前言#

时隔 Hostbrr 二周年，终于开启了新系列的销售（AMD EPYC 9655），包含和一周年一样的闪购优惠。有需要的可以自行关注。还是老样子，机器在dataforest，同时提供块存储，最大20T，每T是2欧元。相较于ComputeBox ，带宽给的比较多，但是价格也略贵了点。

套餐#

套餐	CPU	内存	带宽	流量	价格 (每月)有AFF
MINI-EPYC Anniversary	1 vCore AMD EPYC 9655	4 GB DDR5 ECC RAM	10 Gbps	10 TB Bandwidth	€3.00 EUR
NICE-EPYC Anniversary	2 vCore AMD EPYC 9655	6 GB DDR5 ECC RAM	10 Gbps	20 TB Bandwidth	€5.00 EUR
EPIC-EPYC Anniversary	4 vCore AMD EPYC 9655	16 GB DDR5 ECC RAM	10 Gbps	40 TB Bandwidth	€9.00 EUR

Provider : Hostbrr
Type/Plan : EPIC-EPYC Anniversary
Processor : AMD EPYC 9655 96-Core Processor
Num of Core : 4 Cores
Memory : 1****6 GB
Storage : 16****0 GB NVMe
Bandwidth : 40****TB @ 10 Gbps IN | 10 Gbps OUT
Location : DE
Price : $11

测评#

CPU#

1
root@catcat:~/scripts# lscpu
2
Architecture:            x86_64
3
  CPU op-mode(s):        32-bit, 64-bit
4
  Address sizes:         52 bits physical, 57 bits virtual
5
  Byte Order:            Little Endian
6
CPU(s):                  4
7
  On-line CPU(s) list:   0-3
8
Vendor ID:               AuthenticAMD
9
  BIOS Vendor ID:        QEMU
10
  Model name:            AMD EPYC 9655 96-Core Processor
11
    BIOS Model name:     pc-i440fx-7.2  CPU @ 2.0GHz
12
    BIOS CPU family:     1
13
    CPU family:          26
14
    Model:               2
15
    Thread(s) per core:  1
16
    Core(s) per socket:  1
17
    Socket(s):           4
18
    Stepping:            1
19
    BogoMIPS:            5200.00
20
    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_g
21
                         ood nopl cpuid extd_apicid tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c
22
                         rdrand hypervisor lahf_lm cmp_legacy svm cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw perfctr_core invpcid_single ssbd ibrs ibpb stibp vmmcal
23
                         l fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl
24
                          xsaveopt xsavec xgetbv1 xsaves avx_vnni avx512_bf16 clzero xsaveerptr wbnoinvd arat npt lbrv nrip_save tsc_scale vmcb_clean pausefilter pfthreshold
25
                          v_vmsave_vmload vgif avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq la57 rdpid movdiri movd
26
                         ir64b fsrm avx512_vp2intersect arch_capabilities
27
Virtualization features:
28
  Virtualization:        AMD-V
29
  Hypervisor vendor:     KVM
30
  Virtualization type:   full
31
Caches (sum of all):
32
  L1d:                   192 KiB (4 instances)
33
  L1i:                   128 KiB (4 instances)
34
  L2:                    4 MiB (4 instances)
35
  L3:                    1.5 GiB (4 instances)
36
NUMA:
37
  NUMA node(s):          1
38
  NUMA node0 CPU(s):     0-3
39
Vulnerabilities:
40
  Itlb multihit:         Not affected
41
  L1tf:                  Not affected
42
  Mds:                   Not affected
43
  Meltdown:              Not affected
44
  Mmio stale data:       Not affected
45
  Retbleed:              Not affected
46
  Spec store bypass:     Mitigation; Speculative Store Bypass disabled via prctl
47
  Spectre v1:            Mitigation; usercopy/swapgs barriers and __user pointer sanitization
48
  Spectre v2:            Mitigation; Retpolines, IBPB conditional, IBRS_FW, STIBP disabled, RSB filling, PBRSB-eIBRS Not affected
49
  Srbds:                 Not affected
50
  Tsx async abort:       Not affected
51
root@catcat:~/scripts#

Yabs#

1
Basic System Information:
2
---------------------------------
3
Uptime     : 0 days, 0 hours, 13 minutes
4
Processor  : AMD EPYC 9655 96-Core Processor
5
CPU cores  : 4 @ 2600.002 MHz
6
AES-NI     : ✔ Enabled
7
VM-x/AMD-V : ✔ Enabled
8
RAM        : 15.6 GiB
9
Swap       : 1024.0 MiB
10
Disk       : 159.9 GiB
11
Distro     : Debian GNU/Linux 12 (bookworm)
12
Kernel     : 6.1.0-9-amd64
13
VM Type    : KVM
14
IPv4/IPv6  : ✔ Online / ✔ Online
15

16
IPv6 Network Information:
17
---------------------------------
18
ISP        : dataforest GmbH
19
ASN        : AS58212 dataforest GmbH
20
Host       : dataforest GmbH
21
Location   : Kriftel, Hesse (HE)
22
Country    : Germany
23

24
fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/vda3):
25
---------------------------------
26
Block Size | 4k            (IOPS) | 64k           (IOPS)
27
  ------ | --- ---- | ---- ----
28
Read       | 488.54 MB/s (122.1k) | 3.68 GB/s    (57.5k)
29
Write      | 489.83 MB/s (122.4k) | 3.70 GB/s    (57.8k)
30
Total      | 978.37 MB/s (244.5k) | 7.38 GB/s   (115.3k)
31
           |                      |
32
Block Size | 512k          (IOPS) | 1m            (IOPS)
33
  ------ | --- ---- | ---- ----
34
Read       | 685.76 MB/s   (1.3k) | 4.49 GB/s     (4.3k)
35
Write      | 722.19 MB/s   (1.4k) | 4.78 GB/s     (4.6k)
36
Total      | 1.40 GB/s     (2.7k) | 9.27 GB/s     (9.0k)
37

38
iperf3 Network Speed Tests (IPv4):
39
---------------------------------
40
Provider        | Location (Link)           | Send Speed      | Recv Speed      | Ping
41
----- | ----- | ---- | ---- | ----
42
Clouvider       | London, UK (10G)          | 6.81 Gbits/sec  | 5.97 Gbits/sec  | 13.9 ms
43
Eranium         | Amsterdam, NL (100G)      | 9.36 Gbits/sec  | 8.66 Gbits/sec  | 7.69 ms
44
Uztelecom       | Tashkent, UZ (10G)        | 1.43 Gbits/sec  | 1.50 Gbits/sec  | 91.4 ms
45
Leaseweb        | Singapore, SG (10G)       | 499 Mbits/sec   | 1.10 Gbits/sec  | 161 ms
46
Clouvider       | Los Angeles, CA, US (10G) | 956 Mbits/sec   | 995 Mbits/sec   | 149 ms
47
Leaseweb        | NYC, NY, US (10G)         | 2.07 Gbits/sec  | 1.97 Gbits/sec  | 86.4 ms
48
Edgoo           | Sao Paulo, BR (1G)        | 879 Mbits/sec   | 665 Mbits/sec   | 188 ms
49

50
iperf3 Network Speed Tests (IPv6):
51
---------------------------------
52
Provider        | Location (Link)           | Send Speed      | Recv Speed      | Ping
53
----- | ----- | ---- | ---- | ----
54
Clouvider       | London, UK (10G)          | 6.52 Gbits/sec  | 5.87 Gbits/sec  | 13.9 ms
55
Eranium         | Amsterdam, NL (100G)      | 8.29 Gbits/sec  | 14.5 Gbits/sec  | 7.64 ms
56
Uztelecom       | Tashkent, UZ (10G)        | 14.1 Mbits/sec  | 1.72 Gbits/sec  | 90.4 ms
57
Leaseweb        | Singapore, SG (10G)       | 833 Mbits/sec   | 1.10 Gbits/sec  | --
58
Clouvider       | Los Angeles, CA, US (10G) | 1.16 Gbits/sec  | 1.25 Gbits/sec  | 149 ms
59
Leaseweb        | NYC, NY, US (10G)         | 2.12 Gbits/sec  | 2.24 Gbits/sec  | 83.6 ms
60
Edgoo           | Sao Paulo, BR (1G)        | 898 Mbits/sec   | 624 Mbits/sec   | 188 ms
61

62
Geekbench 6 BenchMark Test:
63
---------------------------------
64
Test            | Value
65
                |
66
Single Core     | 2843
67
Multi Core      | 8971
68
Full Test       | https://browser.geekbench.com/v6/cpu/11569747

GeekBench5#

1
Geekbench 5 测试结果
2

3
系统信息
4
  Operating System              Debian GNU/Linux 12 (bookworm)
5
  Kernel                        Linux 6.1.0-9-amd64 x86_64
6
  Model                         QEMU Standard PC (i440FX + PIIX, 1996)
7
  Motherboard                   N/A
8
  BIOS                          SeaBIOS 1.16.2-debian-1.16.2-1
9

10
处理器信息
11
  Name                          AMD EPYC 9655 96-Core Processor
12
  Topology                      4 Processors, 4 Cores
13
  Identifier                    AuthenticAMD Family 26 Model 2 Stepping 1
14
  Base Frequency                2.60 GHz
15
  L1 Instruction Cache          32.0 KB
16
  L1 Data Cache                 48.0 KB
17
  L2 Cache                      1.00 MB
18
  L3 Cache                      384 MB
19

20
内存信息
21
  Size                          15.6 GB
22

23
单核测试分数：1796
24
多核测试分数：6605
25
详细结果链接：https://browser.geekbench.com/v5/cpu/23482496
26
可供参考链接：https://browser.geekbench.com/search?k=v5_cpu&q=AMD%20EPYC%209655

PerformanceTest Linux 测试#

1
AMD EPYC 9655 96-Core Processor (x86_64)
2
4 cores @ 0 MHz  |  15.6 GiB RAM
3
Number of Processes: 4  |  Test Iterations: 1  |  Test Duration: Medium
4
--------------------------------------------------------------------------
5
CPU Mark:                          13651
6
  Integer Math                     30796 Million Operations/s
7
  Floating Point Math              30814 Million Operations/s
8
  Prime Numbers                    121 Million Primes/s
9
  Sorting                          18025 Thousand Strings/s
10
  Encryption                       6374 MB/s
11
  Compression                      135768 KB/s
12
  CPU Single Threaded              3837 Million Operations/s
13
  Physics                          2033 Frames/s
14
  Extended Instructions (SSE)      11901 Million Matrices/s
15

16
Memory Mark:                       2565
17
  Database Operations              6062 Thousand Operations/s
18
  Memory Read Cached               34115 MB/s
19
  Memory Read Uncached             31626 MB/s
20
  Memory Write                     25638 MB/s
21
  Available RAM                    14506 Megabytes
22
  Memory Latency                   85 Nanoseconds
23
  Memory Threaded                  127675 MB/s
24
--------------------------------------------------------------------------

UnixBench#

1
   BYTE UNIX BenchMarks (Version 5.1.3)
2

3
   System: catcat: GNU/Linux
4
   OS: GNU/Linux -- 6.1.0-9-amd64 -- #1 SMP PREEMPT_DYNAMIC Debian 6.1.27-1 (2023-05-08)
5
   Machine: x86_64 (unknown)
6
   Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
7
   CPU 0: AMD EPYC 9655 96-Core Processor (5200.0 bogomips)
8
          x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
9
   CPU 1: AMD EPYC 9655 96-Core Processor (5200.0 bogomips)
10
          x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
11
   CPU 2: AMD EPYC 9655 96-Core Processor (5200.0 bogomips)
12
          x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
13
   CPU 3: AMD EPYC 9655 96-Core Processor (5200.0 bogomips)
14
          x86-64, MMX, AMD MMX, Physical Address Ext, SYSENTER/SYSEXIT, AMD virtualization, SYSCALL/SYSRET
15
   15:57:35 up  1:19,  2 users,  load average: 0.21, 0.05, 0.06; runlevel 2025-04-18
16

17
------------------------------------------------------------------------
18
BenchMark Run: Fri Apr 18 2025 15:57:35 - 16:25:31
19
4 CPUs in system; running 1 parallel copy of tests
20

21
Dhrystone 2 using register variables       88148615.4 lps   (10.0 s, 7 samples)
22
Double-Precision Whetstone                    11071.9 MWIPS (10.0 s, 7 samples)
23
Execl Throughput                               7383.2 lps   (30.0 s, 2 samples)
24
File Copy 1024 bufsize 2000 maxblocks       1904581.4 KBps  (30.0 s, 2 samples)
25
File Copy 256 bufsize 500 maxblocks          484757.8 KBps  (30.0 s, 2 samples)
26
File Copy 4096 bufsize 8000 maxblocks       6968637.3 KBps  (30.0 s, 2 samples)
27
Pipe Throughput                             3350920.7 lps   (10.0 s, 7 samples)
28
Pipe-based Context Switching                 181980.3 lps   (10.0 s, 7 samples)
29
Process Creation                               9623.9 lps   (30.0 s, 2 samples)
30
Shell Scripts (1 concurrent)                   8255.6 lpm   (60.0 s, 2 samples)
31
Shell Scripts (8 concurrent)                   5854.5 lpm   (60.0 s, 2 samples)
32
System Call Overhead                        3048128.9 lps   (10.0 s, 7 samples)
33

34
System BenchMarks Index Values               BASELINE       RESULT    INDEX
35
Dhrystone 2 using register variables         116700.0   88148615.4   7553.4
36
Double-Precision Whetstone                       55.0      11071.9   2013.1
37
Execl Throughput                                 43.0       7383.2   1717.0
38
File Copy 1024 bufsize 2000 maxblocks          3960.0    1904581.4   4809.5
39
File Copy 256 bufsize 500 maxblocks            1655.0     484757.8   2929.1
40
File Copy 4096 bufsize 8000 maxblocks          5800.0    6968637.3  12014.9
41
Pipe Throughput                               12440.0    3350920.7   2693.7
42
Pipe-based Context Switching                   4000.0     181980.3    455.0
43
Process Creation                                126.0       9623.9    763.8
44
Shell Scripts (1 concurrent)                     42.4       8255.6   1947.1
45
Shell Scripts (8 concurrent)                      6.0       5854.5   9757.6
46
System Call Overhead                          15000.0    3048128.9   2032.1
47
                                                                   ========
48
System BenchMarks Index Score                                        2714.0
49

50
------------------------------------------------------------------------
51
BenchMark Run: Fri Apr 18 2025 16:25:31 - 16:53:28
52
4 CPUs in system; running 4 parallel copies of tests
53

54
Dhrystone 2 using register variables      349054606.9 lps   (10.0 s, 7 samples)
55
Double-Precision Whetstone                    44341.5 MWIPS (10.0 s, 7 samples)
56
Execl Throughput                              15961.4 lps   (30.0 s, 2 samples)
57
File Copy 1024 bufsize 2000 maxblocks       7382311.8 KBps  (30.0 s, 2 samples)
58
File Copy 256 bufsize 500 maxblocks         1933223.9 KBps  (30.0 s, 2 samples)
59
File Copy 4096 bufsize 8000 maxblocks      24514369.9 KBps  (30.0 s, 2 samples)
60
Pipe Throughput                            13353957.9 lps   (10.0 s, 7 samples)
61
Pipe-based Context Switching                1188394.8 lps   (10.0 s, 7 samples)
62
Process Creation                              49987.6 lps   (30.0 s, 2 samples)
63
Shell Scripts (1 concurrent)                  47257.0 lpm   (60.0 s, 2 samples)
64
Shell Scripts (8 concurrent)                   6849.4 lpm   (60.0 s, 2 samples)
65
System Call Overhead                       12142661.2 lps   (10.0 s, 7 samples)
66

67
System BenchMarks Index Values               BASELINE       RESULT    INDEX
68
Dhrystone 2 using register variables         116700.0  349054606.9  29910.4
69
Double-Precision Whetstone                       55.0      44341.5   8062.1
70
Execl Throughput                                 43.0      15961.4   3712.0
71
File Copy 1024 bufsize 2000 maxblocks          3960.0    7382311.8  18642.2
72
File Copy 256 bufsize 500 maxblocks            1655.0    1933223.9  11681.1
73
File Copy 4096 bufsize 8000 maxblocks          5800.0   24514369.9  42266.2
74
Pipe Throughput                               12440.0   13353957.9  10734.7
75
Pipe-based Context Switching                   4000.0    1188394.8   2971.0
76
Process Creation                                126.0      49987.6   3967.3
77
Shell Scripts (1 concurrent)                     42.4      47257.0  11145.5
78
Shell Scripts (8 concurrent)                      6.0       6849.4  11415.7
79
System Call Overhead                          15000.0   12142661.2   8095.1
80
                                                                   ========
81
System BenchMarks Index Score                                       10059.4