大号轻量?腾讯云SA3测试

标准型 SA3 实例采用最新一代 AMD EPYC™ 处理器的实例,提供多种规格,具有超高性价比。此实例提供了平衡的计算、存储、网络资源,是众多应用程序的最佳选择。

其实例采用 AMD EPYC™ Milan 全新处理器,内存采用最新 DDR4,默认网络优化,最高内网收发能力达1900万pps,最高内网带宽可支持100Gbps。

本次测试的服务器为SA3.MEDIUM4,vCPU及内存为2C4G,未安装BBR等加速内核,以下为部分其他配置

规格vCPU内存(GB)网络收发包(pps)(出+入)连接数队列数内网带宽能力(Gbps)(出+入)主频备注
SA3.MEDIUM22230万25万21.52.55GHz-
SA3.MEDIUM42430万25万21.52.55GHz-
SA3.MEDIUM82830万25万21.52.55GHz-
SA3.LARGE84850万25万422.55GHz-
SA3.LARGE1641650万25万422.55GHz-
SA3.2XLARGE1681680万50万842.55GHz-
SA3.2XLARGE3283280万50万842.55GHz-
SA3.4XLARGE321632150万110万1672.55GHz-

完整配置见腾讯云官网的 产品介绍页

Bench.sh 测试

Bench.sh 可以查看 Linux 系统信息,还可以测试网络带宽及硬盘读写速率 由于是国内服务器,所以不对网络进行测试了,这里放出系统信息和硬盘速率

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
-------------------- A Bench.sh Script By Teddysun -------------------
Version : v2022-02-22
Usage : wget -qO- bench.sh | bash
----------------------------------------------------------------------
CPU Model : AMD EPYC 7K83 64-Core Processor
CPU Cores : 2 @ 2545.218 MHz
CPU Cache : 512 KB
AES-NI : Enabled
VM-x/AMD-V : Disabled
Total Disk : 78.7 GB (2.1 GB Used)
Total Mem : 3.5 GB (168.2 MB Used)
System uptime : 0 days, 2 hour 44 min
Load average : 0.00, 0.00, 0.06
OS : Debian GNU/Linux 11
Arch : x86_64 (64 Bit)
Kernel : 5.10.0-12-amd64
TCP CC : cubic
Virtualization : KVM
Organization : AS45090 Shenzhen Tencent Computer Systems Company Limited
Location : Beijing / CN
Region : Beijing
----------------------------------------------------------------------
I/O Speed(1st run) : 301 MB/s
I/O Speed(2nd run) : 278 MB/s
I/O Speed(3rd run) : 278 MB/s
I/O Speed(average) : 285.7 MB/s
----------------------------------------------------------------------

SSD云硬盘速度达到了280MB/s,为普通的SSD水品,介于机械硬盘和NVME硬盘之间,比之前测试的腾讯云轻量应用服务器较好

LemonBench

限于篇幅,此处只放出CPU&内存&硬盘的测试结果,完整结果见UbuntuPastebin

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
-> CPU Performance Test (Standard Mode, 3-Pass @ 30sec)

1 Thread Test: 3671 Scores
2 Threads Test: 4152 Scores

-> Memory Performance Test (Standard Mode, 3-Pass @ 30sec)

1 Thread - Read Test : 46248.03 MB/s
1 Thread - Write Test: 26844.49 MB/s

-> Disk Speed Test (4K Block/1M Block, Direct-Write)

Test Name Write Speed Read Speed
10MB-4K Block 10.0 MB/s (2431 IOPS, 1.05 s) 20.6 MB/s (5022 IOPS, 0.51 s)
10MB-1M Block 135 MB/s (128 IOPS, 0.08 s) 264 MB/s (251 IOPS, 0.04 s)
100MB-4K Block 10.0 MB/s (2446 IOPS, 10.47 s) 20.4 MB/s (4971 IOPS, 5.15 s)
100MB-1M Block 180 MB/s (171 IOPS, 0.58 s) 385 MB/s (367 IOPS, 0.27 s)
1GB-4K Block 9.9 MB/s (2427 IOPS, 105.45 s) 16.1 MB/s (3921 IOPS, 65.29 s)
1GB-1M Block 200 MB/s (191 IOPS, 5.23 s) 315 MB/s (300 IOPS, 3.33 s)

CPU(EPYC 7K83)单核分数达到了3671分,对比曾经测试的硅谷轻量(Gold 6133 CPU)的905分和成都轻量(EPYC 7K62)的1304分有了不小的提升,可见这颗"芯"还是不错的

内网测试

既然提到了内网,那不测一下怎么行,简单跑了下iperf3

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Connecting to host 172.16.xxx.xxx, port 5201
[ 5] local 172.16.xxx.xxx port 44176 connected to 172.16.xxx.xxx port 5201
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 307 MBytes 2.57 Gbits/sec 3144 131 KBytes
[ 5] 1.00-2.00 sec 200 MBytes 1.68 Gbits/sec 1378 48.3 KBytes
[ 5] 2.00-3.00 sec 177 MBytes 1.48 Gbits/sec 1795 68.9 KBytes
[ 5] 3.00-4.00 sec 167 MBytes 1.41 Gbits/sec 790 75.8 KBytes
[ 5] 4.00-5.00 sec 219 MBytes 1.83 Gbits/sec 1926 97.9 KBytes
[ 5] 5.00-6.00 sec 154 MBytes 1.30 Gbits/sec 2055 60.7 KBytes
[ 5] 6.00-7.00 sec 213 MBytes 1.79 Gbits/sec 2433 97.9 KBytes
[ 5] 7.00-8.00 sec 155 MBytes 1.30 Gbits/sec 1052 262 KBytes
[ 5] 8.00-9.00 sec 211 MBytes 1.77 Gbits/sec 1478 81.4 KBytes
[ 5] 9.00-10.00 sec 158 MBytes 1.32 Gbits/sec 2225 24.8 KBytes

---

[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 1.92 GBytes 1.65 Gbits/sec 18276 sender
[ 5] 0.00-10.00 sec 1.91 GBytes 1.64 Gbits/sec receiver

iperf Done.

达到了1.5~2.0Gbits/sec 符合预期水平

UnixBench

UnixBench 是一款开源的测试 Unix 系统基本性能的工具,是比较通用的测试VPS性能的工具 完整数据见 UbuntuPastebin

单任务:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Benchmark Run: Fri Apr 29 2022 22:53:07 - 23:21:05
2 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables 44734686.7 lps (10.0 s, 7 samples)
Double-Precision Whetstone 7618.1 MWIPS (9.9 s, 7 samples)
Execl Throughput 5861.2 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 1217379.1 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 318649.6 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 3650531.6 KBps (30.0 s, 2 samples)
Pipe Throughput 2405640.6 lps (10.0 s, 7 samples)
Pipe-based Context Switching 50809.9 lps (10.0 s, 7 samples)
Process Creation 11187.2 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 12000.6 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1855.0 lpm (60.0 s, 2 samples)
System Call Overhead 2576839.6 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 44734686.7 3833.3
Double-Precision Whetstone 55.0 7618.1 1385.1
Execl Throughput 43.0 5861.2 1363.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 1217379.1 3074.2
File Copy 256 bufsize 500 maxblocks 1655.0 318649.6 1925.4
File Copy 4096 bufsize 8000 maxblocks 5800.0 3650531.6 6294.0
Pipe Throughput 12440.0 2405640.6 1933.8
Pipe-based Context Switching 4000.0 50809.9 127.0
Process Creation 126.0 11187.2 887.9
Shell Scripts (1 concurrent) 42.4 12000.6 2830.3
Shell Scripts (8 concurrent) 6.0 1855.0 3091.7
System Call Overhead 15000.0 2576839.6 1717.9
========
System Benchmarks Index Score 1760.1

多任务

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Benchmark Run: Fri Apr 29 2022 23:21:05 - 23:49:06
2 CPUs in system; running 2 parallel copies of tests

Dhrystone 2 using register variables 68864489.9 lps (10.0 s, 7 samples)
Double-Precision Whetstone 13038.2 MWIPS (9.9 s, 7 samples)
Execl Throughput 7893.6 lps (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 1514472.5 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 398094.6 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 4636873.7 KBps (30.0 s, 2 samples)
Pipe Throughput 3149205.4 lps (10.0 s, 7 samples)
Pipe-based Context Switching 439878.2 lps (10.0 s, 7 samples)
Process Creation 19942.3 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 14120.6 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1881.0 lpm (60.0 s, 2 samples)
System Call Overhead 3376550.5 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 68864489.9 5901.0
Double-Precision Whetstone 55.0 13038.2 2370.6
Execl Throughput 43.0 7893.6 1835.7
File Copy 1024 bufsize 2000 maxblocks 3960.0 1514472.5 3824.4
File Copy 256 bufsize 500 maxblocks 1655.0 398094.6 2405.4
File Copy 4096 bufsize 8000 maxblocks 5800.0 4636873.7 7994.6
Pipe Throughput 12440.0 3149205.4 2531.5
Pipe-based Context Switching 4000.0 439878.2 1099.7
Process Creation 126.0 19942.3 1582.7
Shell Scripts (1 concurrent) 42.4 14120.6 3330.3
Shell Scripts (8 concurrent) 6.0 1881.0 3135.0
System Call Overhead 15000.0 3376550.5 2251.0
========
System Benchmarks Index Score 2760.2

作为对比,上海地区轻量的得分为单核1478.6,多核2666.8,SA3依然更胜一筹

总结

SA3 作为采用最新一代 AMD EPYC™ 处理器的实例,绝不仅仅是一个“大号轻量”,对比轻量应用服务器更能胜任需要运算性能的环境,但相应的价格也会更高。针对不同的使用环境来选择依然是亘古不变的道理。

大号轻量?腾讯云SA3测试

http://www.oplog.cn/archives/55678.html

作者

啊不都

发布于

2022-04-29

更新于

2022-07-05

许可协议

CC BY 4.0

评论

查看更多