push

quisby/benchmarks/auto_hpl/extract.py

@@ -3,13 +3,21 @@
 from quisby.pricing import cloud_pricing
 from quisby.benchmarks.linpack.extract import linpack_format_data
 
+from quisby.util import read_config
+
 
 def extract_auto_hpl_data(path, system_name):
 
+    summary_data = []
+    server = read_config("server", "name")
+    result_dir = read_config("server", "result_dir")
+
     if path.endswith(".csv"):
         with open(path) as file:
             results = []
             file_data = file.readlines()
+            sum_path = path.split("/./")[1]
+            summary_data.append([system_name, "http://" + server + "/results/" + result_dir + "/" + sum_path])
 
             if len(file_data) > 1:
                 header_row = file_data[-2].strip().split(":")
@@ -24,8 +32,8 @@ def extract_auto_hpl_data(path, system_name):
                 )
 
                 if results:
-                    return results
+                    return results, summary_data
 
             else:
-                return None
+                return None, None
 

quisby/benchmarks/coremark/compare.py

@@ -17,8 +17,9 @@ def extract_prefix_and_number(input_string):
     match = re.search(r'^(.*?)(\d+)(.*?)$', input_string)
     if match:
         prefix = match.group(1)
-        return prefix
-    return None
+        suffix = match.group(3)  # Extracts the suffix after the number
+        return prefix, suffix
+    return None, None
 
 
 def compare_inst(item1, item2):
@@ -33,7 +34,7 @@ def compare_inst(item1, item2):
         return extract_prefix_and_number(item1) == extract_prefix_and_number(item2)
 
 
-def compare_coremark_results(spreadsheets, spreadsheetId, test_name, table_name=["System name","Price/perf"]):
+def compare_coremark_results(spreadsheets, spreadsheetId, test_name, table_name=["System name","Price-perf"]):
     values = []
     results = []
     spreadsheet_name = []
@@ -51,12 +52,13 @@ def compare_coremark_results(spreadsheets, spreadsheetId, test_name, table_name=
         for ele in list_2:
             # Check max throughput
             if value[0][0] in table_name and ele[0][0] in table_name and value[0][0] == ele[0][0]:
-                results.append([""])
-                for item1 in value:
-                    for item2 in ele:
-                        if item1[0] == item2[0]:
-                            results = merge_lists_alternately(results, item1, item2)
-                break
+                if compare_inst(value[1][0], ele[1][0]):
+                    results.append([""])
+                    for item1 in value:
+                        for item2 in ele:
+                            if item1[0] == item2[0]:
+                                results = merge_lists_alternately(results, item1, item2)
+                    break
 
             elif value[0][0] == "Cost/Hr" and ele[0][0] == "Cost/Hr":
                 if compare_inst(value[1][0], ele[1][0]):

quisby/benchmarks/coremark/coremark.py

@@ -1,11 +1,17 @@
 
 """ Custom key to sort the data base don instance name """
+from itertools import groupby
+
 from quisby import custom_logger
 import re
 
 from quisby.util import read_config
 from quisby.pricing.cloud_pricing import get_cloud_pricing
 
+from quisby.util import process_instance
+
+from quisby.util import mk_int
+
 
 def extract_prefix_and_number(input_string):
     match = re.search(r'^(.*?)(\d+)(.*?)$', input_string)
@@ -18,27 +24,29 @@ def extract_prefix_and_number(input_string):
 
 
 def custom_key(item):
-    cloud_type = read_config("cloud","cloud_type")
+    cloud_type = read_config("cloud", "cloud_type")
     if item[1][0] == "localhost":
         return item[1][0]
     elif cloud_type == "aws":
-        instance_type =item[1][0].split(".")[0]
-        instance_number = item[1][0].split(".")[1]
+        instance_name = item[1][0]
+        instance_type = instance_name.split(".")[0]
+        instance_number = instance_name.split(".")[1]
         return instance_type, instance_number
     elif cloud_type == "gcp":
          instance_type = item[1][0].split("-")[0]
          instance_number = int(item[1][0].split('-')[-1])
          return instance_type, instance_number
     elif cloud_type == "azure":
-        instance_type, instance_number, version=extract_prefix_and_number(item[1][0])
-        return instance_type, instance_number
+        instance_type, instance_number, version = extract_prefix_and_number(item[1][0])
+        return instance_type, version, instance_number
 
 
 def calc_price_performance(inst, avg):
     region = read_config("cloud", "region")
     cloud_type = read_config("cloud", "cloud_type")
     os_type = read_config("test", "os_type")
     cost_per_hour = None
+    price_perf = 0.0
     try:
         cost_per_hour = get_cloud_pricing(
             inst, region, cloud_type.lower(), os_type)
@@ -49,53 +57,84 @@ def calc_price_performance(inst, avg):
     return cost_per_hour, price_perf
 
 
-def create_summary_coremark_data(results, OS_RELEASE):
+def group_data(results):
+    cloud_type = read_config("cloud", "cloud_type")
+    if cloud_type == "aws":
+        return groupby(results, key=lambda x: process_instance(x[1][0], "family", "version", "feature", "machine_type"))
+    elif cloud_type == "azure":
+        results = sorted(results, key=lambda x: process_instance(x[1][0], "family", "feature"))
+        return groupby(results, key=lambda x: process_instance(x[1][0], "family", "feature"))
+    elif cloud_type == "gcp":
+        return groupby(results, key=lambda x: process_instance(x[1][0], "family", "version","sub_family","feature"))
+
+
+def sort_data(results):
+    cloud_type = read_config("cloud", "cloud_type")
+    if cloud_type == "aws":
+        results.sort(key=lambda x: str(process_instance(x[1][0], "family")))
+    elif cloud_type == "azure":
+        results.sort(key=lambda x: str(process_instance(x[1][0], "family", "version", "feature")))
+    elif cloud_type == "gcp":
+        results.sort(key=lambda x: str(process_instance(x[1][0], "family", "version", "sub_family")))
+
+def create_summary_coremark_data(results, OS_RELEASE, sorted_results=None):
     final_results = []
-    cal_data = [["System name", "Test_passes-"+OS_RELEASE]]
 
     # Sort data based on instance name
-    sorted_data = sorted(results, key=custom_key)
-    cost_per_hour, price_per_perf = [], []
-
-    # Add summary data
-    for item in sorted_data:
-        sum = 0
-        avg = 0
-        iterations = 0
-        for index in range(3, len(item)):
-            sum = sum + float(item[index][1])
-            iterations = iterations + 1
-        avg = float(sum/iterations)
-        try:
-            cph, pp = calc_price_performance(item[1][0], avg)
-        except Exception as exc:
-            custom_logger.error(str(exc))
-            break
-        cal_data.append([item[1][0], avg])
-        price_per_perf.append([item[1][0], pp])
-        cost_per_hour.append([item[1][0], cph])
-
-    final_results += [[""]]
-    final_results += cal_data
-    final_results.append([""])
-    final_results.append(["Cost/Hr"])
-    final_results += cost_per_hour
-    final_results.append([""])
-    final_results.append(["Price/perf", f"Passes/$-{OS_RELEASE}"])
-    final_results += price_per_perf
+    results = list(filter(None, results))
+    sort_data(results)
+
+    for _, items in group_data(results):
+        cal_data = [["System name", "Test_passes-" + OS_RELEASE]]
+        items = list(items)
+        sorted_data = sorted(items, key=lambda x: mk_int(process_instance(x[1][0], "size")))
+        # sorted_results.extend(sorted_data)
+        cost_per_hour, price_per_perf = [], []
+
+        # Add summary data
+        for item in sorted_data:
+            sum = 0
+            avg = 0
+            iterations = 0
+            for index in range(3, len(item)):
+                sum = sum + float(item[index][1])
+                iterations = iterations + 1
+            avg = float(sum/iterations)
+            try:
+                cph, pp = calc_price_performance(item[1][0], avg)
+            except Exception as exc:
+                custom_logger.error(str(exc))
+                break
+            cal_data.append([item[1][0], avg])
+            price_per_perf.append([item[1][0], pp])
+            cost_per_hour.append([item[1][0], cph])
+        sorted_results=[[""]]
+        sorted_results += cal_data
+        sorted_results.append([""])
+        sorted_results.append(["Cost/Hr"])
+        sorted_results += cost_per_hour
+        sorted_results.append([""])
+        sorted_results.append(["Price-perf", f"Passes/$-{OS_RELEASE}"])
+        sorted_results += price_per_perf
+        final_results.extend(sorted_results)
     return final_results
 
 
 def extract_coremark_data(path, system_name, OS_RELEASE):
     """"""
     results = []
     processed_data =[]
+    summary_data = []
+    server = read_config("server", "name")
+    result_dir = read_config("server", "result_dir")
 
     # Extract data from file
     try:
         if path.endswith(".csv"):
             with open(path) as file:
                 coremark_results = file.readlines()
+            sum_path = path.split("/./")[1]
+            summary_data.append([system_name, "http://" + server + "/results/" + result_dir + "/" + sum_path])
         else:
             return None
     except Exception as exc:
@@ -118,7 +157,7 @@ def extract_coremark_data(path, system_name, OS_RELEASE):
             iteration = iteration + 1
     results.append(processed_data)
 
-    return results
+    return results, summary_data
 
 
 

quisby/benchmarks/coremark/graph.py

@@ -108,7 +108,7 @@ def graph_coremark_data(spreadsheetId, range, action):
             header_row.extend(row)
             title = "%s : %s" % (range, "Test Passes")
             subtitle = "Average Test Passes"
-        elif "Price/perf" in row:
+        elif "Price-perf" in row:
             start_index = index
             header_row.extend(row)
             title = "%s : %s" % (range, "Price-Performance")
@@ -133,7 +133,7 @@ def graph_coremark_data(spreadsheetId, range, action):
                     "chart": {
                         "spec": {
                             "title": title,
-                            "subtitle": subtitle,
+                            "subtitle": subtitle + " : "+graph_data[1][0],
                             "basicChart": {
                                 "chartType": "COMBO",
                                 "legendPosition": "RIGHT_LEGEND",