docs: wordsmith benchmark results expln

[skip ci]
dathere · Oct 10, 2023 · a6a9703 · a6a9703
1 parent 4b0c3b9
commit a6a9703
Showing 1 changed file with 3 additions and 5 deletions.
diff --git a/scripts/results/README.md b/scripts/results/README.md
@@ -1,4 +1,4 @@
-These benchmarks results are compiled on an Apple Mac Mini 2023 model with an M2 Pro chip
+These benchmarks are compiled on an Apple Mac Mini 2023 model with an M2 Pro chip
 with 12 CPU cores & 32GB of RAM; a 1TB SSD primary drive & a 1TB Samsung SSD 970 Evo plus
 external drive running MacOS Sonoma 14.0.
 
@@ -10,8 +10,6 @@ found at https://github.com/jqnatividad/qsv/releases/latest.
 The benchmarks were performed on a 1 million row, 41 column, 520 MB sample of NYC's 311 data.
 https://raw.githubusercontent.com/wiki/jqnatividad/qsv/files/NYC_311_SR_2010-2020-sample-1M.7z
 
-Each benchmark is executed five times. Two warmup runs are performed before the three
-benchmark runs. The mean of the three benchmark runs is used to calculate the results.
+Each benchmark is executed five times. Two warmup runs followed by three benchmark runs.
 
-Records per second was calculated by dividing the number of records (1M) by the mean.
-All other measurements are in seconds.
+Records per second is calculated by dividing the number of records (1M) by the mean of the three benchmark runs. All other measurements are in seconds.