examples: update readme files

examples/mixed/README.md

@@ -0,0 +1,73 @@
+This example shows a benchmark that consists of multiple phases (5 seconds
+running time each), benchmarked with 32 threads.
+
+The workload file, `mixed.toml` is as follows:
+
+```toml
+[global]
+threads = 1
+repeat = 5
+klen = 8
+vlen = 16
+kmin = 0
+kmax = 1000000
+report = "repeat"
+
+[[benchmark]]
+set_perc = 100
+get_perc = 0
+del_perc = 0
+repeat = 1
+dist = "incrementp"
+report = "hidden"
+
+# write-intensive, zipfian
+[[benchmark]]
+set_perc = 50
+get_perc = 50
+del_perc = 0
+timeout = 1.0
+dist = "zipfian"
+
+# write-intensive, zipfian, hotspot in middle
+[[benchmark]]
+set_perc = 50
+get_perc = 50
+del_perc = 0
+timeout = 1.0
+dist = "zipfian"
+zipf_hotspot = 0.5
+
+# read-intensive, zipfian
+[[benchmark]]
+set_perc = 5
+get_perc = 95
+del_perc = 0
+timeout = 1.0
+dist = "zipfian"
+
+# read-only, uniform
+[[benchmark]]
+set_perc = 0
+get_perc = 100
+del_perc = 0
+timeout = 1.0
+dist = "uniform"
+```
+
+In the first phase, all worker threads fill the key space of the store, and the metrics are hidden.
+Then, the benchmark consists of 4 parts with 5 seconds running time each:
+
+- Write-intensive on popular keys.
+- Write-intensive on popular keys while the popular keys shifted to the middle of the
+key space.
+- Read-intensive (only 5% writes) on popular keys.
+- Read-only on uniformly random keys.
+
+The script file `run.sh` runs this benchmark against multiple stores with 32 threads.
+Although the number of threads set in the configuration file is only 1, the number of threads are
+dynamically adjusted by setting `global.threads` to 32.
+
+Results:
+
+[mixed](mixed.pdf)

examples/readpopular/README.md

@@ -0,0 +1,42 @@
+This example shows a benchmark that reads popular records in the store, running with different
+number of threads.
+
+The workload file, `readpopular.toml` is as follows:
+
+```toml
+[global]
+threads = 1
+repeat = 1
+klen = 8
+vlen = 16
+kmin = 0
+kmax = 1000000
+
+[[benchmark]]
+set_perc = 100
+get_perc = 0
+del_perc = 0
+repeat = 1
+dist = "incrementp"
+report = "hidden"
+
+[[benchmark]]
+timeout = 1
+set_perc = 0
+get_perc = 100
+del_perc = 0
+dist = "zipfian"
+zipf_theta = 1.0
+report = "finish"
+```
+
+In the first phase, all worker threads fill the key space of the store, and the metrics are hidden.
+In the second phase, worker threads execute the read-only workload that accesses Zipfian keys for
+1 second and report once when finished.
+
+The script file `run.sh` runs this benchmark against multiple stores with different number of
+threads. The number of threads are dynamically adjusted via `global.threads` environment variable.
+
+Results:
+
+[readpopular](readpopular.pdf)

examples/writeheavy/README.md

@@ -0,0 +1,41 @@
+This example shows a benchmark that mixes reads and writes at 1:1 ratio accessing a random record
+in the store, running with different number of threads.
+
+The workload file, `writeheavy.toml` is as follows:
+
+```toml
+[global]
+threads = 1
+repeat = 1
+klen = 8
+vlen = 16
+kmin = 0
+kmax = 1000000
+
+[[benchmark]]
+set_perc = 100
+get_perc = 0
+del_perc = 0
+repeat = 1
+dist = "incrementp"
+report = "hidden"
+
+[[benchmark]]
+timeout = 1
+set_perc = 50
+get_perc = 50
+del_perc = 0
+dist = "uniform"
+report = "finish"
+```
+
+In the first phase, all worker threads fill the key space of the store, and the metrics are hidden.
+In the second phase, worker threads execute the write-heavy workload for 1 second and report once
+when finished.
+
+The script file `run.sh` runs this benchmark against multiple stores with different number of
+threads. The number of threads are dynamically adjusted via `global.threads` environment variable.
+
+Results:
+
+[writeheavy](writeheavy.pdf)

examples/your-kv-store/README.md

@@ -1,3 +1,5 @@
+This example shows how to integrate `kvbench` into your own key-value store implementations.
+
 To compile, simply use:
 
 ```

examples/your-kv-store/src/main.rs

@@ -5,7 +5,7 @@ extern crate kvbench;
 use kvbench::inventory;
 use kvbench::toml;
 
-use kvbench::bench::{BenchKVMap, Registry};
+use kvbench::stores::{BenchKVMap, Registry};
 use kvbench::{KVMap, KVMapHandle};
 use std::collections::HashMap;
 use std::sync::{Arc, RwLock};