본문 바로가기

클라우드

iwinv High Performance Gen 2 인스턴스 벤치마크

728x90

iwinv High Performance Gen 2가 추가되어서 벤치를 돌려봤다.

 

2025.05.24 - [클라우드] - iwinv 인스턴스 타입 비교 벤치마크

 

iwinv 인스턴스 타입 비교 벤치마크

이전에 Vultr 벤치마크를 돌려보고 타 클라우드 서비스 성능이 궁금하기도 했고, 최근에 잠깐 iwinv 인스턴스를 생성해서 잠깐 사용하는 김에 벤치를 돌려봤는데 의외로(?) 점수가 꽤 잘 나와서 아

blog.sysbot32.com

 

대상

  • High Performance - Gen2 Dedicated
    • 2 vCPUs, 4GB, 41,800원/월 (hpa_2.4-g2)

이전과 마찬가지로 KR1-Z01, Ubuntu 24.04 LTS에서 진행했다.

 

CPU 모델 확인

더보기

processor : 0
vendor_id : AuthenticAMD
cpu family : 26
model : 68
model name : AMD EPYC 4545P 16-Core Processor
stepping : 0
microcode : 0xb40401c
cpu MHz : 2994.374
cache size : 512 KB
physical id : 0
siblings : 1
core id : 0
cpu cores : 1
apicid : 0
initial apicid : 0
fpu : yes
fpu_exception : yes
cpuid level : 16
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm rep_good nopl cpuid extd_apicid tsc_known_freq pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy svm cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw perfctr_core ssbd ibrs ibpb stibp ibrs_enhanced vmmcall fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves avx_vnni avx512_bf16 clzero xsaveerptr wbnoinvd arat npt lbrv nrip_save tsc_scale vmcb_clean flushbyasid pausefilter pfthreshold v_vmsave_vmload vgif vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq rdpid movdiri movdir64b overflow_recov succor fsrm avx512_vp2intersect flush_l1d arch_capabilities
bugs : sysret_ss_attrs spectre_v1 spectre_v2 spec_store_bypass ibpb_no_ret
bogomips : 5988.74
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 48 bits physical, 48 bits virtual
power management:

processor : 1
vendor_id : AuthenticAMD
cpu family : 26
model : 68
model name : AMD EPYC 4545P 16-Core Processor
stepping : 0
microcode : 0xb40401c
cpu MHz : 2994.374
cache size : 512 KB
physical id : 1
siblings : 1
core id : 0
cpu cores : 1
apicid : 1
initial apicid : 1
fpu : yes
fpu_exception : yes
cpuid level : 16
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm rep_good nopl cpuid extd_apicid tsc_known_freq pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy svm cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw perfctr_core ssbd ibrs ibpb stibp ibrs_enhanced vmmcall fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves avx_vnni avx512_bf16 clzero xsaveerptr wbnoinvd arat npt lbrv nrip_save tsc_scale vmcb_clean flushbyasid pausefilter pfthreshold v_vmsave_vmload vgif vnmi avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq rdpid movdiri movdir64b overflow_recov succor fsrm avx512_vp2intersect flush_l1d arch_capabilities
bugs : sysret_ss_attrs spectre_v1 spectre_v2 spec_store_bypass ibpb_no_ret
bogomips : 5988.74
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 48 bits physical, 48 bits virtual
power management:

 

싱글 스레드 벤치마크

더보기

sysbench 1.0.20 (using system LuaJIT 2.1.0-beta3)

Running the test with following options:
Number of threads: 1
Initializing random number generator from current time


Prime numbers limit: 20000

Initializing worker threads...

Threads started!

CPU speed:
    events per second:  2540.76

General statistics:
    total time:                          3.9351s
    total number of events:              10000

Latency (ms):
         min:                                    0.39
         avg:                                    0.39
         max:                                    0.45
         95th percentile:                        0.39
         sum:                                 3934.30

Threads fairness:
    events (avg/stddev):           10000.0000/0.00
    execution time (avg/stddev):   3.9343/0.00

 

멀티 스레드 벤치마크 (2)

더보기

sysbench 1.0.20 (using system LuaJIT 2.1.0-beta3)

Running the test with following options:
Number of threads: 2
Initializing random number generator from current time


Prime numbers limit: 20000

Initializing worker threads...

Threads started!

CPU speed:
    events per second:  4974.11

General statistics:
    total time:                          2.0098s
    total number of events:              10000

Latency (ms):
         min:                                    0.39
         avg:                                    0.40
         max:                                    0.80
         95th percentile:                        0.42
         sum:                                 4017.30

Threads fairness:
    events (avg/stddev):           5000.0000/109.00
    execution time (avg/stddev):   2.0086/0.00

 

멀티 스레드 벤치마크 (4)

더보기

sysbench 1.0.20 (using system LuaJIT 2.1.0-beta3)

Running the test with following options:
Number of threads: 4
Initializing random number generator from current time


Prime numbers limit: 20000

Initializing worker threads...

Threads started!

CPU speed:
    events per second:  4984.42

General statistics:
    total time:                          2.0056s
    total number of events:              10000

Latency (ms):
         min:                                    0.39
         avg:                                    0.80
         max:                                    6.22
         95th percentile:                        2.43
         sum:                                 8004.39

Threads fairness:
    events (avg/stddev):           2500.0000/41.02
    execution time (avg/stddev):   2.0011/0.00

 

메모리 읽기 테스트

더보기

sysbench 1.0.20 (using system LuaJIT 2.1.0-beta3)

Running the test with following options:
Number of threads: 1
Initializing random number generator from current time


Running memory speed test with the following options:
  block size: 1KiB
  total size: 102400MiB
  operation: read
  scope: global

Initializing worker threads...

Threads started!

Total operations: 104857600 (14003909.04 per second)

102400.00 MiB transferred (13675.69 MiB/sec)


General statistics:
    total time:                          7.4866s
    total number of events:              104857600

Latency (ms):
         min:                                    0.00
         avg:                                    0.00
         max:                                    0.04
         95th percentile:                        0.00
         sum:                                 2901.57

Threads fairness:
    events (avg/stddev):           104857600.0000/0.00
    execution time (avg/stddev):   2.9016/0.00

 

메모리 쓰기 테스트

더보기

sysbench 1.0.20 (using system LuaJIT 2.1.0-beta3)

Running the test with following options:
Number of threads: 1
Initializing random number generator from current time


Running memory speed test with the following options:
  block size: 1KiB
  total size: 102400MiB
  operation: write
  scope: global

Initializing worker threads...

Threads started!

Total operations: 104857600 (11358955.11 per second)

102400.00 MiB transferred (11092.73 MiB/sec)


General statistics:
    total time:                          9.2306s
    total number of events:              104857600

Latency (ms):
         min:                                    0.00
         avg:                                    0.00
         max:                                    0.08
         95th percentile:                        0.00
         sum:                                 4435.83

Threads fairness:
    events (avg/stddev):           104857600.0000/0.00
    execution time (avg/stddev):   4.4358/0.00

 

디스크 테스트

더보기

sysbench 1.0.20 (using system LuaJIT 2.1.0-beta3)

Running the test with following options:
Number of threads: 1
Initializing random number generator from current time


Extra file open flags: (none)
128 files, 64MiB each
8GiB total file size
Block size 16KiB
Number of IO requests: 0
Read/Write ratio for combined random IO test: 1.50
Periodic FSYNC enabled, calling fsync() each 100 requests.
Calling fsync() at the end of test, Enabled.
Using synchronous I/O mode
Doing random r/w test
Initializing worker threads...

Threads started!


File operations:
    reads/s:                      871.52
    writes/s:                     581.01
    fsyncs/s:                     1859.37

Throughput:
    read, MiB/s:                  13.62
    written, MiB/s:               9.08

General statistics:
    total time:                          300.0754s
    total number of events:              993692

Latency (ms):
         min:                                    0.00
         avg:                                    0.30
         max:                                   88.35
         95th percentile:                        1.70
         sum:                               299789.43

Threads fairness:
    events (avg/stddev):           993692.0000/0.00
    execution time (avg/stddev):   299.7894/0.00

 

결론

  hpa_2.4-g2
명시된 스펙 2 vCPUs, 4GB
CPU 모델 AMD EPYC 4545P 16-Core Processor
싱글 스레드 2540.76
멀티 스레드 (2) 4974.11
멀티 스레드 (4) 4984.42
메모리 읽기 13675.69
메모리 쓰기 11092.73
디스크 13.62
9.08
가격 41,800원/월

 

이전 세대 High Performance에 비하면 성능이 굉장히 좋아졌다!

하지만, General Purpose와의 가격 차이에 비하면 성능 차이는 크지 않은 것으로 보인다.

 

서버용 CPU, 10Gbps가 필요하다 👉 High Performance Gen 2

일반 CPU, 2.5Gbps로도 충분하다 👉 General Purpose

 

인터넷 속도 테스트

인터넷 속도 테스트는 외부 영향을 많이 받아서 테스트 결과가 항상 일정하게 나온다는 보장이 없어서 그동안 안 했는데, 이번에는 번외로 한 번 돌려보고자 한다.

 

ping 테스트

AWS 리전별 S3 엔드포인트에 ping을 10번씩 날렸다.

 

🇰🇷 서울

더보기

PING s3.ap-northeast-2.amazonaws.com (3.5.189.41) 56(84) bytes of data.
64 bytes from 3.5.189.41: icmp_seq=1 ttl=53 time=2.52 ms
64 bytes from 3.5.189.41: icmp_seq=2 ttl=53 time=2.40 ms
64 bytes from 3.5.189.41: icmp_seq=3 ttl=53 time=2.42 ms
64 bytes from 3.5.189.41: icmp_seq=4 ttl=53 time=2.50 ms
64 bytes from 3.5.189.41: icmp_seq=5 ttl=53 time=2.43 ms
64 bytes from 3.5.189.41: icmp_seq=6 ttl=53 time=2.42 ms
64 bytes from 3.5.189.41: icmp_seq=7 ttl=53 time=2.45 ms
64 bytes from 3.5.189.41: icmp_seq=8 ttl=53 time=2.45 ms
64 bytes from 3.5.189.41: icmp_seq=9 ttl=53 time=2.44 ms
64 bytes from 3.5.189.41: icmp_seq=10 ttl=53 time=2.43 ms

--- s3.ap-northeast-2.amazonaws.com ping statistics ---
10 packets transmitted, 10 received, 0% packet loss, time 9014ms
rtt min/avg/max/mdev = 2.402/2.446/2.521/0.035 ms

 

🇯🇵 도쿄

더보기

PING s3.ap-northeast-1.amazonaws.com (3.5.155.26) 56(84) bytes of data.
64 bytes from 3.5.155.26: icmp_seq=1 ttl=53 time=35.3 ms
64 bytes from 3.5.155.26: icmp_seq=2 ttl=53 time=35.1 ms
64 bytes from 3.5.155.26: icmp_seq=3 ttl=53 time=35.1 ms
64 bytes from 3.5.155.26: icmp_seq=4 ttl=53 time=35.4 ms
64 bytes from 3.5.155.26: icmp_seq=5 ttl=53 time=35.1 ms
64 bytes from 3.5.155.26: icmp_seq=6 ttl=53 time=35.1 ms
64 bytes from 3.5.155.26: icmp_seq=7 ttl=53 time=35.3 ms
64 bytes from 3.5.155.26: icmp_seq=8 ttl=53 time=35.1 ms
64 bytes from 3.5.155.26: icmp_seq=9 ttl=53 time=35.1 ms
64 bytes from 3.5.155.26: icmp_seq=10 ttl=53 time=35.2 ms

--- s3.ap-northeast-1.amazonaws.com ping statistics ---
10 packets transmitted, 10 received, 0% packet loss, time 9012ms
rtt min/avg/max/mdev = 35.107/35.201/35.421/0.102 ms

 

🇺🇸 버지니아 북부

더보기

PING s3.us-east-1.amazonaws.com (16.15.253.255) 56(84) bytes of data.
64 bytes from 16.15.253.255: icmp_seq=1 ttl=44 time=191 ms
64 bytes from 16.15.253.255: icmp_seq=2 ttl=44 time=191 ms
64 bytes from 16.15.253.255: icmp_seq=3 ttl=44 time=194 ms
64 bytes from 16.15.253.255: icmp_seq=4 ttl=44 time=194 ms
64 bytes from 16.15.253.255: icmp_seq=5 ttl=44 time=194 ms
64 bytes from 16.15.253.255: icmp_seq=6 ttl=44 time=194 ms
64 bytes from 16.15.253.255: icmp_seq=7 ttl=44 time=194 ms
64 bytes from 16.15.253.255: icmp_seq=8 ttl=44 time=194 ms
64 bytes from 16.15.253.255: icmp_seq=9 ttl=44 time=194 ms
64 bytes from 16.15.253.255: icmp_seq=10 ttl=44 time=194 ms

--- s3.us-east-1.amazonaws.com ping statistics ---
10 packets transmitted, 10 received, 0% packet loss, time 9012ms
rtt min/avg/max/mdev = 190.817/193.592/194.445/1.366 ms

 

🇺🇸 캘리포니아 북부

더보기

PING s3.us-west-1.amazonaws.com (16.15.0.167) 56(84) bytes of data.
64 bytes from 16.15.0.167: icmp_seq=1 ttl=51 time=134 ms
64 bytes from 16.15.0.167: icmp_seq=2 ttl=51 time=134 ms
64 bytes from 16.15.0.167: icmp_seq=3 ttl=51 time=131 ms
64 bytes from 16.15.0.167: icmp_seq=4 ttl=51 time=134 ms
64 bytes from 16.15.0.167: icmp_seq=5 ttl=51 time=131 ms
64 bytes from 16.15.0.167: icmp_seq=6 ttl=51 time=131 ms
64 bytes from 16.15.0.167: icmp_seq=7 ttl=51 time=134 ms
64 bytes from 16.15.0.167: icmp_seq=8 ttl=51 time=134 ms
64 bytes from 16.15.0.167: icmp_seq=9 ttl=51 time=134 ms
64 bytes from 16.15.0.167: icmp_seq=10 ttl=51 time=131 ms

--- s3.us-west-1.amazonaws.com ping statistics ---
10 packets transmitted, 10 received, 0% packet loss, time 9013ms
rtt min/avg/max/mdev = 130.612/132.932/134.498/1.842 ms

 

🇸🇬 싱가포르

더보기

PING s3.ap-southeast-1.amazonaws.com (3.5.151.68) 56(84) bytes of data.
64 bytes from 3.5.151.68: icmp_seq=1 ttl=48 time=66.9 ms
64 bytes from 3.5.151.68: icmp_seq=2 ttl=48 time=66.8 ms
64 bytes from 3.5.151.68: icmp_seq=3 ttl=48 time=66.8 ms
64 bytes from 3.5.151.68: icmp_seq=4 ttl=48 time=66.8 ms
64 bytes from 3.5.151.68: icmp_seq=5 ttl=48 time=66.8 ms
64 bytes from 3.5.151.68: icmp_seq=6 ttl=48 time=66.8 ms
64 bytes from 3.5.151.68: icmp_seq=7 ttl=48 time=66.8 ms
64 bytes from 3.5.151.68: icmp_seq=8 ttl=48 time=66.8 ms
64 bytes from 3.5.151.68: icmp_seq=9 ttl=48 time=66.8 ms
64 bytes from 3.5.151.68: icmp_seq=10 ttl=48 time=66.8 ms

--- s3.ap-southeast-1.amazonaws.com ping statistics ---
10 packets transmitted, 10 received, 0% packet loss, time 9007ms
rtt min/avg/max/mdev = 66.766/66.798/66.933/0.046 ms

 

다운로드 테스트

원래 speedtest-cli 쓰려고 했으나, 계속 에러나서 그냥 Vultr 속도 테스트용 파일 다운 받아 보는 걸로.

 

🇰🇷 서울

더보기

--2026-04-18 11:56:06--  https://sel-kor-ping.vultr.com/vultr.com.1000MB.bin
Resolving sel-kor-ping.vultr.com (sel-kor-ping.vultr.com)... 141.164.34.61
Connecting to sel-kor-ping.vultr.com (sel-kor-ping.vultr.com)|141.164.34.61|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 1048576000 (1000M) [application/octet-stream]
Saving to: ‘vultr.cohttp://m.1000MB.bin

vultr.cohttp://m.1000MB.bi 100%[===================>]   1000M   515MB/s    in 1.9s    

2026-04-18 11:56:08 (515 MB/s) - ‘vultr.cohttp://m.1000MB.bin’ saved [1048576000/1048576000]

 

🇯🇵 도쿄

더보기

--2026-04-18 11:57:24--  https://hnd-jp-ping.vultr.com/vultr.com.1000MB.bin
Resolving hnd-jp-ping.vultr.com (hnd-jp-ping.vultr.com)... 108.61.201.151
Connecting to hnd-jp-ping.vultr.com (hnd-jp-ping.vultr.com)|108.61.201.151|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 1048576000 (1000M) [application/octet-stream]
Saving to: ‘vultr.cohttp://m.1000MB.bin.1

vultr.cohttp://m.1000MB.bi 100%[===================>]   1000M   122MB/s    in 9.2s    

2026-04-18 11:57:33 (109 MB/s) - ‘vultr.cohttp://m.1000MB.bin.1’ saved [1048576000/1048576000]

 

🇺🇸 뉴욕

더보기

--2026-04-18 12:01:03--  https://nj-us-ping.vultr.com/vultr.com.1000MB.bin
Resolving nj-us-ping.vultr.com (nj-us-ping.vultr.com)... 108.61.149.182
Connecting to nj-us-ping.vultr.com (nj-us-ping.vultr.com)|108.61.149.182|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 1048576000 (1000M) [application/octet-stream]
Saving to: ‘vultr.cohttp://m.1000MB.bin.2

vultr.cohttp://m.1000MB.bi 100%[===================>]   1000M  5.98MB/s    in 83s     

2026-04-18 12:02:27 (12.0 MB/s) - ‘vultr.cohttp://m.1000MB.bin.2’ saved [1048576000/1048576000]

 

🇺🇸 실리콘밸리 (산호세)

더보기

--2026-04-18 12:02:51--  https://sjo-ca-us-ping.vultr.com/vultr.com.1000MB.bin
Resolving sjo-ca-us-ping.vultr.com (sjo-ca-us-ping.vultr.com)... 104.156.230.107
Connecting to sjo-ca-us-ping.vultr.com (sjo-ca-us-ping.vultr.com)|104.156.230.107|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 1048576000 (1000M) [application/octet-stream]
Saving to: ‘vultr.cohttp://m.1000MB.bin.3

vultr.cohttp://m.1000MB.bi 100%[===================>]   1000M  21.6MB/s    in 48s     

2026-04-18 12:03:39 (20.9 MB/s) - ‘vultr.cohttp://m.1000MB.bin.3’ saved [1048576000/1048576000]

 

🇸🇬 싱가포르

더보기

--2026-04-18 12:05:16--  https://sgp-ping.vultr.com/vultr.com.1000MB.bin
Resolving sgp-ping.vultr.com (sgp-ping.vultr.com)... 45.32.100.168
Connecting to sgp-ping.vultr.com (sgp-ping.vultr.com)|45.32.100.168|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 1048576000 (1000M) [application/octet-stream]
Saving to: ‘vultr.cohttp://m.1000MB.bin.4

vultr.cohttp://m.1000MB.bi 100%[===================>]   1000M  33.4MB/s    in 97s     

2026-04-18 12:06:53 (10.3 MB/s) - ‘vultr.cohttp://m.1000MB.bin.4’ saved [1048576000/1048576000]

 

728x90