Esiste un modo per conoscere le dimensioni delle cache L1, L2, L3 e RAM in Linux?
Esiste un modo per conoscere le dimensioni delle cache L1, L2, L3 e RAM in Linux?
Risposte:
Se hai lshw
installato:
$ sudo lshw -C memory
$ sudo lshw -C memory
...
*-cache:0
description: L1 cache
physical id: a
slot: Internal L1 Cache
size: 32KiB
capacity: 32KiB
capabilities: asynchronous internal write-through data
*-cache:1
description: L2 cache
physical id: b
slot: Internal L2 Cache
size: 256KiB
capacity: 256KiB
capabilities: burst internal write-through unified
*-cache:2
description: L3 cache
physical id: c
slot: Internal L3 Cache
size: 3MiB
capacity: 8MiB
capabilities: burst internal write-back
*-memory
description: System Memory
physical id: 2a
slot: System board or motherboard
size: 8GiB
*-bank:0
description: SODIMM DDR3 Synchronous 1334 MHz (0.7 ns)
product: M471B5273CH0-CH9
vendor: Samsung
physical id: 0
serial: 67010644
slot: DIMM 1
size: 4GiB
width: 64 bits
clock: 1334MHz (0.7ns)
*-bank:1
description: SODIMM DDR3 Synchronous 1334 MHz (0.7 ns)
product: 16JTF51264HZ-1G4H1
vendor: Micron Technology
physical id: 1
serial: 3749C127
slot: DIMM 2
size: 4GiB
width: 64 bits
clock: 1334MHz (0.7ns)
lshw
(eseguito con il permesso di root ovviamente) non mi ha dato informazioni sulla cache. ma lscpu
e gli dmidecode
strumenti mi hanno dato i risultati.
Se ti preoccupi solo delle dimensioni, prova lscpu
da util-linux
.
$ lscpu
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
CPU(s): 4
On-line CPU(s) list: 0-3
Thread(s) per core: 2
Core(s) per socket: 2
Socket(s): 1
NUMA node(s): 1
Vendor ID: GenuineIntel
CPU family: 6
Model: 37
Model name: Intel(R) Core(TM) i5 CPU M 560 @ 2.67GHz
Stepping: 5
CPU MHz: 1199.000
BogoMIPS: 5319.88
Virtualization: VT-x
L1d cache: 32K
L1i cache: 32K
L2 cache: 256K
L3 cache: 3072K
NUMA node0 CPU(s): 0-3
Dovrebbe esserci anche un pacchetto / comando chiamato x86info. Supponendo che abbiate i386 / x86_64, x86info -c
dovreste fornire informazioni più dettagliate sulle cache.
$ x86info -c
x86info v1.30. Dave Jones 2001-2011
Feedback to <davej@redhat.com>.
Found 4 identical CPUs
Extended Family: 0 Extended Model: 2 Family: 6 Model: 37 Stepping: 5
Type: 0 (Original OEM)
CPU Model (x86info's best guess): Core i7 (Nehalem) [Clarkdale/Arrandale]
Processor name string (BIOS programmed): Intel(R) Core(TM) i5 CPU M 560 @ 2.67GHz
Cache info
L1 Instruction cache: 32KB, 4-way associative. 64 byte line size.
L1 Data cache: 32KB, 8-way associative. 64 byte line size.
L2 (MLC): 256KB, 8-way associative. 64 byte line size.
TLB info
Instruction TLB: 2MB or 4MB pages, fully associative, 7 entries
Instruction TLB: 4K pages, 4-way associative, 64 entries.
Data TLB: 4KB or 4MB pages, fully associative, 32 entries.
Data TLB: 4KB pages, 4-way associative, 64 entries
Data TLB: 4K pages, 4-way associative, 512 entries.
Data TLB: 4KB or 4MB pages, fully associative, 32 entries.
Data TLB: 4KB pages, 4-way associative, 64 entries
64 byte prefetching.
Data TLB: 4K pages, 4-way associative, 512 entries.
Found unknown cache descriptors: dd
Total processor threads: 4
This system has 1 dual-core processor with hyper-threading (2 threads per core) running at an estimated 2.65GHz
Potresti provare questo comando.
$sudo dmidecode -t cache
$ sudo dmidecode -t cache | grep -iE "leve|installed"
Configuration: Enabled, Socketed, Level 1
Installed Size: 32 kB
Installed SRAM Type: Asynchronous
Configuration: Enabled, Socketed, Level 2
Installed Size: 256 kB
Installed SRAM Type: Burst
Configuration: Enabled, Socketed, Level 3
Installed Size: 3072 kB
Installed SRAM Type: Burst
Per vedere la RAM è sufficiente aggiungere lo switch aggiuntivo -t memory
.
$ sudo dmidecode -t cache -t memory
getconf
getconf -a | grep CACHE
dà:
LEVEL1_ICACHE_SIZE 32768
LEVEL1_ICACHE_ASSOC 8
LEVEL1_ICACHE_LINESIZE 64
LEVEL1_DCACHE_SIZE 32768
LEVEL1_DCACHE_ASSOC 8
LEVEL1_DCACHE_LINESIZE 64
LEVEL2_CACHE_SIZE 262144
LEVEL2_CACHE_ASSOC 8
LEVEL2_CACHE_LINESIZE 64
LEVEL3_CACHE_SIZE 20971520
LEVEL3_CACHE_ASSOC 20
LEVEL3_CACHE_LINESIZE 64
LEVEL4_CACHE_SIZE 0
LEVEL4_CACHE_ASSOC 0
LEVEL4_CACHE_LINESIZE 0
O per un singolo livello:
getconf LEVEL2_CACHE_SIZE
La cosa interessante di questa interfaccia è che è solo un wrapper per la sysconf
funzione POSIX C (gli argomenti della cache sono estensioni non POSIX) e quindi può essere utilizzato anche dal codice C.
Testato in Ubuntu 16.04.
istruzione CPUID x86
L'istruzione CPUID x86 offre anche informazioni sulla cache e può essere letta direttamente dall'utenteland: https://en.wikipedia.org/wiki/CPUID
sembra che glibc usi quel metodo per x86. Non ho confermato dal debugging / traccia delle istruzioni, ma il sorgente per 2.28 lo sysdeps/x86/cacheinfo.c
fa:
__cpuid (2, eax, ebx, ecx, edx);
TODO crea un esempio C minimo, pigro ora, richiesto a: /programming/14283171/how-to-receive-l1-l2-l3-cache-size-using-cpuid-instruction-in-x86
ARM ha anche un meccanismo definito dall'architettura per trovare le dimensioni della cache attraverso registri come il registro ID dimensioni cache (CCSIDR), vedere il Manuale dei programmatori ARMv8 11.6 " Rilevamento della cache" per una panoramica.
Esistono file speciali esportati nel sysfs
filesystem / sys Linux dal 2008:
https://www.kernel.org/doc/Documentation/ABI/testing/sysfs-devices-system-cpu
What: /sys/devices/system/cpu/cpu*/cache/index*/<set_of_attributes_mentioned_below>
Date: July 2014(documented, existed before August 2008)
Description: Parameters for the CPU cache attributes
allocation_policy:
- WriteAllocate: allocate a memory location to a cache line
on a cache miss because of a write
- ReadAllocate: allocate a memory location to a cache line
on a cache miss because of a read
- ReadWriteAllocate: both writeallocate and readallocate
coherency_line_size: the minimum amount of data in bytes that gets
transferred from memory to cache
level: the cache hierarchy in the multi-level cache configuration
number_of_sets: total number of sets in the cache, a set is a
collection of cache lines with the same cache index
physical_line_partition: number of physical cache line per cache tag
shared_cpu_list: the list of logical cpus sharing the cache
shared_cpu_map: logical cpu mask containing the list of cpus sharing
the cache
size: the total cache size in kB
type:
- Instruction: cache that only holds instructions
- Data: cache that only caches data
- Unified: cache that holds both data and instructions
ways_of_associativity: degree of freedom in placing a particular block
of memory in the cache
write_policy:
- WriteThrough: data is written to both the cache line
and to the block in the lower-level memory
- WriteBack: data is written only to the cache line and
the modified cache line is written to main
memory only when it is replaced
File ID:
What: /sys/devices/system/cpu/cpu*/cache/index*/id
Date: September 2016
Contact: Linux kernel mailing list <linux-kernel@vger.kernel.org>
Description: Cache id
The id provides a unique number for a specific instance of
a cache of a particular type. E.g. there may be a level
3 unified cache on each socket in a server and we may
assign them ids 0, 1, 2, ...
Note that id value can be non-contiguous. E.g. level 1
caches typically exist per core, but there may not be a
power of two cores on a socket, so these caches may be
numbered 0, 1, 2, 3, 4, 5, 8, 9, 10, ...
index0
ed index1
entrambi lo facciano level
1
sul kernel v4.15, è confuso. Codifica preventiva 0-index? :-)
Un'altra opzione è il programma cpuid . Utilizza le CPUID
istruzioni e non richiede root. Può anche funzionare attraverso il cpuid
modulo del kernel Linux.
cache and TLB information (2):
0x59: data TLB: 4K pages, 16 entries
0xba: data TLB: 4K pages, 4-way, 64 entries
0x4f: instruction TLB: 4K pages, 32 entries
0xc0: data TLB: 4K & 4M pages, 4-way, 8 entries
0x80: L2 cache: 512K, 8-way, 64 byte lines
0x30: L1 cache: 32K, 8-way, 64 byte lines
0x0e: L1 data cache: 24K, 6-way, 64 byte lines
Notare che nelle comuni CPU consumer le cache L1 e L2 sono per core, mentre la cache L3 è condivisa da tutti i core.
se vuoi solo L3, allora grep "cache size" < /proc/cpuinfo
dovrebbe essere sufficiente.
Tuttavia, poiché il modo in cui la cache L3 è condivisa tra gli archi cpu differisce, potrebbe essere necessario normalizzarne il valore