Add kaval support.

.gitmodules

@@ -7,3 +7,6 @@
 [submodule "external/KaGen"]
 	path = external/KaGen
 	url = https://github.com/KarlsruheGraphGeneration/KaGen.git
+[submodule "kaval"]
+	path = kaval
+	url = https://github.com/niklas-uhl/kaval

eval.R

@@ -0,0 +1,15 @@
+#!/usr/bin/env Rscript
+library(plyr, warn.conflict = FALSE)
+library(dplyr, warn.conflict = FALSE)
+
+aggregator <- function(df) data.frame(AvgTime = mean(df$LabelProbTime, na.rm = TRUE))
+load <- function(file) ddply(read.csv(file), "Graph", aggregator) 
+
+mpi <- load("PlainMPI.csv")
+kamping <- load("KaMPIngWrapper.csv")
+kaminpar <- load("dKaMinParWrapper.csv")
+
+cat("PlainMPI:\t\t", mean(mpi$AvgTime), "\n")
+cat("KaMPIngWrapper:\t\t", mean(kamping$AvgTime), "\n")
+cat("dKaMinParWrapper:\t", mean(kaminpar$AvgTime), "\n")
+

label_propagation.suite.yaml

@@ -0,0 +1,15 @@
+name: label_propagation
+executable: KaMPIngLabelPropagation
+ncores: [1, 2, 4, 8, 16, 48, 192, 768, 3072]
+time_limit: 60
+threads_per_rank: [1]
+graphs:
+  - generator: kagen
+    type: rgg2d
+    N: 20
+    M: 25
+    scale_weak: TRUE
+config:
+  - preset: [mpi, kamping, kaminpar]
+    k: [16]
+    seed: [1, 2, 3]

logs2results.sh

@@ -0,0 +1,30 @@
+#!/usr/bin/env bash
+log_dir="${1:-"."}"
+
+SCRIPT_DIR=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )
+
+configs_for_algo() {
+    cat $log_dir/config.json | \
+         jq ' to_entries | group_by(.value.preset) | map({key: .[0].value.preset, value: [.[] | .key]}) | from_entries' | jq -r " .[\"$1\"] | join (\" \")"
+}
+
+
+parse_logs() {
+    gawk -f "$SCRIPT_DIR/parse_logfile.awk" <(
+    for x in $(configs_for_algo $1); do
+        for file in $log_dir/*-c$x-log.txt; do
+        # echo $file
+        echo "__BEGIN_FILE__" $(basename "$file")
+        cat "$file" | sed "s,\x1B\[[0-9;]*[a-zA-Z],,g"
+        echo "__END_FILE__"
+        done
+    done
+    )
+}
+
+parse_logs "mpi" > PlainMPI.csv
+parse_logs "kamping" > KaMPIngWrapper.csv
+parse_logs "kaminpar" > dKaMinParWrapper.csv
+
+exit 0
+

main.cc

@@ -143,7 +143,7 @@ int main(int argc, char* argv[]) {
                              "kamping-dispatch", "kamping", "kamping-sparse",
                              "kaminpar"}))
       ->required();
-  cli.add_option("-G,--graph", app.graph_filename)->required();
+  cli.add_option("-G,--graph,--kagen_option_string", app.graph_filename)->required();
 
   cli.add_option("-k,--k", app.k, "Number of blocks in the partition.")
       ->capture_default_str();

parse_logfile.awk

@@ -0,0 +1,107 @@
+# -*-awk-*-
+BEGIN {
+    header()
+    reset()
+}
+
+/__BEGIN_FILE__/ {
+    sub(/__BEGIN_FILE__ /, "", $0)
+    split($0, parts, "___")
+    data["Graph"] = parts[1]
+
+    split(parts[2], parts, "_")
+    for (i in parts) {
+        if (match(parts[i], /k([0-9]+)/, m)) {
+            data["K"] = m[1]
+        } else if (match(parts[i], /P([0-9]+)x([0-9]+)x([0-9]+)/, m)) {
+            data["NumNodes"] = m[1]
+            data["NumMPIsPerNode"] = m[2]
+            data["NumThreadsPerMPI"] = m[3]
+        } else if (match(parts[i], /seed([0-9]+)/, m)) {
+            data["Seed"] = m[1]
+        } else if (match(parts[i], /eps([0-9\.\-e]+)/, m)) {
+            data["Epsilon"] = m[1]
+        }
+    }
+
+    level = 0
+}
+
+/__END_FILE__/ {
+    yield()
+}
+
+match($0, /Number of global nodes: *([0-9]+)/, m) {
+    data["N"] = m[1]
+}
+
+match($0, /Number of global edges: *([0-9]+)/, m) {
+    data["M"] = m[1]
+}
+
+match($0, /Seed: *([0-9]+)/, m) {
+    data["Seed"] = m[1]
+}
+
+match($0, /Partition summary:/, m) {
+    Summary = 1
+}
+
+match($0, /Imbalance: *([0-9\.\-e]+)/, m) {
+    data["Balance"] = m[1]
+}
+
+match($0, /Edge cut: *([0-9]+)/, m) {
+    data["Cut"] = m[1]
+}
+
+match($0, /\|- Partitioning: \.* ([0-9\.\-e]+) s/, m) {
+    data["Time"] = m[1]
+}
+match($0, /`- Partitioning: \.* ([0-9\.\-e]+) s/, m) {
+    data["Time"] = m[1]
+}
+
+match($0, /\|- Label propagation clustering: \.* ([0-9\.\-e]+) s/, m) {
+    data["LPTime"] = m[1]
+}
+
+END {
+    yield()
+}
+
+function header() {
+    printf "Graph,"
+    printf "K,"
+    printf "Seed,"
+    printf "Cut,"
+    printf "Epsilon,"
+    printf "Balance,"
+    printf "TotalTime,"
+    printf "LabelProbTime,"
+    printf "NumNodes"
+    printf "\n"
+}
+
+function yield() {
+    if (length(data) == 0) {
+        return
+    }
+
+    printf "%s,", data["Graph"]
+    printf "%d,", data["K"]
+    printf "%d,", data["Seed"]
+    printf "%d,", data["Cut"]
+    printf "%f,", data["Epsilon"]
+    printf "%f,", data["Balance"]
+    printf "%f,", data["Time"]
+    printf "%f,", data["LPTime"]
+    printf "%d", data["NumNodes"]
+    printf "\n"
+
+    reset()
+}
+
+function reset() {
+    split("", data)
+}