Introduction to CBEA SDK

Moscow-Bavarian Joint Advanced Student School19-29 March 2006, Moscow, Russia

Introduction to CBEA SDK

Veselin Dikov

Moscow-Bavarian Joint Advanced Student School19-29. March 2006

Moscow, Russia


Outline

● Getting started● SPU Language Extensions● SPE Library● SDK Libraries● Remote Procedure Calls


Getting started

● Two executables formats: ppu vs. spu Two different compilers: ppu-gcc and spu-gcc Check the format of an executable

# file <executable>

● Makefile headers Reside in SDK root directory

make.headermake.footermake.env


Getting started

# Subdirectories ###############DIRS := spu

# Target #######################PROGRAM_ppu:= simple

# Local Defines ################

IMPORTS := spu/lib_simple_spu.a \-lspe

# imports the embedded simple_spu # library allows consolidation of # spu program into ppe binary

# make.footer ##################include ../make.footer

# make.footer is in the top of # the SDK

# Target #######################PROGRAMS_spu := simple_spu# created embedded libraryLIBRARY_embed:= lib_simple_spu.a

# Local Defines################################IMPORTS = $(SDKLIB_spu)/libc.a

# make.footer ################################include ../../make.footer

# make.footer is in the top of # the SDK

● Makefile - ppu ● Makefile - spu


Outline



SPU Language Extensions

● Provides SPU functionality● Extended RISC instruction set● Extended Data types: vector● C/C++ intrinsic routines● Memory Flow Control (MFC)


SIMD Vectorization

● vector - Reserved as a C/C++ keyword● 128 bits of size

16 bytes, or 4 32-bit words, or etc.

● Single instruction acts on the entire vector

vector unsigned int vec = (vector unsigned int)

(1,2,3,4);vector unsigned int v_ones =

(vector unsigned int)(1);vector unsigned int vdest =

spu_add(vec, v_ones);


SIMD Vectorization

● Overridden C/C++ operators sizeof() (=16 for any vector) operator= operator&

● Various intrinsic functions for vectors Arithmetical – spu_add, spu_mult, etc. Logical – spu_and, spu_nor, etc. Comparison – spu_cmpeq Scalar – spu_extract, spu_insert, spu_promote Etc.


MFC routines

● Local Store (LS) vs. Effective Address (EA)● Data transport via DMA

Up to 16,384 bytes on DMA message Asynchronous Part of the instruction set, C/C++ intrinsics

mfc_get(&data,addr+16384*i,16384,20,0,0);

● What is __attribute__ ((aligned (128)));?

LS addressEA address


MFC example

int main() {

…

/* the array needs to be aligned on a 128-byte cache line */ data = (int *) malloc(127 + DATASIZE*sizeof(int)); while (((int) data) & 0x7f) ++data;

…

spe_create_thread(gid,&sample_spu,data,NULL,-1,0);

● ppu main file


MFC example

int main() {

…

/* the array needs to be aligned on a 128-byte cache line */ data = (int *)malloc_align(DATASIZE*sizeof(int),7);

…

spe_create_thread(gid,&sample_spu,data,NULL,-1,0);

● ppu main file


MFC example

int databuf [DATASIZE]__attribute__ ((aligned (128)));

int main(void * arg) { int *data = &databuf[0];

…

mfc_get(data,arg,DATASIZE*sizeof(int),20,0,0); mfc_write_tag_mask(1<<20); mfc_read_tag_status_all();

● spu main file


Outline



● Provides PPU functionality● Two sets of functions

Thread management – POSIX like MFC access functions – access to mailboxes

● Header: <libspe.h>


Thread management

● Threads are organized in thread groups

Thread Group 0

Th0 Th1 Th2 ThNThread Group 1

Th0 Th1 Th2 ThN

Thread scheduling:•SCHED_RR•SCHED_FIFO•SCHED_OTHER

Priority:•RR and FIFO – 1 to 99•OTHER – 0 to 99


Thread management

● Threads are organized in thread groups

spe_gid_t gid; // group handle speid_t speids[1]; // thread handle

// Create an SPE groupgid = spe_create_group (SCHED_OTHER, 0, 1);if (gid == NULL) exit(-1); // failed

// allocate the SPE taskspeids[0] = spe_create_thread(gid,&sample_spu,0,0,-1,0);if (speids[0] == NULL) exit (-1);

// wait for the single SPE to completespe_wait(speids[0], &status[0], 0);


Thread management

● spe_get_affinity, spe_set_affinity 8-bit mask specifying SPE units where to start the

thread

● spe_get_ls Gets access to SPU’s local store Dirty!! Not supported by all operating systems Used for RPC communication

● spe_get/set_context, spe_get_ps


Thread management

● Two ways of communicating with threads via signals via mailboxes

● POSIX-like Signals spe_get_event - check for signals from a thread spe_kill – send signal to a thread spe_write_signal – writes a signal through MFC

● MFC mailboxes spe_read_out_mbox, spe_write_in_mbox spe_stat_in_mbox, spe_stat_out_mbox,

spe_stat_out_intr_mbox


Outline



SDK Libraries

●The SDK comes with various applied libraries

Library name Short Description PPE SPE

C Library standard C99 functionality. POSIX.1 functions. x x

Audio Resample Library

audio resampling functionality for PPE and SPE x x

Curves and Surfaces Library

quadratic and cubic Bezier curves. Biquadric and bicubic Bezier surfaces, and curved point-normal triangles.

x x

FFT Library 1-D FFT and kernel functions for 2-D FFT x x

Game Math Library

math routines applicable to game performance needs x x

Image Library routines for processing images - convolutions and histograms

x x

Large Matrix Library

basic linear algebra routines on large vectors and matrices

x

Math Library general purpose math routines tuned to exploit SIMD x x


SDK Libraries

●The SDK comes with various applied libraries

Library name Short Description PPE SPE

Matrix Library routines for operations on 4x4 Matrices and quaternions x x

Misc Library set of general purpose routines that don’t logically fit within any other

x x

Multi-Precision Math Library

operations on unsigned integer numbers with large number of bits

x

Noise LibraryPPE

1-,2-, 3-, 4-D noise, Lattice and non-lattice noise, Turbulance

x x

Oscillator Libraries

definition of sound sources x x

Simulation Library

functionality related to the Full-Simulator - -

Sync Library synchronization primitives, like atomic operations, mutex x x

Vector Library a set of general purpose routines that operate on vectors. x x


Exampletypedef union {

vector float fv;vector unsigned int uiv;unsigned int ui[4];float f[4];

} v128;typedef vector float floatvec;

…

v128 A[4] = {(floatvec)(0.9501,0.8913,0.8214,0.9218), (floatvec)(0.2311,0.7621,0.4447,0.7382), (floatvec)(0.6068,0.4565,0.6154,0.1763), (floatvec)(0.4860,0.0185,0.7919,0.4057)};

v128 invA[4];v128 x, y;x.fv = (floatvec)(1, 5, 6, 7);y.fv = (floatvec)(0);

inverse_matrix4x4(&invA[0].fv, &A[0].fv);

madd_vector_matrix(4,4,&invA[0].fv,4,(float *)&x,(float *)&y);

yAx ● Solving

from Matrix library

from LargeMatrix library. Solves:

y = A*x + y


Outline



Remote Procedure Calls

● Function-Offload Model SPE threads as services PPE communicates with them thought RPC

calls● Remote Procedure Calls

stubs



● Preparing stubs Interface Description Language (IDL) idl files

interface add{ import "../stub.h"; const int ARRAY_SIZE = 1000; [sync] idl_id_t do_inv ([in] int array_size,

[in, size_is(array_size)] int array_a[], [out, size_is(array_size)] int array_res[]);

…}

idl compiler# idl -p ppe_sample.c -s spe_sample.c sample.idl



● Preparing stubs


Example

● .idl file

interface add{ import "../stub.h"; const int ARRAY_SIZE = 1000; [sync] idl_id_t do_add ([in] int array_size, [in, size_is(array_size)] int array_a[], [in, size_is(array_size)] int array_b[], [out, size_is(array_size)] int array_c[]);

[sync] idl_id_t do_sub ([in] int array_size, [in, size_is(array_size)] int array_a[], [in, size_is(array_size)] int array_b[], [out, size_is(array_size)] int array_c[]);}


Example

● spe do_add.c file (user file)#include "../stub.h"

idl_id_t do_add ( int array_size, int array_a[], int array_b[], int array_c[])

{ int i; for (i = 0; i < array_size; i++) { array_c[i] = array_a[i] + array_b[i]; } return 0;}

idl_id_t do_sub ( int array_size, int array_a[], int array_b[], int array_c[])

{ ... return 0;}


Example

● ppe main.c file (user file)

#include "../stub.h"

int array_a[ARRAY_SIZE] __attribute ((aligned(128)));int array_b[ARRAY_SIZE] __attribute ((aligned(128)));int array_add[ARRAY_SIZE] __attribute ((aligned(128)));

int main(){ ...

do_add (ARRAY_SIZE, array_a, array_b, array_add); ...}


Conclusion

● Powerful SDK vector data type Extensive Matrix routines Extensive Signal Processing routines

● Game-industry driven But applicable for scientific work as well


References

● CBEA-Tutorial.pdf, SDK documentation● idl.pdf, SDK documentation● libraries_SDK.pdf, SDK Documentation● libspe_v1.0.pdf, SDK Documentation● SPU_language_extensions_v21.pdf

Sony online resources http://cell.scei.co.jp/pdf/SPU_language_extensions_v21.pdf, 15.03.2006

Thank you for your attention!

Introduction to CBEA SDK

Documents

Transcript of Introduction to CBEA SDK