Update SDRBench metadata files

metadata/SDRBENCH-QMCPACK-115x69x69x288-property.txt

@@ -0,0 +1,14 @@
+=============================================
+The property of einspline_115_69_69_288.f32 :
+This is little-endian system.
+The first 10 values are: 
+-2.341361 -2.341361 2.177117 2.496278 -2.342006 2.342006 2.529034 -2.141112 2.177645 2.496978 ....
+------------------------
+numOfElem = 157684320
+totalDataSize = 630737280 bytes (601.517944 MB)
+min = -16.505009
+max = 16.505009
+valueRange = 33.010017
+avgValue = 0.008473
+entropy = 7.565581
+zeromean_variance = 0.561939

metadata/template_data.txt

@@ -0,0 +1,29 @@
+DATASET NAME: QMCPACK
+
+DATASET OWNER’S NAME AND CONTACT: Dingwen Tao (dingwen.tao@anl.gov)
+
+ARRAY DIMENSION (1D, 2D,…): 4D 
+
+Dimension sizes: 
+original data: 115x69x69x288
+preconditioned data: 288x115x69x69
+
+DATA NATURE (INTEGER, DECIMAL,…): DECIMAL
+ 
+ROW AND COLUMN NAMES:  
+
+PHYSICAL DESCRIPTION OF THE DATASET (DIMENSION 1: TIME, DIMENSION 2: SPACE X, DIMENSION 3: SPACE Y, DIMENSION 4: TEMPERATURE, …:
+			DIMENSION 1: x, DIMENSION 2: y, DIMENSION 3: z, DIMENSION 4: orbital index
+
+ENDIAN FORMAT (LITTLE, BIG): LITTLE
+
+PRECISION (DOUBLE, SINGLE): SINGLE
+
+ADDITIONAL GRAPHS: 
+
+
+Further description: 
+The original raw data are 115x69x69x288 in dimensions. 
+However, according to the users, the data would become smoother in space after preconditioning, so they suggest to compress preconditioned data instead of original data. 
+They are the same dataset, and the only difference is the different orders of the data in space. 
+