registers
Fragment Program
Output Registers
Constants
Texture
Temp Registers
per thread
per Shader
per Context
FB Memory

5
CUDA
• “Compute Unified Device Architecture”
• General purpose programming model
– User kicks off batches of threads on the GPU
– GPU = dedicated super-threaded, massively data parallel co-processor
• Targeted software stack
– Compute oriented drivers, language, and tools
• Driver for loading computation programs into GPU
– Standalone Driver - Optimized for computation
– Interface designed for compute – graphics-free API
– Data sharing with OpenGL buffer objects
– Guaranteed maximum download & readback speeds
– Explicit GPU memory management

6
An Example of Physical Reality
Behind CUDA CPU
(host)
GPU w/
local DRAM
(device)

7
Parallel Computing on a GPU
• 8-series GPUs deliver 25 to 200+ GFLOPS
on compiled parallel C applications
– Available in laptops, desktops, and clusters
• GPU parallelism is doubling every year
• Programming model scales transparently
• Programmable in C with CUDA tools
• Multithreaded SPMD model uses application
data parallelism and thread parallelism
GeForce 8800
Tesla S870
Tesla D870

8
Overview
• CUDA programming model – basic concepts and
data types
• CUDA application programming interface - basic
• Simple examples to illustrate basic concepts and
functionalities
• Performance features will be covered later

9
CUDA – C with no shader limitations!
• Integrated host+device app C program
– Serial or modestly parallel parts in host C code
– Highly parallel parts in device SPMD kernel C code
Serial Code (host)
. . .
. . .
Parallel Kernel (device)
KernelA<<< nBlk, nTid >>>(args);
Serial Code (host)
Parallel Kernel (device)
KernelB<<< nBlk, nTid >>>(args);

10
CUDA Devices and Threads
• A compute device
– Is a coprocessor to the CPU or host
– Has its own DRAM (device memory)
– Runs many threads in parallel
– Is typically a GPU but can also be another type of parallel processing
device
• Data-parallel portions of an application are expressed as device
kernels which run on many threads
• Differences between GPU and CPU threads
– GPU threads are extremely lightweight
• Very little creation overhead
– GPU needs 1000s of threads for full efficiency
• Multi-core CPU needs only a few

11
L2
FB
SP SP
L1
TF
ThreadProcessor
Vtx Thread Issue
Setup / Rstr / ZCull
Geom Thread Issue Pixel Thread Issue
Input Assembler
Host
SP SP
L1
TF
SP SP
L1
TF
SP SP
L1
TF
SP SP
L1
TF
SP SP
L1
TF
SP SP
L1
TF
SP SP
L1
TF
L2
FB
L2
FB
L2
FB
L2
FB
L2
FB
• The future of GPUs is programmable processing
• So – build the architecture around the processor
G80 – Graphics Mode

12
G80 CUDA mode – A Device Example
• Processors execute computing threads
• New operating mode/HW interface for computing
Load/store
Global Memory
Thread Execution Manager
Input Assembler
Host
Texture Texture Texture Texture Texture Texture Texture TextureTexture
Parallel Data
Cache
Parallel Data
Cache
Parallel Data
Cache
Parallel Data
Cache
Parallel Data
Cache
Parallel Data
Cache
Parallel Data
Cache
Parallel Data
Cache
Load/store Load/store Load/store Load/store Load/store

13
Extended C
• Type Qualifiers
– global, device, shared,
local, constant
• Keywords
– threadIdx, blockIdx
• Intrinsics
– __syncthreads
• Runtime API
– Memory, symbol,
execution management
• Function launch
__device__ float filter[N];
__global__ void convolve (float *image) {
__shared__ float region[M];
...
region[threadIdx] = image[i];
__syncthreads()
...
image[j] = result;
}
// Allocate GPU memory
void *myimage = cudaMalloc(bytes)
// 100 blocks, 10 threads per block
convolve<<<100, 10>>> (myimage);

14
gcc / cl
G80 SASS
foo.sass
OCG
Extended C
cudacc
EDG C/C++ frontend
Open64 Global Optimizer
GPU Assembly
foo.s
CPU Host Code
foo.cpp
Integrated source
(foo.cu)
Mark Murphy, “
NVIDIA’s Experience with Open64,”
www.capsl.udel.edu/conferences/open64/2
008/Papers/101.doc© David Kirk/NVIDIA and Wen-mei W. Hwu, 2007-2009

15
Arrays of Parallel Threads
• A CUDA kernel is executed by an array of
threads
– All threads run the same code (SPMD)
– Each thread has an ID that it uses to compute
memory addresses and make control decisions
76543210
…
float x = input[threadID];
float y = func(x);
output[threadID] = y;
…
threadID

16
…
float x =
input[threadID];
float y = func(x);
…
threadID
Thread Block 0
…
…
float x =
input[threadID];
float y = func(x);
…
Thread Block 1
…
float x =
input[threadID];
float y = func(x);
…
Thread Block N - 1
Thread Blocks: Scalable Cooperation
• Divide monolithic thread array into multiple blocks
– Threads within a block cooperate via shared memory,
atomic operations and barrier synchronization
– Threads in different blocks cannot cooperate
76543210 76543210 76543210

17
Host
Kernel
1
Kernel
2
Device
Grid 1
Block
(0, 0)
Block
(1, 0)
Block
(0, 1)
Block
(1, 1)
Grid 2
Courtesy:NDVIA
Figure 3.2. An Example of CUDA Thread Org
Block(1, 1)
Thread
(0,1,0)
Thread
(1,1,0)
Thread
(2,1,0)
Thread
(3,1,0)
Thread
(0,0,0)
Thread
(1,0,0)
Thread
(2,0,0)
Thread
(3,0,0)
(0,0,1) (1,0,1) (2,0,1) (3,0,1)
Block IDs and Thread IDs
• Each thread uses IDs to decide
what data to work on
– Block ID: 1D or 2D
– Thread ID: 1D, 2D, or 3D
• Simplifies memory
addressing when processing
multidimensional data
– Image processing
– Solving PDEs on volumes
– …

18
CUDA Memory Model Overview
• Global memory
– Main means of
communicating R/W
Data between host and
device
– Contents visible to all
threads
– Long latency access
• We will focus on global
memory for now
– Constant and texture
memory will come later
Grid
Global Memory
Block (0, 0)
Shared Memory
Thread (0, 0)
Registers
Thread (1, 0)
Registers
Block (1, 0)
Shared Memory
Thread (0, 0)
Registers
Thread (1, 0)
Registers
Host

19
CUDA API Highlights:
Easy and Lightweight
• The API is an extension to the ANSI C programming
language
Low learning curve
• The hardware is designed to enable lightweight
runtime and driver
High performance

20
CUDA Device Memory Allocation
• cudaMalloc()
– Allocates object in the
device Global MemoryGlobal Memory
– Requires two parameters
• Address of a pointer to the
allocated object
• Size of of allocated object
• cudaFree()
– Frees object from device
Global Memory
• Pointer to freed object
Grid
Global
Memory
Block (0, 0)
Shared Memory
Thread (0, 0)
Registers
Thread (1, 0)
Registers
Block (1, 0)
Shared Memory
Thread (0, 0)
Registers
Thread (1, 0)
Registers
Host

21
CUDA Device Memory Allocation (cont.)
• Code example:
– Allocate a 64 * 64 single precision float array
– Attach the allocated storage to Md
– “d” is often used to indicate a device data structure
TILE_WIDTH = 64;
Float* Md
int size = TILE_WIDTH * TILE_WIDTH * sizeof(float);
cudaMalloc((void**)&Md, size);
cudaFree(Md);

22
CUDA Host-Device Data Transfer
• cudaMemcpy()
– memory data transfer
– Requires four parameters
• Pointer to destination
• Pointer to source
• Number of bytes copied
• Type of transfer
– Host to Host
– Host to Device
– Device to Host
– Device to Device
• Asynchronous transfer
Grid
Global
Memory
Block (0, 0)
Shared Memory
Thread (0, 0)
Registers
Thread (1, 0)
Registers
Block (1, 0)
Shared Memory
Thread (0, 0)
Registers
Thread (1, 0)
Registers
Host

23
CUDA Host-Device Data Transfer
(cont.)
• Code example:
– Transfer a 64 * 64 single precision float array
– M is in host memory and Md is in device memory
– cudaMemcpyHostToDevice and
cudaMemcpyDeviceToHost are symbolic constants
cudaMemcpy(Md, M, size, cudaMemcpyHostToDevice);
cudaMemcpy(M, Md, size, cudaMemcpyDeviceToHost);

24
CUDA Keywords

25
CUDA Function Declarations
hosthost__host__ float HostFunc()
hostdevice__global__ void KernelFunc()
devicedevice__device__ float DeviceFunc()
Only callable
from the:
Executed
on the:
• __global__ defines a kernel function
– Must return void
• __device__ and __host__ can be used
together

26
CUDA Function Declarations (cont.)
• __device__ functions cannot have their
address taken
• For functions executed on the device:
– No recursion
– No static variable declarations inside the function
– No variable number of arguments

27
Calling a Kernel Function – Thread Creation
• A kernel function must be called with an execution
configuration:
__global__ void KernelFunc(...);
dim3 DimGrid(100, 50); // 5000 thread blocks
dim3 DimBlock(4, 8, 8); // 256 threads per
block
size_t SharedMemBytes = 64; // 64 bytes of shared
memory
KernelFunc<<< DimGrid, DimBlock, SharedMemBytes
>>>(...);
• Any call to a kernel function is asynchronous from
CUDA 1.0 on, explicit synch needed for blocking

28
A Simple Running Example
Matrix Multiplication
• A simple matrix multiplication example that illustrates
the basic features of memory and thread management
in CUDA programs
– Leave shared memory usage until later
– Local, register usage
– Thread ID usage
– Memory data transfer API between host and device
– Assume square matrix for simplicity

29
Programming Model:
Square Matrix Multiplication Example
• P = M * N of size WIDTH x WIDTH
• Without tiling:
– One thread calculates one element of P
– M and N are loaded WIDTH times from
global memory
M
N
P
WIDTHWIDTH
WIDTH WIDTH

30
M0,2
M1,1
M0,1M0,0
M1,0
M0,3
M1,2 M1,3
Memory Layout of a Matrix in C
M0,2M0,1M0,0 M0,3 M1,1M1,0 M1,2 M1,3 M2,1M2,0 M2,2 M2,3
M2,1M2,0 M2,2 M2,3
M3,1M3,0 M3,2 M3,3
M3,1M3,0 M3,2 M3,3
M

31
Step 1: Matrix Multiplication
A Simple Host Version in C
M
N
P
WIDTHWIDTH
WIDTH WIDTH
// Matrix multiplication on the (CPU) host in double
precision
void MatrixMulOnHost(float* M, float* N, float* P, int Width)
{
for (int i = 0; i < Width; ++i)
for (int j = 0; j < Width; ++j) {
double sum = 0;
for (int k = 0; k < Width; ++k) {
double a = M[i * width + k];
double b = N[k * width + j];
sum += a * b;
}
P[i * Width + j] = sum;
}
}
i
k
k
j

32
void MatrixMulOnDevice(float* M, float* N, float* P, int Width)
{
int size = Width * Width * sizeof(float);
float* Md, Nd, Pd;
…
1. // Allocate and Load M, N to device memory
cudaMalloc(&Md, size);
cudaMemcpy(Md, M, size, cudaMemcpyHostToDevice);
cudaMalloc(&Nd, size);
cudaMemcpy(Nd, N, size, cudaMemcpyHostToDevice);
// Allocate P on the device
cudaMalloc(&Pd, size);
Step 2: Input Matrix Data Transfer
(Host-side Code)

33
Step 3: Output Matrix Data Transfer
(Host-side Code)
2. // Kernel invocation code – to be shown later
…
3. // Read P from the device
cudaMemcpy(P, Pd, size, cudaMemcpyDeviceToHost);
// Free device matrices
cudaFree(Md); cudaFree(Nd); cudaFree (Pd);
}

34
Step 4: Kernel Function
// Matrix multiplication kernel – per thread code
__global__ void MatrixMulKernel(float* Md, float* Nd, float* Pd, int Width)
{
// Pvalue is used to store the element of the matrix
// that is computed by the thread
float Pvalue = 0;

35
Nd
Md Pd
WIDTHWIDTH
WIDTH WIDTH
Step 4: Kernel Function (cont.)
for (int k = 0; k < Width; ++k) {
float Melement = Md[threadIdx.y*Width+k];
float Nelement = Nd[k*Width+threadIdx.x];
Pvalue += Melement * Nelement;
}
Pd[threadIdx.y*Width+threadIdx.x] = Pvalue;
}
ty
tx
ty
tx
k
k

36
// Setup the execution configuration
dim3 dimGrid(1, 1);
dim3 dimBlock(Width, Width);
// Launch the device computation threads!
MatrixMulKernel<<<dimGrid, dimBlock>>>(Md, Nd, Pd, Width);
Step 5: Kernel Invocation
(Host-side Code)

37
Only One Thread Block Used
• One Block of threads compute
matrix Pd
– Each thread computes one
element of Pd
• Each thread
– Loads a row of matrix Md
– Loads a column of matrix Nd
– Perform one multiply and
addition for each pair of Md and
Nd elements
– Compute to off-chip memory
access ratio close to 1:1 (not very
high)
• Size of matrix limited by the
number of threads allowed in a
thread block
Grid 1
Block 1
3 2 5 4
2
4
2
6
48
Thread
(2, 2)
WIDTH
Md Pd
Nd

38
Step 7: Handling Arbitrary Sized Square
Matrices
• Have each 2D thread block to compute
a (TILE_WIDTH)2
sub-matrix (tile) of
the result matrix
– Each has (TILE_WIDTH)2
threads
• Generate a 2D Grid of
(WIDTH/TILE_WIDTH)2
blocks
Md
Nd
Pd
WIDTHWIDTH
WIDTH WIDTH
ty
tx
by
bx
You still need to put a loop
around the kernel call for
cases where
WIDTH/TILE_WIDTH is
greater than max grid size
(64K)!
TILE_WIDTH

39
Some Useful Information on
Tools

40
Compiling a CUDA Program
NVCC
C/C++ CUDA
Application
PTX to Target
Compiler
G80 … GPU
Target code
PTX Code
Virtual
Physical
CPU Code
• Parallel Thread
eXecution (PTX)
– Virtual Machine
and ISA
– Programming
model
– Execution
resources and
state
float4 me = gx[gtid];
me.x += me.y * me.z;
ld.global.v4.f32 {$f1,$f3,$f5,$f7}, [$r9+0];
mad.f32 $f1, $f5, $f3, $f1;

41
Compilation
• Any source file containing CUDA language
extensions must be compiled with NVCC
• NVCC is a compiler driver
– Works by invoking all the necessary tools and
compilers like cudacc, g++, cl, ...
• NVCC outputs:
– C code (host CPU Code)
• Must then be compiled with the rest of the application using another tool
– PTX
• Object code directly
• Or, PTX source, interpreted at runtime

42
Linking
• Any executable with CUDA code requires two
dynamic libraries:
– The CUDA runtime library (cudart)
– The CUDA core library (cuda)

43
Debugging Using the
Device Emulation Mode
• An executable compiled in device emulation mode
(nvcc -deviceemu) runs completely on the host
using the CUDA runtime
– No need of any device and CUDA driver
– Each device thread is emulated with a host thread
• Running in device emulation mode, one can:
– Use host native debug support (breakpoints, inspection, etc.)
– Access any device-specific data from host code and vice-versa
– Call any host function from device code (e.g. printf) and vice-
versa
– Detect deadlock situations caused by improper usage of
__syncthreads

44
Device Emulation Mode Pitfalls
• Emulated device threads execute sequentially, so
simultaneous accesses of the same memory location
by multiple threads could produce different results.
• Dereferencing device pointers on the host or host
pointers on the device can produce correct results in
device emulation mode, but will generate an error in
device execution mode

45
Floating Point
• Results of floating-point computations will slightly
differ because of:
– Different compiler outputs, instruction sets
– Use of extended precision for intermediate results
• There are various options to force strict single precision on the host

Lecture2 cuda spring 2010

More Related Content

What's hot (20)

Similar to Lecture2 cuda spring 2010 (20)

Recently uploaded (20)

Lecture2 cuda spring 2010

Editor's Notes