|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [Xen-users] Performance of PVM and HVM
Recently, I met a strange issue about vm's performance.
I downloaded Jeremy's branch of xen/next-2.6.32 (version:2.6.32.57)for dom0,
and compiled xen-4.1.3 source code for xend service, then I used Unixbench
to test the performance of native, dom0, and domU(HVM and PVM). By the way,
I set 4 vcpus and 2048M memory for each VM. And I got the following results:
It seems that dom0's performance is so bad from native, and PVM so on. But
HVM's performance is better than my expected, it's so close to native. It's
weird to me, or it's just a normal thing? Is any one else experiencing this?
Any reply or suggests will be appreciated.
========native:(2.6.32.57 without xen)============
System: localhost.localdomain: GNU/Linux
OS: GNU/Linux -- 2.6.32.57 -- #1 SMP Thu Sep 6 10:42:38 CST 2012
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6185.3 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext,
SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 1: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6185.9 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext,
SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 2: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6185.9 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext,
SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 3: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6185.9 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext,
SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
11:01:15 up 9 min, 2 users, load average: 0.15, 0.41, 0.26; runlevel 5
Double-Precision Whetstone 4023.5 MWIPS (9.9 s, 7
samples)
Execl Throughput 2384.2 lps (29.7 s, 2
samples)
File Copy 1024 bufsize 2000 maxblocks 1134757.4 KBps (30.0 s, 2
samples)
File Copy 256 bufsize 500 maxblocks 312913.0 KBps (30.0 s, 2
samples)
File Copy 4096 bufsize 8000 maxblocks 2731953.1 KBps (30.0 s, 2
samples)
Pipe Throughput 2043709.3 lps (10.0 s, 7
samples)
Pipe-based Context Switching 105547.0 lps (10.0 s, 7
samples)
Process Creation 10013.4 lps (30.0 s, 2
samples)
Shell Scripts (1 concurrent) 4158.8 lpm (60.0 s, 2
samples)
Shell Scripts (8 concurrent) 2961.2 lpm (60.0 s, 2
samples)
System Call Overhead 4431920.0 lps (10.0 s, 7
samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 37435723.9 3207.9
Double-Precision Whetstone 55.0 4023.5 731.5
Execl Throughput 43.0 2384.2 554.5
File Copy 1024 bufsize 2000 maxblocks 3960.0 1134757.4 2865.5
File Copy 256 bufsize 500 maxblocks 1655.0 312913.0 1890.7
File Copy 4096 bufsize 8000 maxblocks 5800.0 2731953.1 4710.3
Pipe Throughput 12440.0 2043709.3 1642.9
Pipe-based Context Switching 4000.0 105547.0 263.9
Process Creation 126.0 10013.4 794.7
Shell Scripts (1 concurrent) 42.4 4158.8 980.9
Shell Scripts (8 concurrent) 6.0 2961.2 4935.4
System Call Overhead 15000.0 4431920.0 2954.6
========
System Benchmarks Index Score 1529.3
------------------------------------------------------------------------
Benchmark Run: Thu Sep 06 2012 11:30:07 - 11:59:09
4 CPUs in system; running 4 parallel copies of tests
Dhrystone 2 using register variables 141361626.6 lps (10.0 s, 7
samples)
Double-Precision Whetstone 15160.0 MWIPS (9.9 s, 7
samples)
Execl Throughput 20000.9 lps (29.6 s, 2
samples)
File Copy 1024 bufsize 2000 maxblocks 1127449.6 KBps (30.0 s, 2
samples)
File Copy 256 bufsize 500 maxblocks 299709.9 KBps (30.0 s, 2
samples)
File Copy 4096 bufsize 8000 maxblocks 3197973.3 KBps (30.0 s, 2
samples)
Pipe Throughput 7732066.4 lps (10.0 s, 7
samples)
Pipe-based Context Switching 1801378.7 lps (10.0 s, 7
samples)
Process Creation 75567.5 lps (30.0 s, 2
samples)
Shell Scripts (1 concurrent) 25427.3 lpm (60.0 s, 2
samples)
Shell Scripts (8 concurrent) 3397.8 lpm (60.0 s, 2
samples)
System Call Overhead 12278676.0 lps (10.0 s, 7
samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 141361626.6 12113.2
Double-Precision Whetstone 55.0 15160.0 2756.4
Execl Throughput 43.0 20000.9 4651.4
File Copy 1024 bufsize 2000 maxblocks 3960.0 1127449.6 2847.1
File Copy 256 bufsize 500 maxblocks 1655.0 299709.9 1810.9
File Copy 4096 bufsize 8000 maxblocks 5800.0 3197973.3 5513.7
Pipe Throughput 12440.0 7732066.4 6215.5
Pipe-based Context Switching 4000.0 1801378.7 4503.4
Process Creation 126.0 75567.5 5997.4
Shell Scripts (1 concurrent) 42.4 25427.3 5997.0
Shell Scripts (8 concurrent) 6.0 3397.8 5663.1
System Call Overhead 15000.0 12278676.0 8185.8
========
System Benchmarks Index Score 4929.8
========native:(2.6.32.57 with xen)============
System: localhost.localdomain: GNU/Linux
OS: GNU/Linux -- 2.6.32.57 -- #1 SMP Wed Sep 5 16:17:40 CST 2012
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6185.9 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext,
SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6185.9 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext,
SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6185.9 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext,
SYSENTER/SYSEXIT, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6185.9 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext,
SYSENTER/SYSEXIT, SYSCALL/SYSRET
16:39:07 up 5 min, 2 users, load average: 0.41, 0.14, 0.05; runlevel 5
Double-Precision Whetstone 4014.3 MWIPS (10.0 s, 7
samples)
Execl Throughput 1539.9 lps (29.7 s, 2
samples)
File Copy 1024 bufsize 2000 maxblocks 230816.4 KBps (30.0 s, 2
samples)
File Copy 256 bufsize 500 maxblocks 59384.4 KBps (30.0 s, 2
samples)
File Copy 4096 bufsize 8000 maxblocks 799672.5 KBps (30.0 s, 2
samples)
Pipe Throughput 464331.1 lps (10.0 s, 7
samples)
Pipe-based Context Switching 84464.6 lps (10.0 s, 7
samples)
Process Creation 1416.4 lps (30.0 s, 2
samples)
Shell Scripts (1 concurrent) 1777.9 lpm (60.0 s, 2
samples)
Shell Scripts (8 concurrent) 1077.6 lpm (60.1 s, 2
samples)
System Call Overhead 588951.1 lps (10.0 s, 7
samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 37551405.9 3217.8
Double-Precision Whetstone 55.0 4014.3 729.9
Execl Throughput 43.0 1539.9 358.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 230816.4 582.9
File Copy 256 bufsize 500 maxblocks 1655.0 59384.4 358.8
File Copy 4096 bufsize 8000 maxblocks 5800.0 799672.5 1378.7
Pipe Throughput 12440.0 464331.1 373.3
Pipe-based Context Switching 4000.0 84464.6 211.2
Process Creation 126.0 1416.4 112.4
Shell Scripts (1 concurrent) 42.4 1777.9 419.3
Shell Scripts (8 concurrent) 6.0 1077.6 1796.0
System Call Overhead 15000.0 588951.1 392.6
========
System Benchmarks Index Score 541.5
------------------------------------------------------------------------
Benchmark Run: Wed Sep 05 2012 17:08:17 - 17:37:42
4 CPUs in system; running 4 parallel copies of tests
Dhrystone 2 using register variables 141651867.8 lps (10.0 s, 7
samples)
Double-Precision Whetstone 15164.7 MWIPS (9.9 s, 7
samples)
Execl Throughput 4973.0 lps (29.5 s, 2
samples)
File Copy 1024 bufsize 2000 maxblocks 314355.0 KBps (30.0 s, 2
samples)
File Copy 256 bufsize 500 maxblocks 80269.7 KBps (30.0 s, 2
samples)
File Copy 4096 bufsize 8000 maxblocks 1157706.6 KBps (30.0 s, 2
samples)
Pipe Throughput 1733674.4 lps (10.0 s, 7
samples)
Pipe-based Context Switching 336552.6 lps (10.0 s, 7
samples)
Process Creation 10207.0 lps (30.0 s, 2
samples)
Shell Scripts (1 concurrent) 8880.3 lpm (60.0 s, 2
samples)
Shell Scripts (8 concurrent) 1197.1 lpm (60.1 s, 2
samples)
System Call Overhead 2126022.5 lps (10.0 s, 7
samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 141651867.8 12138.1
Double-Precision Whetstone 55.0 15164.7 2757.2
Execl Throughput 43.0 4973.0 1156.5
File Copy 1024 bufsize 2000 maxblocks 3960.0 314355.0 793.8
File Copy 256 bufsize 500 maxblocks 1655.0 80269.7 485.0
File Copy 4096 bufsize 8000 maxblocks 5800.0 1157706.6 1996.0
Pipe Throughput 12440.0 1733674.4 1393.6
Pipe-based Context Switching 4000.0 336552.6 841.4
Process Creation 126.0 10207.0 810.1
Shell Scripts (1 concurrent) 42.4 8880.3 2094.4
Shell Scripts (8 concurrent) 6.0 1197.1 1995.2
System Call Overhead 15000.0 2126022.5 1417.3
========
System Benchmarks Index Score 1532.1
========native:(PVM rhel6.1)============
System: xenpv1.myxen.com: GNU/Linux
OS: GNU/Linux -- 2.6.32-131.0.15.el6.x86_64 -- #1 SMP Tue May 10 15:42:40
EDT 2011
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6186.4 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6186.4 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6186.4 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6186.4 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
09:21:10 up 1 min, 1 user, load average: 0.09, 0.03, 0.00; runlevel 3
------------------------------------------------------------------------
Benchmark Run: Fri Sep 14 2012 09:21:10 - 09:49:19
4 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 36560766.3 lps (10.0 s, 7
samples)
Double-Precision Whetstone 3962.1 MWIPS (9.9 s, 7
samples)
Execl Throughput 798.0 lps (30.0 s, 2
samples)
File Copy 1024 bufsize 2000 maxblocks 307212.1 KBps (30.0 s, 2
samples)
File Copy 256 bufsize 500 maxblocks 77826.9 KBps (30.0 s, 2
samples)
File Copy 4096 bufsize 8000 maxblocks 989297.2 KBps (30.0 s, 2
samples)
Pipe Throughput 415705.7 lps (10.0 s, 7
samples)
Pipe-based Context Switching 82646.7 lps (10.0 s, 7
samples)
Process Creation 1643.2 lps (30.0 s, 2
samples)
Shell Scripts (1 concurrent) 2111.5 lpm (60.0 s, 2
samples)
Shell Scripts (8 concurrent) 1222.8 lpm (60.1 s, 2
samples)
System Call Overhead 442797.7 lps (10.0 s, 7
samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 36560766.3 3132.9
Double-Precision Whetstone 55.0 3962.1 720.4
Execl Throughput 43.0 798.0 185.6
File Copy 1024 bufsize 2000 maxblocks 3960.0 307212.1 775.8
File Copy 256 bufsize 500 maxblocks 1655.0 77826.9 470.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 989297.2 1705.7
Pipe Throughput 12440.0 415705.7 334.2
Pipe-based Context Switching 4000.0 82646.7 206.6
Process Creation 126.0 1643.2 130.4
Shell Scripts (1 concurrent) 42.4 2111.5 498.0
Shell Scripts (8 concurrent) 6.0 1222.8 2038.0
System Call Overhead 15000.0 442797.7 295.2
========
System Benchmarks Index Score 546.1
------------------------------------------------------------------------
Benchmark Run: Fri Sep 14 2012 09:49:19 - 10:17:31
4 CPUs in system; running 4 parallel copies of tests
Dhrystone 2 using register variables 139304316.2 lps (10.0 s, 7
samples)
Double-Precision Whetstone 14960.7 MWIPS (9.9 s, 7
samples)
Execl Throughput 5607.6 lps (30.0 s, 2
samples)
File Copy 1024 bufsize 2000 maxblocks 498246.2 KBps (30.0 s, 2
samples)
File Copy 256 bufsize 500 maxblocks 121580.0 KBps (30.0 s, 2
samples)
File Copy 4096 bufsize 8000 maxblocks 1649657.6 KBps (30.1 s, 2
samples)
Pipe Throughput 1574427.7 lps (10.0 s, 7
samples)
Pipe-based Context Switching 327122.9 lps (10.0 s, 7
samples)
Process Creation 10641.0 lps (30.0 s, 2
samples)
Shell Scripts (1 concurrent) 9917.6 lpm (60.0 s, 2
samples)
Shell Scripts (8 concurrent) 1351.8 lpm (60.1 s, 2
samples)
System Call Overhead 1605775.2 lps (10.0 s, 7
samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 139304316.2 11937.0
Double-Precision Whetstone 55.0 14960.7 2720.1
Execl Throughput 43.0 5607.6 1304.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 498246.2 1258.2
File Copy 256 bufsize 500 maxblocks 1655.0 121580.0 734.6
File Copy 4096 bufsize 8000 maxblocks 5800.0 1649657.6 2844.2
Pipe Throughput 12440.0 1574427.7 1265.6
Pipe-based Context Switching 4000.0 327122.9 817.8
Process Creation 126.0 10641.0 844.5
Shell Scripts (1 concurrent) 42.4 9917.6 2339.0
Shell Scripts (8 concurrent) 6.0 1351.8 2253.0
System Call Overhead 15000.0 1605775.2 1070.5
========
System Benchmarks Index Score 1691.6
========native:(HVM rhel6.1)============
System: hvmserver0.myxen.com: GNU/Linux
OS: GNU/Linux -- 2.6.32-131.0.15.el6.x86_64 -- #1 SMP Tue May 10 15:42:40
EDT 2011
Machine: x86_64 (x86_64)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6186.4 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
CPU 1: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6185.2 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
CPU 2: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6185.4 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
CPU 3: Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz (6185.4 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSCALL/SYSRET
10:21:14 up 2 min, 1 user, load average: 0.13, 0.12, 0.05; runlevel 3
------------------------------------------------------------------------
Benchmark Run: Fri Sep 14 2012 10:21:14 - 10:49:17
4 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 35232485.6 lps (10.0 s, 7
samples)
Double-Precision Whetstone 3877.4 MWIPS (9.7 s, 7
samples)
Execl Throughput 1939.8 lps (30.0 s, 2
samples)
File Copy 1024 bufsize 2000 maxblocks 966506.7 KBps (30.0 s, 2
samples)
File Copy 256 bufsize 500 maxblocks 266621.2 KBps (30.0 s, 2
samples)
File Copy 4096 bufsize 8000 maxblocks 2444019.5 KBps (30.0 s, 2
samples)
Pipe Throughput 1590900.5 lps (10.0 s, 7
samples)
Pipe-based Context Switching 368541.3 lps (10.0 s, 7
samples)
Process Creation 5577.5 lps (30.0 s, 2
samples)
Shell Scripts (1 concurrent) 3881.2 lpm (60.0 s, 2
samples)
Shell Scripts (8 concurrent) 2650.3 lpm (60.0 s, 2
samples)
System Call Overhead 2509799.8 lps (10.0 s, 7
samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 35232485.6 3019.1
Double-Precision Whetstone 55.0 3877.4 705.0
Execl Throughput 43.0 1939.8 451.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 966506.7 2440.7
File Copy 256 bufsize 500 maxblocks 1655.0 266621.2 1611.0
File Copy 4096 bufsize 8000 maxblocks 5800.0 2444019.5 4213.8
Pipe Throughput 12440.0 1590900.5 1278.9
Pipe-based Context Switching 4000.0 368541.3 921.4
Process Creation 126.0 5577.5 442.7
Shell Scripts (1 concurrent) 42.4 3881.2 915.4
Shell Scripts (8 concurrent) 6.0 2650.3 4417.2
System Call Overhead 15000.0 2509799.8 1673.2
========
System Benchmarks Index Score 1398.8
------------------------------------------------------------------------
Benchmark Run: Fri Sep 14 2012 10:49:17 - 11:17:26
4 CPUs in system; running 4 parallel copies of tests
Dhrystone 2 using register variables 133445381.8 lps (10.0 s, 7
samples)
Double-Precision Whetstone 14620.1 MWIPS (9.9 s, 7
samples)
Execl Throughput 17582.8 lps (30.0 s, 2
samples)
File Copy 1024 bufsize 2000 maxblocks 1126655.3 KBps (30.0 s, 2
samples)
File Copy 256 bufsize 500 maxblocks 301996.7 KBps (30.0 s, 2
samples)
File Copy 4096 bufsize 8000 maxblocks 3221205.1 KBps (30.0 s, 2
samples)
Pipe Throughput 6244159.7 lps (10.0 s, 7
samples)
Pipe-based Context Switching 1508547.2 lps (10.0 s, 7
samples)
Process Creation 57311.4 lps (30.0 s, 2
samples)
Shell Scripts (1 concurrent) 21936.7 lpm (60.0 s, 2
samples)
Shell Scripts (8 concurrent) 2931.2 lpm (60.0 s, 2
samples)
System Call Overhead 7935430.4 lps (10.0 s, 7
samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 133445381.8 11434.9
Double-Precision Whetstone 55.0 14620.1 2658.2
Execl Throughput 43.0 17582.8 4089.0
File Copy 1024 bufsize 2000 maxblocks 3960.0 1126655.3 2845.1
File Copy 256 bufsize 500 maxblocks 1655.0 301996.7 1824.8
File Copy 4096 bufsize 8000 maxblocks 5800.0 3221205.1 5553.8
Pipe Throughput 12440.0 6244159.7 5019.4
Pipe-based Context Switching 4000.0 1508547.2 3771.4
Process Creation 126.0 57311.4 4548.5
Shell Scripts (1 concurrent) 42.4 21936.7 5173.8
Shell Scripts (8 concurrent) 6.0 2931.2 4885.4
System Call Overhead 15000.0 7935430.4 5290.3
========
System Benchmarks Index Score 4311.5
--
View this message in context:
http://xen.1045712.n5.nabble.com/Performance-of-PVM-and-HVM-tp5711309.html
Sent from the Xen - User mailing list archive at Nabble.com.
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxx
http://lists.xen.org/xen-users
|
 |
Lists.xenproject.org is hosted with RackSpace, monitoring our |