From 5bd4dd3964c798ae37f94d77a42d15b44546ff11 Mon Sep 17 00:00:00 2001
From: Armando Luiz Nicolini Delgado <nicolui@inf.ufpr.br>
Date: Mon, 21 Jun 2021 14:47:58 -0300
Subject: [PATCH] =?UTF-8?q?Novos=20arquivos=20para=20exerc=C3=ADcio.?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
---
HP_EliteDesk-Lab3.txt | 117 ++++++++++++++++++++
Makefile | 44 ++++++++
README.md | 4 +
matmult.c | 89 +++++++++++++++
matriz.c | 246 ++++++++++++++++++++++++++++++++++++++++++
matriz.h | 35 ++++++
perfctr | 12 +++
7 files changed, 547 insertions(+)
create mode 100644 HP_EliteDesk-Lab3.txt
create mode 100644 Makefile
create mode 100644 matmult.c
create mode 100644 matriz.c
create mode 100644 matriz.h
create mode 100755 perfctr
diff --git a/HP_EliteDesk-Lab3.txt b/HP_EliteDesk-Lab3.txt
new file mode 100644
index 0000000..cfc0149
--- /dev/null
+++ b/HP_EliteDesk-Lab3.txt
@@ -0,0 +1,117 @@
+--------------------------------------------------------------------------------
+CPU name: Intel(R) Core(TM) i5-7500 CPU @ 3.40GHz
+CPU type: Intel Coffeelake processor
+CPU clock: 3.41 GHz
+CPU family: 6
+CPU model: 158
+CPU short: skylake
+CPU stepping: 9
+CPU features:
+--------------------------------------------------------------------------------
+ Group name Description
+--------------------------------------------------------------------------------
+ ICACHE Instruction cache miss rate/ratio
+ PORT_USAGE Execution port utilization
+ FLOPS_SP Single Precision MFLOP/s
+ TLB_INSTR L1 Instruction TLB miss rate/ratio
+ TLB_DATA L2 data TLB miss rate/ratio
+ UOPS_ISSUE UOPs issueing
+ UOPS_RETIRE UOPs retirement
+ FLOPS_DP Double Precision MFLOP/s
+ TMA Top down cycle allocation
+ L2CACHE L2 cache miss rate/ratio
+CYCLE_ACTIVITY Cycle Activities
+ CLOCK Power and Energy consumption
+ FLOPS_AVX Packed AVX MFLOP/s
+ L3 L3 cache bandwidth in MBytes/s
+ FALSE_SHARE False sharing
+ DIVIDE Divide unit information
+ CYCLE_STALLS Cycle Activities (Stalls)
+ UOPS_EXEC UOPs execution
+ L3CACHE L3 cache miss rate/ratio
+ L2 L2 cache bandwidth in MBytes/s
+ DATA Load to store ratio
+ RECOVERY Recovery duration
+ BRANCH Branch prediction miss rate/ratio
+ UOPS UOPs execution info
+ ENERGY Power and Energy consumption
+
+********************************************************************************
+Hardware Thread Topology
+********************************************************************************
+Sockets: 1
+Cores per socket: 4
+Threads per core: 1
+--------------------------------------------------------------------------------
+HWThread Thread Core Socket Available
+0 0 0 0 *
+1 0 1 0 *
+2 0 2 0 *
+3 0 3 0 *
+--------------------------------------------------------------------------------
+Socket 0: ( 0 1 2 3 )
+--------------------------------------------------------------------------------
+********************************************************************************
+Cache Topology
+********************************************************************************
+Level: 1
+Size: 32 kB
+Type: Data cache
+Associativity: 8
+Number of sets: 64
+Cache line size: 64
+Cache type: Non Inclusive
+Shared by threads: 1
+Cache groups: ( 0 ) ( 1 ) ( 2 ) ( 3 )
+--------------------------------------------------------------------------------
+Level: 2
+Size: 256 kB
+Type: Unified cache
+Associativity: 4
+Number of sets: 1024
+Cache line size: 64
+Cache type: Non Inclusive
+Shared by threads: 1
+Cache groups: ( 0 ) ( 1 ) ( 2 ) ( 3 )
+--------------------------------------------------------------------------------
+Level: 3
+Size: 6 MB
+Type: Unified cache
+Associativity: 12
+Number of sets: 8192
+Cache line size: 64
+Cache type: Inclusive
+Shared by threads: 4
+Cache groups: ( 0 1 2 3 )
+--------------------------------------------------------------------------------
+********************************************************************************
+NUMA Topology
+********************************************************************************
+NUMA domains: 1
+--------------------------------------------------------------------------------
+Domain: 0
+Processors: ( 0 1 2 3 )
+Distances: 10
+Free memory: 3774.68 MB
+Total memory: 7858 MB
+--------------------------------------------------------------------------------
+
+
+********************************************************************************
+Graphical Topology
+********************************************************************************
+Socket 0:
++---------------------------------------------+
+| +--------+ +--------+ +--------+ +--------+ |
+| | 0 | | 1 | | 2 | | 3 | |
+| +--------+ +--------+ +--------+ +--------+ |
+| +--------+ +--------+ +--------+ +--------+ |
+| | 32 kB | | 32 kB | | 32 kB | | 32 kB | |
+| +--------+ +--------+ +--------+ +--------+ |
+| +--------+ +--------+ +--------+ +--------+ |
+| | 256 kB | | 256 kB | | 256 kB | | 256 kB | |
+| +--------+ +--------+ +--------+ +--------+ |
+| +-----------------------------------------+ |
+| | 6 MB | |
+| +-----------------------------------------+ |
++---------------------------------------------+
diff --git a/Makefile b/Makefile
new file mode 100644
index 0000000..e2d1543
--- /dev/null
+++ b/Makefile
@@ -0,0 +1,44 @@
+ PROG = matmult
+ MODULOS = matriz
+
+ CC = gcc -std=c11 -g
+ OBJS = $(addsuffix .o,$(MODULOS))
+
+ LIKWID = /home/soft/likwid
+ LIKWID_FLAGS = -I$(LIKWID)/include
+ LIKWID_LIBS = -L$(LIKWID)/lib
+
+ AVX_FLAGS = -march=native -mavx -O3 -falign-functions=32 -falign-loops=32 -fstrict-aliasing
+ AVX_LOG_FLAGS = -fopt-info-vec -fopt-info-vec-missed
+
+ CFLAGS = $(LIKWID_FLAGS)
+ LFLAGS = $(LIKWID_LIBS) -lm
+
+.PHONY: all clean limpa purge faxina distclean debug avx likwid
+
+%.o: %.c %.h
+ $(CC) $(CFLAGS) -c $<
+
+all: $(PROG)
+
+debug: CFLAGS += -DDEBUG
+
+avx: CFLAGS += $(AVX_FLAGS) $(AVX_LOG_FLAGS)
+avx likwid: CFLAGS += -DLIKWID_PERFMON
+avx likwid: LFLAGS += -llikwid
+
+likwid avx debug: $(PROG)
+
+$(PROG): $(PROG).o
+
+$(PROG): $(OBJS)
+ $(CC) $(CFLAGS) -o $@ $^ $(LFLAGS)
+
+clean:
+ @echo "Limpando ...."
+ @rm -f *~ *.bak *.tmp
+
+purge distclean: clean
+ @echo "Faxina ...."
+ @rm -f $(PROG) *.o core a.out
+ @rm -f *.png marker.out *.log
diff --git a/README.md b/README.md
index f418fbe..ebbf64e 100644
--- a/README.md
+++ b/README.md
@@ -1 +1,5 @@
O enunciado do exercício está <A HREF="https://moodle.c3sl.ufpr.br/mod/assign/view.php?id=24939">aqui</a>
+<BR>
+<BR>
+O arquivo <B>perfctr</B> é um <I>script</I> shell para facilitar o uso de <I>likwid-perfctr</I>.
+<BR><BR>
diff --git a/matmult.c b/matmult.c
new file mode 100644
index 0000000..91595cc
--- /dev/null
+++ b/matmult.c
@@ -0,0 +1,89 @@
+#include <stdio.h>
+#include <stdlib.h> /* exit, malloc, calloc, etc. */
+#include <string.h>
+#include <getopt.h> /* getopt */
+
+#include <likwid.h>
+
+#include "matriz.h"
+
+/**
+ * Exibe mensagem de erro indicando forma de uso do programa e termina
+ * o programa.
+ */
+
+static void usage(char *progname)
+{
+ fprintf(stderr, "Forma de uso: %s [ -n <ordem> ] \n", progname);
+ exit(1);
+}
+
+
+
+/**
+ * Programa principal
+ * Forma de uso: matmult [ -n <ordem> ]
+ * -n <ordem>: ordem da matriz quadrada e dos vetores
+ *
+ */
+
+int main (int argc, char *argv[])
+{
+ int c, n=DEF_SIZE;
+ double norma;
+
+ MatPtr mPtr;
+ MatRow mRow;
+ Vetor vet, resPtr, resRow, resCol;
+
+ /* =============== TRATAMENTO DE LINHA DE COMAANDO =============== */
+
+ char *opts = "n:";
+ c = getopt (argc, argv, opts);
+
+ while ( c != -1 ) {
+ switch (c) {
+ case 'n': n = atoi(optarg); break;
+ default: usage(argv[0]);
+ }
+
+ c = getopt (argc, argv, opts);
+ }
+
+ /* ================ FIM DO TRATAMENTO DE LINHA DE COMANDO ========= */
+
+ resPtr = (double *) malloc (n * sizeof(double));
+ resRow = (double *) malloc (n * sizeof(double));
+
+ srand(20191);
+
+ mPtr = geraMatPtr (n, n);
+ mRow = geraMatRow (n, n);
+ vet = geraVetor (n);
+
+#ifdef DEBUG
+ prnMatPtr (mPtr, n, n);
+ prnMatRow (mRow, n, n);
+ prnVetor (vet, n);
+ printf ("=================================\n\n");
+#endif /* DEBUG */
+
+ multMatPtrVet (mPtr, vet, n, n, resPtr);
+
+ multMatRowVet (mRow, vet, n, n, resRow);
+
+#ifdef DEBUG
+ prnVetor (resPtr, n);
+ prnVetor (resRow, n);
+#endif /* DEBUG */
+
+ liberaMatPtr (mPtr, n);
+ liberaVetor ((void*)mRow);
+ liberaVetor ((void*)vet);
+
+ free(resRow);
+ free(resPtr);
+
+ return 0;
+}
+
diff --git a/matriz.c b/matriz.c
new file mode 100644
index 0000000..8f7508a
--- /dev/null
+++ b/matriz.c
@@ -0,0 +1,246 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <math.h>
+
+#include "matriz.h"
+
+/**
+ * Função que gera valores para para ser usado em uma matriz
+ * @param i,j coordenadas do elemento a ser calculado (0<=i,j<n)
+* @return valor gerado para a posição i,j
+ */
+static inline double generateRandomA( unsigned int i, unsigned int j)
+{
+ static double invRandMax = 1.0 / (double)RAND_MAX;
+ return ( (i==j) ? (double)(BASE<<1) : 1.0 ) * (double)rand() * invRandMax;
+}
+
+/**
+ * Função que gera valores aleatórios para ser usado em um vetor
+ * @return valor gerado
+ *
+ */
+static inline double generateRandomB( )
+{
+ static double invRandMax = 1.0 / (double)RAND_MAX;
+ return (double)(BASE<<2) * (double)rand() * invRandMax;
+}
+
+
+
+/* ----------- FUNÇÕES ---------------- */
+
+/**
+ * Funcao geraMatPtr: gera matriz como vetor de ponteiros para as suas linhas
+ *
+ * @param m número de linhas da matriz
+ * @param n número de colunas da matriz
+ * @return ponteiro para a matriz gerada
+ *
+ */
+
+MatPtr geraMatPtr (int m, int n)
+{
+ MatPtr matriz = (double **) malloc(m*sizeof(double));
+
+ if (matriz) {
+ for (int i=0; i < m; ++i) {
+ if (matriz[i] = (double *) malloc(n*sizeof(double)))
+ for (int j=0; matriz[i] && j < n; ++j)
+ matriz[i][j] = generateRandomA(i, j);
+ else
+ return NULL;
+ }
+ }
+
+ return (matriz);
+}
+
+/**
+ * \brief: libera matriz alocada como vetor de ponteiros para as suas linhas
+ *
+ * @param ponteiro para matriz
+ *
+ */
+void liberaMatPtr (MatPtr matriz, int m)
+{
+ if (matriz) {
+ for (int i=0; i < m; ++i) {
+ free (matriz[i]);
+ }
+ free (matriz);
+ }
+}
+
+
+
+/**
+ * Funcao geraMatRow: gera matriz como vetor único, 'row-oriented'
+ *
+ * @param m número de linhas da matriz
+ * @param n número de colunas da matriz
+ * @return ponteiro para a matriz gerada
+ *
+ */
+
+MatRow geraMatRow (int m, int n)
+{
+ MatRow matriz = (double *) malloc(m*n*sizeof(double));
+
+ if (matriz) {
+ for (int i=0; i < m; ++i) {
+ for (int j=0; j < n; ++j)
+ matriz[i*m + j] = generateRandomA(i, j);
+ }
+ }
+
+ return (matriz);
+}
+
+
+/**
+ * Funcao geraVetor: gera vetor de tamanho 'n'
+ *
+ * @param n número de elementos do vetor
+ * @return ponteiro para vetor gerado
+ *
+ */
+
+Vetor geraVetor (int n)
+{
+ Vetor vetor = (double *) malloc(n*sizeof(double));
+
+ if (vetor)
+ for (int i=0; i < n; ++i)
+ vetor[i] = generateRandomB();
+
+ return (vetor);
+}
+
+/**
+ * \brief: libera vetor
+ *
+ * @param ponteiro para vetor
+ *
+ */
+void liberaVetor (void *vet)
+{
+ free(vet);
+}
+
+
+/**
+ * Funcao multMatPtrVet: Efetua multiplicacao entre matriz 'mxn' por vetor
+ * de 'n' elementos
+ * @param mat matriz 'mxn'
+ * @param m número de linhas da matriz
+ * @param n número de colunas da matriz
+ * @param res vetor que guarda o resultado. Deve estar previamente alocado e com
+ * seus elementos inicializados em 0.0 (zero)
+ * @return vetor de 'm' elementos
+ *
+ */
+
+void multMatPtrVet (MatPtr mat, Vetor v, int m, int n, Vetor res)
+{
+
+ /* Efetua a multiplicação */
+ if (res) {
+ for (int i=0; i < m; ++i)
+ for (int j=0; j < n; ++j)
+ res[i] += mat[i][j] * v[j];
+ }
+}
+
+/**
+ * Funcao prnMatPtr: Imprime o conteudo de uma matriz em stdout
+ * @param mat matriz
+ * @param m número de linhas da matriz
+ * @param n número de colunas da matriz
+ *
+ */
+
+void prnMatPtr (MatPtr mat, int m, int n)
+{
+ for (int i=0; i < m; ++i) {
+ for (int j=0; j < n; ++j)
+ printf(DBL_FIELD, mat[i][j]);
+ printf("\n");
+ }
+ printf(SEP_RES);
+}
+
+/**
+ * Funcao multMatRowVet: Efetua multiplicacao entre matriz 'mxn' por vetor
+ * de 'n' elementos
+ * @param mat matriz 'mxn'
+ * @param m número de linhas da matriz
+ * @param n número de colunas da matriz
+ * @param res vetor que guarda o resultado. Deve estar previamente alocado e com
+ * seus elementos inicializados em 0.0 (zero)
+ * @return vetor de 'm' elementos
+ *
+ */
+
+void multMatRowVet (MatRow mat, Vetor v, int m, int n, Vetor res)
+{
+
+ /* Efetua a multiplicação */
+ if (res) {
+ for (int i=0; i < m; ++i)
+ for (int j=0; j < n; ++j)
+ res[i] += mat[m*i + j] * v[j];
+ }
+}
+
+/**
+ * Funcao prnMatRow: Imprime o conteudo de uma matriz em stdout
+ * @param mat matriz
+ * @param m número de linhas da matriz
+ * @param n número de colunas da matriz
+ *
+ */
+
+void prnMatRow (MatRow mat, int m, int n)
+{
+ for (int i=0; i < m; ++i) {
+ for (int j=0; j < n; ++j)
+ printf(DBL_FIELD, mat[m*i + j]);
+ printf("\n");
+ }
+ printf(SEP_RES);
+}
+
+/**
+ * Funcao prodEscalar: Calcula o produto escalar entre 2 vetores
+ * @param v1 vetor com 'n' elementos
+ * @param v2 vetor com 'n' elementos
+ * @param n número de elementos dos vetores
+ * @return Valor do produto escalar
+ */
+
+double prodEscalar (Vetor v1, Vetor v2, int n)
+{
+ double prod = 0.0;
+
+ for (int i=0; i < n; ++i)
+ prod += v1[i]*v2[i];
+
+ return prod;
+}
+
+
+/**
+ * Funcao prnVetor: Imprime o conteudo de vetor em stdout
+ * @param vet vetor com 'n' elementos
+ * @param n número de elementos do vetor
+ *
+ */
+
+void prnVetor (Vetor vet, int n)
+{
+ for (int i=0; i < n; ++i)
+ printf(DBL_FIELD, vet[i]);
+ printf(SEP_RES);
+}
+
diff --git a/matriz.h b/matriz.h
new file mode 100644
index 0000000..11a2025
--- /dev/null
+++ b/matriz.h
@@ -0,0 +1,35 @@
+/* Constantes */
+
+#define DBL_FIELD "%12.7lg"
+#define SEP_RES "\n\n\n"
+
+#define DEF_SIZE 128
+#define BASE 32
+
+
+#define ABS(num) ((num) < 0.0 ? -(num) : (num))
+
+/* Implementações para matrizes e vetores */
+
+typedef double ** MatPtr;
+typedef double * MatRow;
+typedef double * Vetor;
+
+/* ----------- FUNÇÕES ---------------- */
+
+MatPtr geraMatPtr (int m, int n);
+MatRow geraMatRow (int m, int n);
+Vetor geraVetor (int n);
+
+void liberaMatPtr (MatPtr mPtr, int n);
+void liberaVetor (void *vet);
+
+void multMatPtrVet (MatPtr mat, Vetor v, int m, int n, Vetor res);
+void prnMatPtr (MatPtr mat, int m, int n);
+
+void multMatRowVet (MatRow mat, Vetor v, int m, int n, Vetor res);
+void prnMatRow (MatRow mat, int m, int n);
+
+double prodEscalar (Vetor v1, Vetor v2, int n);
+void prnVetor (Vetor vet, int n);
+
diff --git a/perfctr b/perfctr
new file mode 100755
index 0000000..24db0a1
--- /dev/null
+++ b/perfctr
@@ -0,0 +1,12 @@
+#!/bin/bash
+
+LIKWID_CMD="likwid-perfctr -C $1 -g $2 -m"
+shift 2
+
+${LIKWID_CMD} $*
+
+# Para obter lista de grupos de indicadores de performance:
+# likwid-perfctr -a
+
+# Para obter topologia dos processadores
+# likwid-topology -c -g
--
GitLab