.gitignore

@@ -1,4 +1,4 @@
-**/.Rproj.user
-**/.Rhistory
-**/.RData
-**/.Ruserdata
+.Rproj.user
+.Rhistory
+.RData
+.Ruserdata

w1/README.rmd

@@ -1,89 +0,0 @@
-# Comparison between C-implemented and R-implemented dual-loop matrix summing function performance
-
-## Running
-
-To run this project, run the following commands:
-
-```bash
-# matrix test
-Rscript mat_tests.R
-```
-
-### Building and running
-
-If you edit the C code, to recompile run:
-
-```bash
-bash make_c.sh
-```
-
-### View Evaluation
-
-To install packages necessary for this .rmd document, run:
-
-```bash
-Rscript install_libs.R
-```
-
-## Evaluation 
-The experiment shows the performance comparison between the R-implemented and C-implemented matrix summing functions 
-for different matrix sizes. As the matrix size increases, the C-implemented function demonstrates significantly 
-better performance compared to the R-implemented function.
-Surprisingly, the speedup remains fairly constant in relative terms, stabilizing at about 4x
-
-
-
-
-> Note: Evaluation script run on an AMD Ryzen 9 7950X3D cpu with enough RAM for all matrix sizes
-
-
-| Matrix size | sum1 run duration (secs) | sum2 run duration (secs) |
-|-------------|---------------------------|---------------------------|
-| 5x5         | 7.152557e-06              | 7.867813e-06              |
-| 10x10       | 9.775162e-06              | 5.00679e-06               |
-| 50x50       | 9.346008e-05              | 8.106232e-06              |
-| 100x100     | 0.0003376007              | 1.955032e-05              |
-| 500x500     | 0.007472992               | 0.001415014               |
-| 1000x1000   | 0.03007007                | 0.005748034               |
-| 5000x5000   | 0.6559205                 | 0.1854615                 |
-| 10000x10000 | 2.692389                  | 0.6747584                 |
-| 20000x20000 | 10.67763                  | 2.615553                  |
-| 30000x30000 | 24.33534                  | 5.987761                  |
-
-```{r diagram}
-library(ggplot2)
-library(dplyr)
-library(tidyr)
-
-# prepare data
-data <- tribble(
-  ~Matrix.size, ~R.sum, ~C.sum,
-  "5", 7.152557e-06, 7.867813e-06,
-  "10", 9.775162e-06, 5.00679e-06,
-  "50", 9.346008e-05, 8.106232e-06,
-  "100", 0.0003376007, 1.955032e-05,
-  "500", 0.007472992, 0.001415014,
-  "1000", 0.03007007, 0.005748034,
-  "5000", 0.6559205, 0.1854615,
-  "10000", 2.692389, 0.6747584,
-  "20000", 10.67763, 2.615553,
-  "30000", 24.33534, 5.987761
-)
-
-# Convert Matrix.size to factor with desired order
-data$Matrix.size <- factor(data$Matrix.size, levels = data$Matrix.size)
-
-# rearrange data
-data_long <- data %>%
-  pivot_longer(cols = c(R.sum, C.sum),
-               names_to = "Method",
-               values_to = "Duration")
-
-# Create the plot
-ggplot(data_long, aes(x = Matrix.size, y = Duration, color = Method)) +
-  geom_point() +
-  scale_y_log10() +
-  labs(x = "Matrix Size", y = "Duration (seconds)", color = "Method") +
-  ggtitle("Calculation Time per (square) matrix size") +
-  theme_minimal()
-```

w1/install_libs.R

@@ -1,4 +0,0 @@
-install.packages("rmarkdown")
-install.packages("ggplot2")
-install.packages("dplyr")
-install.packages("tidyr")

w1/make_c.sh

@@ -1 +0,0 @@
-R CMD SHLIB mat.c

w1/mat.R

@@ -1,19 +0,0 @@
-dyn.load("mat.so")
-
-
-sum1 <- function(matrix) {
-    result <- 0
-    for (i in seq_len(nrow(matrix))) {
-        for (j in seq_len(ncol(matrix))) {
-            result <- result + matrix[i, j]
-        }
-    }
-    return(result)
-}
-
-sum2 <- function(matrix) {
-    nrow <- nrow(matrix)
-    ncol <- ncol(matrix)
-    result <- .C("c_sum_matrix", as.double(matrix), as.integer(nrow), as.integer(ncol), result = double(1))$result
-    return(result)
-}

w1/mat.c

@@ -1,12 +0,0 @@
-#include <R.h>
-#include <Rinternals.h>
-
-void c_sum_matrix(double *matrix, int *nrow, int *ncol, double *result) {
-    int i, j;
-    *result = 0.0;
-    for (i = 0; i < *nrow; i++) {
-        for (j = 0; j < *ncol; j++) {
-            *result += matrix[i * (*ncol) + j];
-        }
-    }
-}

w1/mat_tests.R

@@ -1,32 +0,0 @@
-source("mat.R")
-
-time <- function(f) {
-  start_time <- Sys.time()
-  val <- f()
-  end_time <- Sys.time()
-  print(end_time - start_time)
-  return(val)
-}
-
-sizes <- c(5, 10, 50, 100, 500, 1000, 5000, 10000, 20000, 30000)
-# sizes <- c(5, 10, 50, 100, 500, 1000, 5000, 10000)
-
-# warm up
-
-for (size in 30:50) {
-  m <- matrix(runif(size * size), nrow = size, ncol = size)
-  sum1(m)
-  sum2(m)
-}
-
-for (size in sizes) {
-  m <- matrix(runif(size * size), nrow = size, ncol = size)
-  
-  cat(sprintf("Matrix size: %dx%d\n", size, size))
-  cat("sum1 run duration: ")
-  time(function() sum1(m))
-  cat("sum2 run duration: ")
-  time(function() sum2(m))
-  cat("\n")
-}
-