Linux – How to find the L2 cache size in Linux

Use the lscpu command.

32bit example output:

$ lscpu
Architecture:          i686
CPU op-mode(s):        32-bit
Byte Order:            Little Endian
CPU(s):                4
On-line CPU(s) list:   0-3
Thread(s) per core:    4
Core(s) per socket:    1
Socket(s):             1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 26
Stepping:              5
CPU MHz:               2260.998
BogoMIPS:              4521.99
L1d cache:             32K
L1i cache:             32K
L2 cache:              256K
L3 cache:              8192K

64bit example:

$ lscpu
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
CPU(s):                4
Thread(s) per core:    1
Core(s) per socket:    4
CPU socket(s):         1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 23
Stepping:              6
CPU MHz:               2327.533
Virtualization:        VT-x
L1d cache:             32K
L1i cache:             32K
L2 cache:              6144K

From http://www.kernel.org/doc/Documentation/filesystems/tmpfs.txt:

Further on you can check the actual RAM+swap use of a tmpfs instance with df(1) and du(1).

So simply use df or du (you can also use them for ramfs):

$ df | grep tmpfs
tmpfs             205032      1136    203896   1% /run

so 1136 KB is in use.

$ sudo du -s /run
[sudo] password for jaume: 
1416    /run

so 1416 KB is in use.

Thats's interesting... df and du report different sizes. What is happening here and which one is right?

/run has a subfolder called /run/shm, which is itself a separate tmpfs filesytem (although shown as none):

$ df | grep run
tmpfs             205032      1136    203896   1% /run
none                5120         0      5120   0% /run/lock
none              512572       280    512292   1% /run/shm

Adding both amounts you get the size reported by du:

$ expr 1136 + 280
1416

So the whole story is that 2 tmpfs filesystems use 1416 KB.

(Here's another reason why du and df outputs may differ.)