From: Ondrej Lichtner <olichtne(a)redhat.com>
In our usecases we're interested in calculating the standard deviation
of the individual averages of the one-level-lower measurements not the
standard deviation of the absolute values of the one-level-lower
measurements. This fixes the issue.
Signed-off-by: Ondrej Lichtner <olichtne(a)redhat.com>
---
lnst/RecipeCommon/PerfResult.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lnst/RecipeCommon/PerfResult.py b/lnst/RecipeCommon/PerfResult.py
index a8d2245..9afac16 100644
--- a/lnst/RecipeCommon/PerfResult.py
+++ b/lnst/RecipeCommon/PerfResult.py
@@ -8,7 +8,7 @@ class PerfStatMixin(object):
@property
def std_deviation(self):
- return std_deviation([i.value for i in self])
+ return std_deviation([i.average for i in self])
class PerfInterval(PerfStatMixin):
def __init__(self, value, duration, unit):
--
2.19.0