[redis.git] / doc / Benchmarks.html


<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN">
<html>
    <head>
        <link type="text/css" rel="stylesheet" href="style.css" />
    </head>
    <body>
        <div id="page">
        
            <div id='header'>
            <a href="index.html">
            <img style="border:none" alt="Redis Documentation" src="redis.png">
            </a>
            </div>
        
            <div id="pagecontent">
                <div class="index">
<!-- This is a (PRE) block.  Make sure it's left aligned or your toc title will be off. -->
<b>Benchmarks: Contents</b><br>&nbsp;&nbsp;<a href="#How Fast is Redis?">How Fast is Redis?</a><br>&nbsp;&nbsp;<a href="#Latency percentiles">Latency percentiles</a>
                </div>
                
                <h1 class="wikiname">Benchmarks</h1>

                <div class="summary">
                    
                </div>

                <div class="narrow">
                    <h1><a name="How Fast is Redis?">How Fast is Redis?</a></h1>Redis includes the <code name="code" class="python">redis-benchmark</code> utility that simulates SETs/GETs done by N clients at the same time sending M total queries (it is similar to the Apache's <code name="code" class="python">ab</code> utility). Below you'll find the full output of the benchmark executed against a Linux box.<br/><br/><ul><li> The test was done with 50 simultaneous clients performing 100000 requests.</li><li> The value SET and GET is a 256 bytes string.</li><li> The Linux box is running <b>Linux 2.6</b>, it's <b>Xeon X3320 2.5Ghz</b>.</li><li> Text executed using the loopback interface (127.0.0.1).</li></ul>
Results: <b>about 110000 SETs per second, about 81000 GETs per second.</b><h1><a name="Latency percentiles">Latency percentiles</a></h1><pre class="codeblock python" name="code">
./redis-benchmark -n 100000

====== SET ======
  100007 requests completed in 0.88 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

58.50% &lt;= 0 milliseconds
99.17% &lt;= 1 milliseconds
99.58% &lt;= 2 milliseconds
99.85% &lt;= 3 milliseconds
99.90% &lt;= 6 milliseconds
100.00% &lt;= 9 milliseconds
114293.71 requests per second

====== GET ======
  100000 requests completed in 1.23 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

43.12% &lt;= 0 milliseconds
96.82% &lt;= 1 milliseconds
98.62% &lt;= 2 milliseconds
100.00% &lt;= 3 milliseconds
81234.77 requests per second

====== INCR ======
  100018 requests completed in 1.46 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

32.32% &lt;= 0 milliseconds
96.67% &lt;= 1 milliseconds
99.14% &lt;= 2 milliseconds
99.83% &lt;= 3 milliseconds
99.88% &lt;= 4 milliseconds
99.89% &lt;= 5 milliseconds
99.96% &lt;= 9 milliseconds
100.00% &lt;= 18 milliseconds
68458.59 requests per second

====== LPUSH ======
  100004 requests completed in 1.14 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

62.27% &lt;= 0 milliseconds
99.74% &lt;= 1 milliseconds
99.85% &lt;= 2 milliseconds
99.86% &lt;= 3 milliseconds
99.89% &lt;= 5 milliseconds
99.93% &lt;= 7 milliseconds
99.96% &lt;= 9 milliseconds
100.00% &lt;= 22 milliseconds
100.00% &lt;= 208 milliseconds
88109.25 requests per second

====== LPOP ======
  100001 requests completed in 1.39 seconds
  50 parallel clients
  3 bytes payload
  keep alive: 1

54.83% &lt;= 0 milliseconds
97.34% &lt;= 1 milliseconds
99.95% &lt;= 2 milliseconds
99.96% &lt;= 3 milliseconds
99.96% &lt;= 4 milliseconds
100.00% &lt;= 9 milliseconds
100.00% &lt;= 208 milliseconds
71994.96 requests per second
</pre>Notes: changing the payload from 256 to 1024 or 4096 bytes does not change the numbers significantly (but reply packets are glued together up to 1024 bytes so GETs may be slower with big payloads). The same for the number of clients, from 50 to 256 clients I got the same numbers. With only 10 clients it starts to get a bit slower.<br/><br/>You can expect different results from different boxes. For example a low profile box like <b>Intel core duo T5500 clocked at 1.66Ghz running Linux 2.6</b> will output the following:
<pre class="codeblock python python" name="code">
 ./redis-benchmark -q -n 100000
SET: 53684.38 requests per second
GET: 45497.73 requests per second
INCR: 39370.47 requests per second
LPUSH: 34803.41 requests per second
LPOP: 37367.20 requests per second
</pre>
                </div>
        
            </div>
        </div>
    </body>
</html>
Commit	Line	Data
ed9b544e	1
	2	<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN">
	3	<html>
	4	<head>
	5	<link type="text/css" rel="stylesheet" href="style.css" />
	6	</head>
	7	<body>
	8	<div id="page">
	9
	10	<div id='header'>
	11	<a href="index.html">
	12	<img style="border:none" alt="Redis Documentation" src="redis.png">
	13	</a>
	14	</div>
	15
	16	<div id="pagecontent">
	17	<div class="index">
	18	<!-- This is a (PRE) block. Make sure it's left aligned or your toc title will be off. -->
	19	<b>Benchmarks: Contents</b><br>  <a href="#How Fast is Redis?">How Fast is Redis?</a><br>  <a href="#Latency percentiles">Latency percentiles</a>
	20	</div>
	21
	22	<h1 class="wikiname">Benchmarks</h1>
	23
	24	<div class="summary">
	25
	26	</div>
	27
	28	<div class="narrow">
	29	<h1><a name="How Fast is Redis?">How Fast is Redis?</a></h1>Redis includes the <code name="code" class="python">redis-benchmark</code> utility that simulates SETs/GETs done by N clients at the same time sending M total queries (it is similar to the Apache's <code name="code" class="python">ab</code> utility). Below you'll find the full output of the benchmark executed against a Linux box.<br/><br/><ul><li> The test was done with 50 simultaneous clients performing 100000 requests.</li><li> The value SET and GET is a 256 bytes string.</li><li> The Linux box is running <b>Linux 2.6</b>, it's <b>Xeon X3320 2.5Ghz</b>.</li><li> Text executed using the loopback interface (127.0.0.1).</li></ul>
	30	Results: <b>about 110000 SETs per second, about 81000 GETs per second.</b><h1><a name="Latency percentiles">Latency percentiles</a></h1><pre class="codeblock python" name="code">
	31	./redis-benchmark -n 100000
	32
	33	====== SET ======
	34	100007 requests completed in 0.88 seconds
	35	50 parallel clients
	36	3 bytes payload
	37	keep alive: 1
	38
	39	58.50% <= 0 milliseconds
	40	99.17% <= 1 milliseconds
	41	99.58% <= 2 milliseconds
	42	99.85% <= 3 milliseconds
	43	99.90% <= 6 milliseconds
	44	100.00% <= 9 milliseconds
	45	114293.71 requests per second
	46
	47	====== GET ======
	48	100000 requests completed in 1.23 seconds
	49	50 parallel clients
	50	3 bytes payload
	51	keep alive: 1
	52
	53	43.12% <= 0 milliseconds
	54	96.82% <= 1 milliseconds
	55	98.62% <= 2 milliseconds
	56	100.00% <= 3 milliseconds
	57	81234.77 requests per second
	58
	59	====== INCR ======
	60	100018 requests completed in 1.46 seconds
	61	50 parallel clients
	62	3 bytes payload
	63	keep alive: 1
	64
65	32.32% <= 0 milliseconds
66	96.67% <= 1 milliseconds
67	99.14% <= 2 milliseconds
68	99.83% <= 3 milliseconds
69	99.88% <= 4 milliseconds
70	99.89% <= 5 milliseconds
71	99.96% <= 9 milliseconds
72	100.00% <= 18 milliseconds
73	68458.59 requests per second
74
75	====== LPUSH ======
76	100004 requests completed in 1.14 seconds
77	50 parallel clients
78	3 bytes payload
79	keep alive: 1
80
81	62.27% <= 0 milliseconds
82	99.74% <= 1 milliseconds
83	99.85% <= 2 milliseconds
84	99.86% <= 3 milliseconds
85	99.89% <= 5 milliseconds
86	99.93% <= 7 milliseconds
87	99.96% <= 9 milliseconds
88	100.00% <= 22 milliseconds
89	100.00% <= 208 milliseconds
90	88109.25 requests per second
91
92	====== LPOP ======
93	100001 requests completed in 1.39 seconds
94	50 parallel clients
95	3 bytes payload
96	keep alive: 1
97
98	54.83% <= 0 milliseconds
99	97.34% <= 1 milliseconds
100	99.95% <= 2 milliseconds
101	99.96% <= 3 milliseconds
102	99.96% <= 4 milliseconds
103	100.00% <= 9 milliseconds
104	100.00% <= 208 milliseconds
105	71994.96 requests per second
106	</pre>Notes: changing the payload from 256 to 1024 or 4096 bytes does not change the numbers significantly (but reply packets are glued together up to 1024 bytes so GETs may be slower with big payloads). The same for the number of clients, from 50 to 256 clients I got the same numbers. With only 10 clients it starts to get a bit slower.<br/><br/>You can expect different results from different boxes. For example a low profile box like <b>Intel core duo T5500 clocked at 1.66Ghz running Linux 2.6</b> will output the following:
107	<pre class="codeblock python python" name="code">
108	./redis-benchmark -q -n 100000
109	SET: 53684.38 requests per second
110	GET: 45497.73 requests per second
111	INCR: 39370.47 requests per second
112	LPUSH: 34803.41 requests per second
113	LPOP: 37367.20 requests per second
114	</pre>
115	</div>
116
117	</div>
118	</div>
119	</body>
120	</html>
121