NUC 折腾笔记 - 储存能力测试
前文提到过,我计划折腾下这种小盒子,作为日常开发过程中的拓展设备。在短暂使用之后,我为它拓展了两块容量更高的硬盘,开始正式使用。
因为使用了一些网上颇有争议的国产硬件,所以我觉得或许可以记录下来,为后来者作为选择参考。
测试环境
硬件测试环境还是上篇文章提到的 NUC8,存储硬件选择了最近国货之光长江存储和光威,软件环境则选择了 Ubuntu 20.04 LTS 版本。
硬件环境
下面的列表可以更直观的看到本次测试都用到了哪些“东西”:
- NUC8 i5beh(这货不是国货,但是是国产,233)
- 【国货】长江存储 ZHITAI PC005 Active 1TB - NVME
- 【国货】长江存储 ZHITAI SC001 Act 1TB - SATA3
- 【国货】光威 战将 16GB DDR4 2666
- 【国货】光威 弈Pro 8GB DDR4 2666
- 【搭配测试】雷克沙 16G DDR4 2666
软件环境
文章中测试使用的软件,则包含了下面的内容:
- Ubuntu 20.04.2 LTS
- dd (coreutils) v8.3
- hdparm v9.58
- nvme v1.9
- dmidecode v3.2
查看存储设备基础信息
在开始测试之前,我们先来了解下磁盘和内存的基础信息,以及如何获取它们。
查看 SATA SSD
非 NVME 协议的 SATA 盘的信息查看,除了使用 smart 外,还可以使用 hdparm
工具来查看,相比较 smart ,你可以通过它直观看到这个磁盘具备的各种能力:
hdparm -I /dev/sda
代码语言:txt复制/dev/sda:
ATA device, with non-removable media
Model Number: ZHITAI SC001 Active 1TB SSD
Serial Number: ZTB11T0KA20[masked]M
Firmware Revision: ZT013200
Transport: Serial, ATA8-AST, SATA 1.0a, SATA II Extensions, SATA Rev 2.5, SATA Rev 2.6, SATA Rev 3.0
Standards:
Supported: 9 8 7 6 5
Likely used: 9
Configuration:
Logical max current
cylinders 16383 16383
heads 16 16
sectors/track 63 63
--
CHS current addressable sectors: 16514064
LBA user addressable sectors: 268435455
LBA48 user addressable sectors: 2000409264
Logical Sector size: 512 bytes
Physical Sector size: 512 bytes
Logical Sector-0 offset: 0 bytes
device size with M = 1024*1024: 976762 MBytes
device size with M = 1000*1000: 1024209 MBytes (1024 GB)
cache/buffer size = unknown
Form Factor: 2.5 inch
Nominal Media Rotation Rate: Solid State Device
Capabilities:
LBA, IORDY(can be disabled)
Queue depth: 32
Standby timer values: spec'd by Standard, no device specific minimum
R/W multiple sector transfer: Max = 1 Current = 1
DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 *udma6
Cycle time: min=120ns recommended=120ns
PIO: pio0 pio1 pio2 pio3 pio4
Cycle time: no flow control=120ns IORDY flow control=120ns
Commands/features:
Enabled Supported:
* SMART feature set
Security Mode feature set
* Power Management feature set
* Write cache
* Look-ahead
* WRITE_BUFFER command
* READ_BUFFER command
* NOP cmd
* DOWNLOAD_MICROCODE
* 48-bit Address feature set
* Mandatory FLUSH_CACHE
* FLUSH_CACHE_EXT
* SMART error logging
* General Purpose Logging feature set
* WRITE_{DMA|MULTIPLE}_FUA_EXT
* 64-bit World wide name
* WRITE_UNCORRECTABLE_EXT command
* {READ,WRITE}_DMA_EXT_GPL commands
* Segmented DOWNLOAD_MICROCODE
* unknown 119[6]
unknown 119[8]
unknown 119[9]
* Gen1 signaling speed (1.5Gb/s)
* Gen2 signaling speed (3.0Gb/s)
* Gen3 signaling speed (6.0Gb/s)
* Native Command Queueing (NCQ)
* Host-initiated interface power management
* Phy event counters
* Host automatic Partial to Slumber transitions
* Device automatic Partial to Slumber transitions
* READ_LOG_DMA_EXT equivalent to READ_LOG_EXT
* DMA Setup Auto-Activate optimization
* Device-initiated interface power management
* Software settings preservation
Device Sleep (DEVSLP)
* SMART Command Transport (SCT) feature set
* SCT Features Control (AC4)
* SCT Data Tables (AC5)
* SANITIZE_ANTIFREEZE_LOCK_EXT command
* SANITIZE feature set
* BLOCK_ERASE_EXT command
* Data Set Management TRIM supported (limit 8 blocks)
Security:
Master password revision code = 65534
supported
not enabled
not locked
frozen
not expired: security count
supported: enhanced erase
2min for SECURITY ERASE UNIT. 2min for ENHANCED SECURITY ERASE UNIT.
Logical Unit WWN Device Identifier: 5000000000000000
NAA : 5
IEEE OUI : 000000
Unique ID : 000000000
Device Sleep:
DEVSLP Exit Timeout (DETO): 240 ms (drive)
Minimum DEVSLP Assertion Time (MDAT): 31 ms (drive)
Checksum: correct
查看 NVME SSD
NVME 协议的磁盘则需要借助 nvme
来进行信息查看,默认系统中不会包含这个软件包,需要先进行安装:
apt install nvme-cli
安装完毕之后,使用 list
命令查看 nvme 设备的名称:
nvme list
Node SN Model Namespace Usage Format FW Rev
---------------- -------------------- ---------------------------------------- --------- -------------------------- ---------------- --------
/dev/nvme0n1 ZTA11T0JA2[masked]EK ZHITAI PC005 Active 1TB 1 92.90 GB / 1.02 TB 512 B 0 B ZT00D212
然后使用 smart-log
命令查看硬盘详细信息:
nvme smart-log /dev/nvme0n1
Smart Log for NVME device:nvme0n1 namespace-id:ffffffff
critical_warning : 0
temperature : 29 C
available_spare : 100%
available_spare_threshold : 10%
percentage_used : 0%
data_units_read : 83,560
data_units_written : 221,989
host_read_commands : 362,055
host_write_commands : 810,814
controller_busy_time : 8
power_cycles : 9
power_on_hours : 25
unsafe_shutdowns : 4
media_errors : 0
num_err_log_entries : 0
Warning Temperature Time : 0
Critical Composite Temperature Time : 0
Temperature Sensor 1 : 29 C
Temperature Sensor 2 : 26 C
Temperature Sensor 3 : 30 C
Thermal Management T1 Trans Count : 0
Thermal Management T2 Trans Count : 0
Thermal Management T1 Total Time : 0
Thermal Management T2 Total Time : 0
查看内存基础信息
查看内存基础信息需要借助 dmidecode
软件的能力,执行 dmidecode -t memory
后,便能看到内存的主要信息了:
dmidecode -t memory
# dmidecode 3.2
Getting SMBIOS data from sysfs.
SMBIOS 3.2.1 present.
# SMBIOS implementations newer than version 3.2.0 are not
# fully supported by this version of dmidecode.
Handle 0x003B, DMI type 16, 23 bytes
Physical Memory Array
Location: System Board Or Motherboard
Use: System Memory
Error Correction Type: None
Maximum Capacity: 32 GB
Error Information Handle: Not Provided
Number Of Devices: 2
Handle 0x003C, DMI type 17, 40 bytes
Memory Device
Array Handle: 0x003B
Error Information Handle: Not Provided
Total Width: 64 bits
Data Width: 64 bits
...
虽然信息很长,但是我们需要关注的数值其实很少,只需要知道下面两项就好:
代码语言:txt复制Type: DDR4
Speed: 2667 MT/s
Configured Memory Speed: 2400 MT/s
Speed 代表了设备本身的速度,而 Configured Memory Speed 则代表了系统实际使用的速度。
使用 hdparm 进行基础读测试
使用 hdparm
测试硬盘缓内缓外速度很容易,但是要注意,如果要测试硬件实打实的性能,需要添加 --direct
参数,激活 O_DIRECT
标志,不走页面缓存。
使用 hdparm 测试 NVME 磁盘读取速度
使用 hdparm
分别对 NVME 盘进行 三次读测试:
hdparm -Tt --direct /dev/nvme0n1
测试结果如下:
代码语言:txt复制/dev/nvme0n1:
Timing O_DIRECT cached reads: 4130 MB in 2.00 seconds = 2067.10 MB/sec
HDIO_DRIVE_CMD(identify) failed: Inappropriate ioctl for device
Timing O_DIRECT disk reads: 5142 MB in 3.00 seconds = 1713.82 MB/sec
/dev/nvme0n1:
Timing O_DIRECT cached reads: 2974 MB in 1.99 seconds = 1490.91 MB/sec
HDIO_DRIVE_CMD(identify) failed: Inappropriate ioctl for device
Timing O_DIRECT disk reads: 5574 MB in 3.00 seconds = 1857.85 MB/sec
/dev/nvme0n1:
Timing O_DIRECT cached reads: 2976 MB in 2.00 seconds = 1491.63 MB/sec
HDIO_DRIVE_CMD(identify) failed: Inappropriate ioctl for device
Timing O_DIRECT disk reads: 5074 MB in 3.00 seconds = 1690.87 MB/sec
可以看到文件系统读取速度基本保持在 1G 以上,磁盘数据读速度在 1.5G 左右。
使用 hdparm 测试 SATA 磁盘读取速度
同样使用 hdparm
对 SATA 磁盘进行测试,和之前的命令类似,只需要修改目标设备名称即可:
hdparm -Tt --direct /dev/sda
测试结果也类似:
代码语言:txt复制 Timing O_DIRECT cached reads: 710 MB in 2.00 seconds = 355.07 MB/sec
Timing O_DIRECT disk reads: 1278 MB in 3.00 seconds = 425.43 MB/sec
/dev/sda:
Timing O_DIRECT cached reads: 710 MB in 2.00 seconds = 354.84 MB/sec
Timing O_DIRECT disk reads: 1278 MB in 3.00 seconds = 425.49 MB/sec
/dev/sda:
Timing O_DIRECT cached reads: 710 MB in 2.00 seconds = 355.41 MB/sec
Timing O_DIRECT disk reads: 1278 MB in 3.00 seconds = 425.66 MB/sec
不过因为系统总线限制,可以看到文件系统读取速度在 350 M/s 左右徘徊,而直接读取设备则稳定在 425 M/s。
使用 dd 进行顺序文件写测试
先聊聊写文件。
日常开发最常见的需求之一便是写文件,尤其是写大文件。所以测试连续读写性能非常重要。为了方便测试,这里可以写一个简单的脚本,分别测试不同场景下的写性能。
代码语言:txt复制echo "test 10G"
time sh -c "dd if=/dev/zero of=./ddfile bs=1M count=10240 oflag=direct && sync";
rm ./ddfile;
echo "test 1G/1M"
time sh -c "dd if=/dev/zero of=./ddfile bs=1M count=1024 oflag=direct && sync";
rm ./ddfile;
echo "test 1G/512K"
time sh -c "dd if=/dev/zero of=./ddfile bs=512k count=2048 oflag=direct && sync";
rm ./ddfile;
echo "test 1G/256K"
time sh -c "dd if=/dev/zero of=./ddfile bs=256k count=4096 oflag=direct && sync";
rm ./ddfile;
echo "test 1G/256K"
time sh -c "dd if=/dev/zero of=./ddfile bs=256k count=8192 oflag=direct && sync";
rm ./ddfile;
echo "test 1G/4K"
time sh -c "dd if=/dev/zero of=./ddfile bs=4k count=262144 oflag=direct && sync";
rm ./ddfile;
在 NVME 磁盘上进行读取测试
在 NVME 磁盘上执行后,你会看到结果会类似这样:
代码语言:txt复制test 10G
10240 0 records in
10240 0 records out
10737418240 bytes (11 GB, 10 GiB) copied, 10.9002 s, 985 MB/s
real 0m10.933s
user 0m0.070s
sys 0m4.581s
test 1G/1M
1024 0 records in
1024 0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.50586 s, 713 MB/s
real 0m1.554s
user 0m0.013s
sys 0m0.498s
test 1G/512K
2048 0 records in
2048 0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.54981 s, 693 MB/s
real 0m1.593s
user 0m0.008s
sys 0m0.572s
test 1G/256K
4096 0 records in
4096 0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.64218 s, 654 MB/s
real 0m1.684s
user 0m0.020s
sys 0m0.660s
test 1G/256K
8192 0 records in
8192 0 records out
2147483648 bytes (2.1 GB, 2.0 GiB) copied, 3.23154 s, 665 MB/s
real 0m3.272s
user 0m0.044s
sys 0m1.261s
test 1G/4K
262144 0 records in
262144 0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 4.76955 s, 225 MB/s
real 0m4.808s
user 0m0.154s
sys 0m2.056s
除了4K 随文件读写外,大于 256K 的测试性能表现都在 650 M/s 以上,1M 的文件读写的速度甚至能够持续在 900 M/s。
在 SATA3 磁盘上进行测试
代码语言:txt复制test 10G
10240 0 records in
10240 0 records out
10737418240 bytes (11 GB, 10 GiB) copied, 30.3082 s, 354 MB/s
real 0m30.372s
user 0m0.136s
sys 0m7.296s
test 1G/1M
1024 0 records in
1024 0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 3.15482 s, 340 MB/s
real 0m3.174s
user 0m0.013s
sys 0m0.816s
test 1G/512K
2048 0 records in
2048 0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 3.32807 s, 323 MB/s
real 0m3.346s
user 0m0.043s
sys 0m0.871s
test 1G/256K
4096 0 records in
4096 0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 3.49806 s, 307 MB/s
real 0m3.512s
user 0m0.039s
sys 0m0.946s
test 1G/256K
8192 0 records in
8192 0 records out
2147483648 bytes (2.1 GB, 2.0 GiB) copied, 6.95447 s, 309 MB/s
real 0m7.021s
user 0m0.060s
sys 0m1.886s
test 1G/4K
262144 0 records in
262144 0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 7.54626 s, 142 MB/s
real 0m7.610s
user 0m0.228s
sys 0m2.311s
SATA 盘写入保持在 300M/s 以上,4K写入也有 140 MB/s,对于 S ATA 盘来说,真的可以了。
使用 dd 进行顺序文件读取测试
和写入一样,为了省事,我们同样可以准备一个脚本。但是和写测试相反,我们需要先准备一个比较大的文件作为读取数据。
代码语言:txt复制dd if=/dev/zero of=./ddfile bs=1M count=10240 oflag=direct && sync
time sh -c "dd if=./ddfile of=/dev/shm/ddfile bs=1M count=10240";
rm ./ddfile;
在 NVME 磁盘上进行读取测试
这里我们仅做一种即可,因为读的下限远远比写高的多:
代码语言:txt复制10240 0 records in
10240 0 records out
10737418240 bytes (11 GB, 10 GiB) copied, 8.71309 s, 1.2 GB/s
real 0m9.332s
user 0m0.021s
sys 0m5.544s
10240 0 records in
10240 0 records out
10737418240 bytes (11 GB, 10 GiB) copied, 3.23808 s, 3.3 GB/s
real 0m3.843s
user 0m0.012s
sys 0m3.826s
10240 0 records in
10240 0 records out
10737418240 bytes (11 GB, 10 GiB) copied, 2.98388 s, 3.6 GB/s
real 0m3.599s
user 0m0.010s
sys 0m3.585s
在 SATA 磁盘上进行读取测试
对于 SATA 盘也是一样,相同的命令执行三次:
代码语言:txt复制10240 0 records in
10240 0 records out
10737418240 bytes (11 GB, 10 GiB) copied, 22.571 s, 476 MB/s
real 0m23.195s
user 0m0.086s
sys 0m13.272s
10240 0 records in
10240 0 records out
10737418240 bytes (11 GB, 10 GiB) copied, 3.27516 s, 3.3 GB/s
real 0m3.877s
user 0m0.004s
sys 0m3.869s
10240 0 records in
10240 0 records out
10737418240 bytes (11 GB, 10 GiB) copied, 3.00661 s, 3.6 GB/s
real 0m3.619s
user 0m0.006s
sys 0m3.608s
内存读写测试
内存读写和写测试一样,跑个开心就好,因为下限远比硬盘要来的高,一般情况都能满足需求,如果满足不了,那么就涉及到改软件架构了:
先跑一个简单的写测试:
代码语言:txt复制time sh -c "dd if=/dev/zero of=/dev/shm/fill bs=1M count=10240"
10240 0 records in
10240 0 records out
10737418240 bytes (11 GB, 10 GiB) copied, 2.44928 s, 4.4 GB/s
real 0m2.456s
user 0m0.015s
sys 0m2.441s
再跑一个读测试:
代码语言:txt复制time sh -c "dd if=/dev/shm/fill of=/dev/null bs=1M count=10240"
10240 0 records in
10240 0 records out
10737418240 bytes (11 GB, 10 GiB) copied, 1.15888 s, 9.3 GB/s
real 0m1.164s
user 0m0.011s
sys 0m1.154s
关于光威两种不同类型内存是否能混用
有群友说,两条相同品牌相同容量的光威内存搭配使用会出现问题。我这里因为没有相同规格的内存,所以没有遇到这个问题:
- 实测光威奕Pro和光威战将、以及和雷克沙混搭组合使用没有问题。
奕 Pro 和战将组合插上,使用 dmidecode
筛选查看内存信息命令:
dmidecode -t memory
完整信息如下:
代码语言:txt复制# dmidecode 3.2
Getting SMBIOS data from sysfs.
SMBIOS 3.2.1 present.
# SMBIOS implementations newer than version 3.2.0 are not
# fully supported by this version of dmidecode.
Handle 0x003B, DMI type 16, 23 bytes
Physical Memory Array
Location: System Board Or Motherboard
Use: System Memory
Error Correction Type: None
Maximum Capacity: 32 GB
Error Information Handle: Not Provided
Number Of Devices: 2
Handle 0x003C, DMI type 17, 40 bytes
Memory Device
Array Handle: 0x003B
Error Information Handle: Not Provided
Total Width: 64 bits
Data Width: 64 bits
Size: 16384 MB
Form Factor: SODIMM
Set: None
Locator: SODIMM1
Bank Locator: Memory Channel A
Type: DDR4
Type Detail: Synchronous
Speed: 2667 MT/s
Manufacturer: 0813
Serial Number: D330011A
Asset Tag: 9876543210
Part Number: WAR4S2666[masked]C
Rank: 1
Configured Memory Speed: 2400 MT/s
Minimum Voltage: 1.2 V
Maximum Voltage: 1.2 V
Configured Voltage: 1.2 V
Handle 0x003D, DMI type 17, 40 bytes
Memory Device
Array Handle: 0x003B
Error Information Handle: Not Provided
Total Width: 64 bits
Data Width: 64 bits
Size: 8192 MB
Form Factor: SODIMM
Set: None
Locator: SODIMM2
Bank Locator: Memory Channel B
Type: DDR4
Type Detail: Synchronous
Speed: 2667 MT/s
Manufacturer: 0813
Serial Number: 96C374D7
Asset Tag: 9876543210
Part Number: YCT4S2666[masked]C
Rank: 1
Configured Memory Speed: 2400 MT/s
Minimum Voltage: 1.2 V
Maximum Voltage: 1.2 V
Configured Voltage: 1.2 V
对两条光威内存进行写测试,性能结果如下:
代码语言:txt复制free -g
total used free shared buff/cache available
Mem: 23 0 21 0 1 22
Swap: 1 0 1
dd if=/dev/zero of=/dev/shm/testfile bs=1M count=1000
1000 0 records in
1000 0 records out
1048576000 bytes (1.0 GB, 1000 MiB) copied, 0.249176 s, 4.2 GB/s
dd if=/dev/zero of=/dev/shm/testfile bs=1M count=10000
10000 0 records in
10000 0 records out
10485760000 bytes (10 GB, 9.8 GiB) copied, 2.56671 s, 4.1 GB/s
dd if=/dev/zero of=/dev/shm/testfile bs=10M count=1000
1000 0 records in
1000 0 records out
10485760000 bytes (10 GB, 9.8 GiB) copied, 3.32415 s, 3.2 GB/s
dd if=/dev/zero of=/dev/shm/testfile bs=100M count=100
100 0 records in
100 0 records out
10485760000 bytes (10 GB, 9.8 GiB) copied, 3.76005 s, 2.8 GB/s
dd if=/dev/zero of=/dev/shm/testfile bs=1000M count=10
10 0 records in
10 0 records out
10485760000 bytes (10 GB, 9.8 GiB) copied, 4.01882 s, 2.6 GB/s
最后
简单说下结论,以上测试都是在模拟密集计算场景、高负载的使用的情况下,我认为作为本地低延时的开发补充,NUC也好,国货内存、SSD也罢都能够满足我们对于 IO 的需求,在它的价格档位,可以谈得上是“真香”。
如果你是作为一般日常使用,那么可以忽略上面的数值,因为这些数值对你来说是高不可攀的天花板。你只需要根据自己的预算情况,选择性入手即可。
如果你不是某些品牌的簇拥的话,可以考虑适当给国产以机会。
--EOF
我现在有一个小小的折腾群,里面聚集了一些喜欢折腾的小伙伴。
在不发广告的情况下,我们在里面会一起聊聊软件、HomeLab、编程上的一些问题,也会在群里不定期的分享一些技术沙龙的资料。
喜欢折腾的小伙伴欢迎扫码添加好友。(请注明来源和目的、实名入群,否则不会通过审核)
关于折腾群入群的那些事
本文使用「署名 4.0 国际 (CC BY 4.0)」许可协议,欢迎转载、或重新修改使用,但需要注明来源。 署名 4.0 国际 (CC BY 4.0)
本文作者: 苏洋
创建时间: 2021年02月02日
统计字数: 12629字
阅读时间: 26分钟阅读
本文链接: https://soulteary.com/2021/02/02/nuc-notes-storage-ability-test.html