0

Lately I've launched a few EC2 c1.xlarge instances with identical setups to ones that I launch a few months ago and I've found that the new ones have 5-10 times the CPU usage. The only difference I can see is that new instances use a 6M cache. Has anyone else experienced this oddness? I've also noticed the new, slower ones have the sse4_1 flag, whereas the old, faster ones don't. Super confusing!

Here is the /proc/cpuinfo for the new, slower one:

processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 23
model name      : Intel(R) Xeon(R) CPU           E5410  @ 2.33GHz
stepping        : 10
cpu MHz         : 2327.506
cache size      : 6144 KB
physical id     : 0
siblings        : 1
core id         : 0
cpu cores       : 1
apicid          : 0
initial apicid  : 1
fpu             : yes
fpu_exception   : yes
cpuid level     : 13
wp              : yes
flags           : fpu de tsc msr pae cx8 apic sep cmov pat clflush acpi mmx fxsr sse sse2 ss ht syscall nx lm constant_tsc rep_good aperfmperf pni ssse3 cx16 sse4_1 hypervisor lahf_lm
bogomips        : 4657.88
clflush size    : 64
cache_alignment : 64
address sizes   : 38 bits physical, 48 bits virtual

And here it is for the old, faster one:

processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 15
model name      : Intel(R) Xeon(R) CPU           E5345  @ 2.33GHz
stepping        : 11
cpu MHz         : 2327.502
cache size      : 4096 KB
physical id     : 0
siblings        : 1
core id         : 0
cpu cores       : 1
apicid          : 0
initial apicid  : 0
fpu             : yes
fpu_exception   : yes
cpuid level     : 10
wp              : yes
flags           : fpu de tsc msr pae cx8 apic sep cmov pat clflush acpi mmx fxsr sse sse2 ss ht syscall nx lm constant_tsc rep_good aperfmperf pni ssse3 cx16 hypervisor lahf_lm
bogomips        : 4662.14
clflush size    : 64
cache_alignment : 64
address sizes   : 38 bits physical, 48 bits virtual

Anyone have any idea what could be going on?

2 Answers 2

3

You have probably observed a known issue with EC2: even within the same EC2 instance type, heterogenous hardware might be used, resulting in potentially high performance variations.

Refer to this HotCloud'12 paper for more information.

0

On the slower instance, you may have a busy neighbor. If it is an EBS backed instance stop it and start it again so it switches to a new host and see if that clears up the problem. If it isn't then you'd have to terminate it and build a new one.

1
  • I've launched 5 new instances and they all seem to experience this problem. Has my EC2 neighborhood just gotten worse? Oct 24, 2012 at 16:32

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .