Your Ntfs disk is fragmented ?
Are these tests valid ?
Oh, haven't heard that term for many years. I've goodled and found some strong opinions against it: Should I defrag my SSD, The short answer is no. The long answer is absolutely no. ; No, you should never defrag an SSD drive, only the old type HDD's need to be defragmented. An SSD uses a different technology named TRIM and that is managed automatically and so on.
-- I'm not sure what you mean. Or, I don't know. What other ways to run speed tests under Linux? But the very reason I started testing speeds is that NTFS (via ntfs-3g at least, ntfs3 to be investigated next) feels slow under Linux.
NTFS3 Kernel Module (Post by babam #79250)babam wrote: ↑26 Oct 2020, 05:56Pros:
1. Very fast, much faster than ntfs-3g.
2. Prevents fragmentation when writing files, whereas ntfs-3g generates a lot of fragmentation.
3. Low in CPU and RAM usage.
Cons:
1. NTFS3 has not been tested for stability, reliability and security. NTFS-3G has been tested for stability and safety.
Code: Select all
# hdparm -t /dev/sdb3
/dev/sdb3:
Timing buffered disk reads: 1270 MB in 3.00 seconds = 423.26 MB/sec
# hdparm -t /dev/sdb1
/dev/sdb1:
Timing buffered disk reads: 1230 MB in 3.00 seconds = 409.93 MB/sec
fulalas, I think those are essentially the same value plus/minus. I'm suspicious about hdparm -- how does it measure what? It doesn't even check whether those partitions are mounted! I think it just measures the device overall speed class somehow.
Code: Select all
dd if=/dev/zero of=/mnt/sda3/tempfile bs=1M count=1024 conv=fdatasync |& tail -1
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.56873 s, 684 MB/s
dd if=/dev/zero of=/mnt/sda4/tempfile bs=1M count=1024 conv=fdatasync |& tail -1
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 3.36937 s, 319 MB/s
Code: Select all
sudo sh -c "/usr/bin/echo 3 > /proc/sys/vm/drop_caches"
dd if=/mnt/sda3/tempfile of=/dev/null bs=1M count=1024 |& tail -1
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.58736 s, 676 MB/s
dd if=/mnt/sda4/tempfile of=/dev/null bs=1M count=1024 |& tail -1
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 4.73717 s, 227 MB/s
a few people were complaining -- yes, but months ago, well before the new fixes in 6.4 and 6.6 https://lore.kernel.org/ntfs3/. It's uncomfortable to be investigating this at frontier of hopefully first bug-free versions I agree My strategy is to enable it on the disk that's daily plugged into Windows and synched there -- Windows and my synchronization software will hopefully catch any file corruption. I'll post it then here and in NTFS3 Kernel Module
trying now my fast NVMe Samsung SSD through the USB enclosure and the 10Gbps port:
Code: Select all
umount /mnt/sda4
mount -t ntfs3 /dev/sda4
Thanks for that nice utility. I've measured the reading speed of ntfs3, it gives an even higher number 851.8 MB/s (100 samples). That's more than I've ever seen from a USB3.1 10Gbps port.
I also have the KDiskMark-3.1.3-fio-3.34-x86_64.AppImage. I personally don't trust it for its bugs? For example, one has to restart the program afresh when choosing a different disk/partition between tests -- otherwise it's just testing the previous partition again. Anyway here're the results that look about right:
KDiskMark (3.1.3): https://github.com/JonMagon/KDiskMark
Flexible I/O Tester (fio-3.34): https://github.com/axboe/fio
--------------------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes
[Read]
Sequential 1 MiB (Q= 8, T= 1): 865.022 MB/s [ 844.7 IOPS] < 9437.22 us>
Sequential 128 KiB (Q= 32, T= 1): 785.451 MB/s [ 6136.3 IOPS] < 5202.63 us>
Random 4 KiB (Q= 32, T=16): 86.009 MB/s [ 21503.2 IOPS] < 5940.20 us>
Random 4 KiB (Q= 1, T= 1): 23.932 MB/s [ 5983.1 IOPS] < 166.23 us>
[Write]
Sequential 1 MiB (Q= 8, T= 1): 699.244 MB/s [ 682.9 IOPS] < 11394.37 us>
Sequential 128 KiB (Q= 32, T= 1): 656.682 MB/s [ 5130.3 IOPS] < 6211.55 us>
Random 4 KiB (Q= 32, T=16): 85.249 MB/s [ 21313.4 IOPS] < 5990.25 us>
Random 4 KiB (Q= 1, T= 1): 46.921 MB/s [ 11730.4 IOPS] < 83.97 us>
Profile: Default
Test: 1 GiB (x5) [Measure: 5 sec / Interval: 5 sec]
Date: 2023-11-01 21:59:25
OS: slackware 5.01 [linux 6.5.5-porteus]
Code: Select all
root@porteus:~# smartctl -x -d sntrealtek /dev/disk/by-label/PorteuX | grep Temp*
Warning Comp. Temp. Threshold: 85 Celsius
Critical Comp. Temp. Threshold: 85 Celsius
Temperature: 55 Celsius
Warning Comp. Temperature Time: 0
Critical Comp. Temperature Time: 0
Temperature Sensor 1: 55 Celsius
Temperature Sensor 2: 74 Celsius
UPDATE: The comparison is now done in the NTFS3 dedicated thread, NTFS3 Kernel Module (Post by rych #97030) and demonstrates significant speed improvementsKDiskMark (3.1.3): https://github.com/JonMagon/KDiskMark
Flexible I/O Tester (fio-3.34): https://github.com/axboe/fio
--------------------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes
[Read]
Sequential 1 MiB (Q= 8, T= 1): 1037.815 MB/s [ 1013.5 IOPS] < 7808.20 us>
Sequential 128 KiB (Q= 32, T= 1): 839.224 MB/s [ 6556.4 IOPS] < 5015.91 us>
Random 4 KiB (Q= 32, T=16): 19.823 MB/s [ 4956.5 IOPS] < 25854.82 us>
Random 4 KiB (Q= 1, T= 1): 28.057 MB/s [ 7014.4 IOPS] < 141.95 us>
[Write]
Sequential 1 MiB (Q= 8, T= 1): 400.840 MB/s [ 391.4 IOPS] < 8382.00 us>
Sequential 128 KiB (Q= 32, T= 1): 407.736 MB/s [ 3185.4 IOPS] < 3536.56 us>
Random 4 KiB (Q= 32, T=16): 52.634 MB/s [ 13159.5 IOPS] < 9316.66 us>
Random 4 KiB (Q= 1, T= 1): 91.410 MB/s [ 22852.7 IOPS] < 26.89 us>
Profile: Default
Test: 1 GiB (x5) [Measure: 5 sec / Interval: 5 sec]
Date: 2023-11-01 22:25:21
OS: slackware 5.01 [linux 6.5.5-porteus]
(my enclosure controller is ASM236X for some reason) -- of a higher class (20Gbps), but still plugged into a 10Gbps USB 3.1 port:
Code: Select all
root@porteus:~# lsusb -t
/: Bus 02.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/10p, 10000M
|__ Port 2: Dev 3, If 0, Class=, Driver=uas, 10000M
Code: Select all
root@porteus:~# smartctl -x -d sntasmedia /dev/disk/by-label/PorteuX | grep Temperature
Temperature: 56 Celsius
Warning Comp. Temperature Time: 0
Critical Comp. Temperature Time: 0
Temperature Sensor 1: 56 Celsius
Temperature Sensor 2: 60 Celsius
Speed tests on its NTFS, ntfs3 driver on /mnt/sda4 partition:KDiskMark (3.1.3): https://github.com/JonMagon/KDiskMark
Flexible I/O Tester (fio-3.34): https://github.com/axboe/fio
--------------------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes
[Read]
Sequential 1 MiB (Q= 8, T= 1): 1015.669 MB/s [ 991.9 IOPS] < 8034.74 us>
Sequential 128 KiB (Q= 32, T= 1): 909.275 MB/s [ 7103.7 IOPS] < 4494.44 us>
Random 4 KiB (Q= 32, T=16): 96.824 MB/s [ 24207.1 IOPS] < 5282.25 us>
Random 4 KiB (Q= 1, T= 1): 25.087 MB/s [ 6271.7 IOPS] < 158.69 us>
[Write]
Sequential 1 MiB (Q= 8, T= 1): 774.676 MB/s [ 756.5 IOPS] < 10327.58 us>
Sequential 128 KiB (Q= 32, T= 1): 733.480 MB/s [ 5730.3 IOPS] < 5551.47 us>
Random 4 KiB (Q= 32, T=16): 98.027 MB/s [ 24507.9 IOPS] < 5213.59 us>
Random 4 KiB (Q= 1, T= 1): 41.485 MB/s [ 10371.3 IOPS] < 94.15 us>
Profile: Default
Test: 1 GiB (x5) [Measure: 5 sec / Interval: 5 sec]
Date: 2024-01-15 20:00:07
OS: slackware 5.01 [linux 6.6.3-porteus]
KDiskMark (3.1.3): https://github.com/JonMagon/KDiskMark
Flexible I/O Tester (fio-3.34): https://github.com/axboe/fio
--------------------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes
[Read]
Sequential 1 MiB (Q= 8, T= 1): 1017.442 MB/s [ 993.6 IOPS] < 8017.45 us>
Sequential 128 KiB (Q= 32, T= 1): 907.701 MB/s [ 7091.4 IOPS] < 4501.56 us>
Random 4 KiB (Q= 32, T=16): 96.970 MB/s [ 24243.4 IOPS] < 5276.51 us>
Random 4 KiB (Q= 1, T= 1): 25.596 MB/s [ 6399.1 IOPS] < 155.56 us>
[Write]
Sequential 1 MiB (Q= 8, T= 1): 774.772 MB/s [ 756.6 IOPS] < 10300.75 us>
Sequential 128 KiB (Q= 32, T= 1): 734.296 MB/s [ 5736.7 IOPS] < 5555.16 us>
Random 4 KiB (Q= 32, T=16): 97.634 MB/s [ 24409.5 IOPS] < 5231.61 us>
Random 4 KiB (Q= 1, T= 1): 47.243 MB/s [ 11810.9 IOPS] < 83.20 us>
Profile: Default
Test: 1 GiB (x5) [Measure: 5 sec / Interval: 5 sec]
Date: 2024-01-15 20:06:34
OS: slackware 5.01 [linux 6.6.3-porteus]