tmpCP job

jobs/bprPoisson.sh → jobs/bprNetwork.sh

@@ -1,5 +1,5 @@
 #!/bin/bash
-#SBATCH --job-name=bprPoisson
+#SBATCH --job-name=bprNetwork
 #SBATCH --partition=general1
 #SBATCH --nodes=1
 #SBATCH --ntasks=60
@@ -9,7 +9,7 @@
 #SBATCH --no-requeue
 #SBATCH --mail-type=FAIL
 
-OUTPUTDIR="/scratch/memhierarchy/geis/bpr/experiments/poisson"
+OUTPUTDIR="/scratch/memhierarchy/geis/bpr/experiments/network"
 
 for NUM in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
 do

jobs/bprGeneral.sh → jobs/bprUniform.sh

@@ -1,5 +1,5 @@
 #!/bin/bash
-#SBATCH --job-name=bprGeneral
+#SBATCH --job-name=bprUniform
 #SBATCH --partition=general1
 #SBATCH --nodes=1
 #SBATCH --ntasks=60
@@ -9,7 +9,7 @@
 #SBATCH --no-requeue
 #SBATCH --mail-type=FAIL
 
-OUTPUTDIR="/scratch/memhierarchy/geis/bpr/experiments/general"
+OUTPUTDIR="/scratch/memhierarchy/geis/bpr/experiments/uniform"
 
 for NUM in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
 do

jobs/tmpCP.sh

@@ -0,0 +1,17 @@
+#!/bin/bash
+#SBATCH --job-name=tmpCP
+#SBATCH --partition=general1
+#SBATCH --nodes=1
+#SBATCH --ntasks=20
+#SBATCH --cpus-per-task=1
+#SBATCH --mem-per-cpu=2000
+#SBATCH --time=100:00:00
+#SBATCH --no-requeue
+#SBATCH --mail-type=FAIL
+
+OUTPUTDIR="/scratch/memhierarchy/geis/bpr/experiments/network"
+
+for NUM in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
+do
+    target/release/bpr --log "${OUTPUTDIR}/cov/OUT_${NUM}" --na $((16 * $NUM)) --nb $((400 * $NUM)) --vs 10 --iterations 500 --instances 10 --goal COV --algorithm ALL --parameters 4 --poisson &
+done

scripts/fitTime.py

@@ -0,0 +1,34 @@
+import pickle
+from math import log
+import matplotlib.pyplot as plt
+
+PATH = "logs/time/T_COV_G"
+
+def fun(na: int, nb: int) -> int:
+    return na * na * na * na * nb
+
+with open(PATH, "rb") as f:
+    data = pickle.load(f)
+
+keys = data["x"]
+
+values = [
+    data["y"][key][0] for key in range(len(keys))
+]
+
+factors = [
+    values[key] / fun(keys[key] * 16, keys[key] * 16) for key in range(len(keys))    
+]
+
+factor = sum(factors) / len(keys)
+
+fitting = [
+    factor * fun(keys[key] * 16, keys[key] * 16) for key in range(len(keys))
+]
+
+plt.plot(keys, values, label = "Values")
+plt.plot(keys, fitting, label = "Fitting")
+plt.yscale("log")
+plt.legend()
+
+plt.show()

scripts/timePlot.py

@@ -38,7 +38,7 @@
 
 values = np.array(inverted_array)
 
-data = pd.DataFrame(values, keys, columns = ["OPT - General", "OPT - Poisson", "AMP - General", "AMP - Poisson", "NAMP - General", "NAMP - Poisson"])
+data = pd.DataFrame(values, keys, columns = ["OPT - Uniform", "OPT - Network", "AMP - Uniform", "AMP - Network", "NAMP - Uniform", "NAMP - Network"])
 
 plot = sns.lineplot(data = data, palette = "Paired", dashes = False, linewidth = 2.5)