Hi!
I bought a rather cheap but not so cheap 1TB SSD drive advertised at rw speeds of 450MB/s. I have 2 thinkpads, one X1 3rd gen and one X1 6th gen. Now here is what happened: 1. On 3rd gen laptop with F32 I tested (and made a full backup) : 45MB/s (NTFS) 2. I installed a the "brand new" F33, restored my backup and...: 85MB/s (NTFS) 3. Since I was a tad late updating I told myself let's do it on the 6th Gen too. Did a full backup on F32 at : ~400MB/s!!! (NTFS and ext4 - did several backups) 4. I installed a fresh new F33 and the restore speed was then... 50MB/s :-( (ext4) 5. I just restested the disk on the 3rd gen laptop and I am now (ext4) at 45MB/s (I also just upgrade my BIOS to its latest version for this test).
So I was wondering if there was any logical explanations and things I could do to ensure I use the disk at its maximum speed.
Thank you.
Fred
On 3/1/21 10:37 PM, Frederic Muller wrote:
I bought a rather cheap but not so cheap 1TB SSD drive advertised at rw speeds of 450MB/s. I have 2 thinkpads, one X1 3rd gen and one X1 6th gen. Now here is what happened:
- On 3rd gen laptop with F32 I tested (and made a full backup) : 45MB/s
(NTFS) 2. I installed a the "brand new" F33, restored my backup and...: 85MB/s (NTFS) 3. Since I was a tad late updating I told myself let's do it on the 6th Gen too. Did a full backup on F32 at : ~400MB/s!!! (NTFS and ext4 - did several backups) 4. I installed a fresh new F33 and the restore speed was then... 50MB/s :-( (ext4) 5. I just restested the disk on the 3rd gen laptop and I am now (ext4) at 45MB/s (I also just upgrade my BIOS to its latest version for this test).
So I was wondering if there was any logical explanations and things I could do to ensure I use the disk at its maximum speed.
There are a lot of factors and you've left out some details. Is this internal or external. If external, then USB2 or USB3 (or something else)? What are you using for the backup and restore? Even though there is almost no seek delay, reading and writing lots of random files involved a lot of IO operations. If you really want to test the speed, you should use something like dd using a big blocksize.
dd if=/dev/zero of=/test.dat bs=8M count=1K dd if=/test.dat of=/dev/null bs=8M
On 3/2/21 2:15 PM, Samuel Sieb wrote:
On 3/1/21 10:37 PM, Frederic Muller wrote:
I bought a rather cheap but not so cheap 1TB SSD drive advertised at rw speeds of 450MB/s. I have 2 thinkpads, one X1 3rd gen and one X1 6th gen. Now here is what happened:
- On 3rd gen laptop with F32 I tested (and made a full backup) :
45MB/s (NTFS) 2. I installed a the "brand new" F33, restored my backup and...: 85MB/s (NTFS) 3. Since I was a tad late updating I told myself let's do it on the 6th Gen too. Did a full backup on F32 at : ~400MB/s!!! (NTFS and ext4
- did several backups)
- I installed a fresh new F33 and the restore speed was then...
50MB/s :-( (ext4) 5. I just restested the disk on the 3rd gen laptop and I am now (ext4) at 45MB/s (I also just upgrade my BIOS to its latest version for this test).
So I was wondering if there was any logical explanations and things I could do to ensure I use the disk at its maximum speed.
There are a lot of factors and you've left out some details. Is this internal or external. If external, then USB2 or USB3 (or something else)?
As the title says it's indeed USB. Both machines have USB 3 ports, which I used without any hub or middle man.
What are you using for the backup and restore? Even though there is almost no seek delay, reading and writing lots of random files involved a lot of IO operations. If you really want to test the speed, you should use something like dd using a big blocksize.
dd if=/dev/zero of=/test.dat bs=8M count=1K dd if=/test.dat of=/dev/null bs=8M
I used mainly the cp command line together with progress to get an idea of the execution status. I also used the benchmark from Disk under GNOME. The results are (were?) very consistent at each described phase.
And while I was not really testing per se (since I was ok with the initial 45MB/s, I really noticed when it went to 80+ and then 400+ ... and then down to 45.
Anyway since I am somehow back to initial lousy speeds and saw it way faster on the same machines, is there a "reliable" way to go back to more decent speed then?
Thank you.
Fred
users mailing list -- users@lists.fedoraproject.org To unsubscribe send an email to users-leave@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org Do not reply to spam on the list, report it: https://pagure.io/fedora-infrastructure
On 3/1/21 11:34 PM, Frederic Muller wrote:
On 3/2/21 2:15 PM, Samuel Sieb wrote:
On 3/1/21 10:37 PM, Frederic Muller wrote:
I bought a rather cheap but not so cheap 1TB SSD drive advertised at rw speeds of 450MB/s. I have 2 thinkpads, one X1 3rd gen and one X1 6th gen. Now here is what happened:
- On 3rd gen laptop with F32 I tested (and made a full backup) :
45MB/s (NTFS) 2. I installed a the "brand new" F33, restored my backup and...: 85MB/s (NTFS) 3. Since I was a tad late updating I told myself let's do it on the 6th Gen too. Did a full backup on F32 at : ~400MB/s!!! (NTFS and ext4
- did several backups)
- I installed a fresh new F33 and the restore speed was then...
50MB/s :-( (ext4) 5. I just restested the disk on the 3rd gen laptop and I am now (ext4) at 45MB/s (I also just upgrade my BIOS to its latest version for this test).
So I was wondering if there was any logical explanations and things I could do to ensure I use the disk at its maximum speed.
There are a lot of factors and you've left out some details. Is this internal or external. If external, then USB2 or USB3 (or something else)?
As the title says it's indeed USB. Both machines have USB 3 ports, which I used without any hub or middle man.
Ugh, sorry, don't know how I missed that. Are you sure that you plugged it into a USB3 port? Some laptops have both. Also, check the log to make sure it is being recognized as a full speed USB3 device.
And while I was not really testing per se (since I was ok with the initial 45MB/s, I really noticed when it went to 80+ and then 400+ ... and then down to 45.
Anyway since I am somehow back to initial lousy speeds and saw it way faster on the same machines, is there a "reliable" way to go back to more decent speed then?
I expect it is something to do with the port, especially since it was faster on the one laptop. I haven't had speed issues like that when using USB3 ports with USB3 external drives.
On 3/2/21 8:34 AM, Frederic Muller wrote:
Anyway since I am somehow back to initial lousy speeds and saw it way faster on the same machines, is there a "reliable" way to go back to more decent speed then?
Before putting filesystems in the middle:
hdparm -t /dev/xxxxxxx
to test sequential read speed.
On 3/2/21 4:23 PM, Roberto Ragusa wrote:
On 3/2/21 8:34 AM, Frederic Muller wrote:
Anyway since I am somehow back to initial lousy speeds and saw it way faster on the same machines, is there a "reliable" way to go back to more decent speed then?
Before putting filesystems in the middle:
hdparm -t /dev/xxxxxxx
to test sequential read speed.
I think there is a misunderstanding about my question: somehow I cannot get reliable read speed as depending on the machine (not sure) or/and the OS (and its patches) used I get very very different real usage speed on 2 different machines (we're talking hour+ to minutes differences to copy those files).
So the same machine can get very very very different real life speed (not testing) in "simply" copying files depending on the version of Fedora used (with its patches or not and....what else).
So now how can configure Fedora 33 to get the optimum speed is the question. Are there BIOS or system settings that make it happen? The X1 3rd gen now has a new bios, not the 6th gen. The 6th gen went from 400MB/s down to 40MB/s switching from F32 to F33, the 3rd gen went from 45 to 85 switching from F32 to F33. I am just trying to understand what is giving those very different speeds. And yes it is really weird.
Now those are 2 "work" machines and I can't really reinstall F32 (or F26-7 and upgrade little by little to F32) to see what happens this time (not that it's even possible). So if you're as puzzled as me, I would definitely understand. What speed do you get with your external USB SSD drives for example, and is it consistent over different setups?
Well.. not a big deal, everything else (not my mouse ;-) ) is working (oh, and not GnunPGP either).
So thank you for your potential tips and if none... well I'll live with a slow SSD for now :D
Thank you.
Fred
One complication. You said you aren't using a HUB, that is probably not exactly true. Almost all of the USB ports on a laptop or a desktop are using a HUB, very few of the ports are dedicated. Typically a laptop or desktop has at most 3 actual real underlying ports, and some do not even have that many.
Because of this there could be some other devices on the HUB effecting it.
Also, the copy speed once you get above 50MB/sec can be affected by the device you are copying from. Also also on the device you are copying if the file sizes are different then you cannot compare MB/sec since on small files (say 64k) the overhead for each new file is significant.
Make sure to test single big files if you want to determine speed, and do not call the test finished until a sync returns as with smaller tests there could be a significant amount of data in the host os cache causing the rate to look much better than it actually is. Checking grep -i Dirty /proc/meminfo will tell you how much is in said cache at the end of the test.
On Tue, Mar 2, 2021 at 5:47 AM Frederic Muller fred@cm17.com wrote:
On 3/2/21 4:23 PM, Roberto Ragusa wrote:
On 3/2/21 8:34 AM, Frederic Muller wrote:
Anyway since I am somehow back to initial lousy speeds and saw it way faster on the same machines, is there a "reliable" way to go back to more decent speed then?
Before putting filesystems in the middle:
hdparm -t /dev/xxxxxxx
to test sequential read speed.
I think there is a misunderstanding about my question: somehow I cannot get reliable read speed as depending on the machine (not sure) or/and the OS (and its patches) used I get very very different real usage speed on 2 different machines (we're talking hour+ to minutes differences to copy those files).
So the same machine can get very very very different real life speed (not testing) in "simply" copying files depending on the version of Fedora used (with its patches or not and....what else).
So now how can configure Fedora 33 to get the optimum speed is the question. Are there BIOS or system settings that make it happen? The X1 3rd gen now has a new bios, not the 6th gen. The 6th gen went from 400MB/s down to 40MB/s switching from F32 to F33, the 3rd gen went from 45 to 85 switching from F32 to F33. I am just trying to understand what is giving those very different speeds. And yes it is really weird.
Now those are 2 "work" machines and I can't really reinstall F32 (or F26-7 and upgrade little by little to F32) to see what happens this time (not that it's even possible). So if you're as puzzled as me, I would definitely understand. What speed do you get with your external USB SSD drives for example, and is it consistent over different setups?
Well.. not a big deal, everything else (not my mouse ;-) ) is working (oh, and not GnunPGP either).
So thank you for your potential tips and if none... well I'll live with a slow SSD for now :D
Thank you.
Fred _______________________________________________ users mailing list -- users@lists.fedoraproject.org To unsubscribe send an email to users-leave@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org Do not reply to spam on the list, report it: https://pagure.io/fedora-infrastructure
On 3/2/21 7:20 PM, Roger Heflin wrote:
One complication. You said you aren't using a HUB, that is probably not exactly true. Almost all of the USB ports on a laptop or a desktop are using a HUB, very few of the ports are dedicated. Typically a laptop or desktop has at most 3 actual real underlying ports, and some do not even have that many.
ok sorry for the unclear description: I am not using any external HUB. The latop has 2 USB ports (3rd gen) and 3 for the 6th gen.
Because of this there could be some other devices on the HUB effecting it.
Also, the copy speed once you get above 50MB/sec can be affected by the device you are copying from. Also also on the device you are copying if the file sizes are different then you cannot compare MB/sec since on small files (say 64k) the overhead for each new file is significant.
In every instance I was making a backup of the same files (not the same on both machines - but then again you can see the performance do not seem to be machine related) stored on the internal SSD disk. Roughly 3-400GB I'd say for both machines.
Make sure to test single big files if you want to determine speed, and do not call the test finished until a sync returns as with smaller tests there could be a significant amount of data in the host os cache causing the rate to look much better than it actually is. Checking grep -i Dirty /proc/meminfo will tell you how much is in said cache at the end of the test.
What is puzzling (to me) is the performance disparity between the 2 machines AND the version of Fedora, especially since I cannot even conclude Fxx is faster/better than Fyy.
Thank you.
Fred
On 2021-03-02 1:37 a.m., Frederic Muller wrote:
So I was wondering if there was any logical explanations and things I could do to ensure I use the disk at its maximum speed.
You are using Fedora, so you have an already-installed graphical utility to run a quick benchmark test on the disk. You can find the disks utility (usually in the utilities icon), select the appropriate disk on the left, and click on the "..." on the top right. You will see a benchmark selection. Start that up and it will run for about a minute, checking random and sequential read and write times, and graphing it nicely for you. Since the disk is on a USB port, you need to be careful about getting the rest of the machine as idle as possible, as almost all USB implementations are CPU-bound.
If you're not getting the specified speeds, then it could be due to a number of factors that are probably unrelated to the disk itself. However, one thing will affect your write speeds: most SSDs have a cache system onboard to mitigate slow write speeds. Writing a much larger file than the RAM buffer in the disk will result in stalling that looks like what you are describing. That's not unexpected, but it does lead to the next problem - what happens if power is lost while writing. Is the drive's onboard cache flushed, or is it lost due to an onboard lack of large power capacitors as a cost reduction? If the latter is true, then this is a point where you're going to appreciate that your use of BTRFS or ZFS will at least prevent disk corruption in this situation. Note that the drive cost is not related to the answer, as e.g. expensive Samsung EVO drives are known to have this issue.
--
John Mellor
On 3/2/21 12:47 PM, Frederic Muller wrote:
On 3/2/21 4:23 PM, Roberto Ragusa wrote:
On 3/2/21 8:34 AM, Frederic Muller wrote:
Anyway since I am somehow back to initial lousy speeds and saw it way faster on the same machines, is there a "reliable" way to go back to more decent speed then?
Before putting filesystems in the middle:
hdparm -t /dev/xxxxxxx
to test sequential read speed.
I think there is a misunderstanding about my question: somehow I cannot get reliable read speed as depending on the machine (not sure) or/and the OS (and its patches) used I get very very different real usage speed on 2 different machines (we're talking hour+ to minutes differences to copy those files).
You just want the result and it is not sure what you are measuring and in which conditions.
That's why you should try things in order: - lsusb -t (guess you may be falling back from USB3 to USB2 for some reason, since 40-50MB/s is typical of USB2) - hdparm -t /dev/xxxxx to see the actual reading speed of the raw device as configured in that moment - then add the filesystem in the middle (caching + possible issues with ntfs that may be fuse, who knows) - then reach the "I'm just copying" files, where we do not know what are you reading from and if it is already cached or not
Regards and good luck.
On 3/2/21 9:58 PM, Roberto Ragusa wrote:
On 3/2/21 12:47 PM, Frederic Muller wrote:
On 3/2/21 4:23 PM, Roberto Ragusa wrote:
On 3/2/21 8:34 AM, Frederic Muller wrote:
Anyway since I am somehow back to initial lousy speeds and saw it way faster on the same machines, is there a "reliable" way to go back to more decent speed then?
Before putting filesystems in the middle:
hdparm -t /dev/xxxxxxx
to test sequential read speed.
I think there is a misunderstanding about my question: somehow I cannot get reliable read speed as depending on the machine (not sure) or/and the OS (and its patches) used I get very very different real usage speed on 2 different machines (we're talking hour+ to minutes differences to copy those files).
You just want the result and it is not sure what you are measuring and in which conditions.
That's why you should try things in order:
- lsusb -t (guess you may be falling back from USB3 to USB2 for some
reason, since 40-50MB/s is typical of USB2)
- hdparm -t /dev/xxxxx to see the actual reading speed of the raw
device as configured in that moment
- then add the filesystem in the middle (caching + possible issues
with ntfs that may be fuse, who knows)
- then reach the "I'm just copying" files, where we do not know what
are you reading from and if it is already cached or not
Regards and good luck.
point taken. I'll try to have a more technical and thourough approach and see if I can sort the issue out then.
Thank you for your patience.
Fred
On 3/2/21 3:47 AM, Frederic Muller wrote:
So now how can configure Fedora 33 to get the optimum speed is the question. Are there BIOS or system settings that make it happen? The X1 3rd gen now has a new bios, not the 6th gen. The 6th gen went from 400MB/s down to 40MB/s switching from F32 to F33, the 3rd gen went from 45 to 85 switching from F32 to F33. I am just trying to understand what is giving those very different speeds. And yes it is really weird.
Your original post was rather unclear on this part. I mostly couldn't tell which computer and OS combination was getting which speed. A table or something like that would have helped.
As Roberto suggested, you should use "lsusb -t". Here's the output from a laptop with a USB3 SSD attached: $ lsusb -t /: Bus 04.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/6p, 5000M |__ Port 6: Dev 2, If 0, Class=Mass Storage, Driver=uas, 5000M /: Bus 03.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/15p, 480M |__ Port 2: Dev 2, If 0, Class=Human Interface Device, Driver=usbhid, 12M |__ Port 5: Dev 3, If 0, Class=Vendor Specific Class, Driver=, 12M |__ Port 6: Dev 28, If 0, Class=Human Interface Device, Driver=usbhid, 1.5M |__ Port 7: Dev 4, If 0, Class=Video, Driver=uvcvideo, 480M |__ Port 7: Dev 4, If 1, Class=Video, Driver=uvcvideo, 480M |__ Port 12: Dev 6, If 1, Class=Wireless, Driver=btusb, 12M |__ Port 12: Dev 6, If 0, Class=Wireless, Driver=btusb, 12M /: Bus 02.Port 1: Dev 1, Class=root_hub, Driver=ehci-pci/3p, 480M |__ Port 1: Dev 2, If 0, Class=Hub, Driver=hub/8p, 480M /: Bus 01.Port 1: Dev 1, Class=root_hub, Driver=ehci-pci/3p, 480M |__ Port 1: Dev 2, If 0, Class=Hub, Driver=hub/6p, 480M
xhci indicates a USB3 interface. The last number on each line shows the speed. You can see the SSD at the top getting an available bandwidth of 5Gb.
I moved it to a different port: /: Bus 03.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/15p, 480M |__ Port 2: Dev 2, If 0, Class=Human Interface Device, Driver=usbhid, 12M |__ Port 5: Dev 3, If 0, Class=Vendor Specific Class, Driver=, 12M |__ Port 6: Dev 29, If 0, Class=Mass Storage, Driver=uas, 480M
It's only getting 480Mb now.
On 3/2/21 9:58 PM, Roberto Ragusa wrote:
You just want the result and it is not sure what you are measuring and in which conditions.
That's why you should try things in order:
- lsusb -t (guess you may be falling back from USB3 to USB2 for some
reason, since 40-50MB/s is typical of USB2)
- hdparm -t /dev/xxxxx to see the actual reading speed of the raw
device as configured in that moment
- then add the filesystem in the middle (caching + possible issues
with ntfs that may be fuse, who knows)
- then reach the "I'm just copying" files, where we do not know what
are you reading from and if it is already cached or not
Regards and good luck.
Hello again!
A bit of spare time and here I come again: I started with lshub -t and am getting this: /: Bus 03.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/4p, 5000M /: Bus 02.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/11p, 480M |__ Port 6: Dev 3, If 0, Class=Vendor Specific Class, Driver=, 12M |__ Port 8: Dev 7, If 0, Class=Video, Driver=uvcvideo, 480M |__ Port 8: Dev 7, If 1, Class=Video, Driver=uvcvideo, 480M /: Bus 01.Port 1: Dev 1, Class=root_hub, Driver=ehci-pci/3p, 480M |__ Port 1: Dev 2, If 0, Class=Hub, Driver=hub/8p, 480M
Now when I plug in a hub on either of the 2 available ports there are connected to Bus 02 Port 1 or Port 2, which states 480M. I've read this is the standard speed for USB 2. Now I checked again the X1 carbon 3rd gen specs and it says it has 2 x USB 3.0 (1 with AOU).
I do see a 5000M port but apparently it is for internal purposes.
So already I am confused. Would you mind explaining me?
Thank you.
Fred
On 3/3/21 2:03 PM, Frederic Muller wrote:
On 3/2/21 9:58 PM, Roberto Ragusa wrote:
You just want the result and it is not sure what you are measuring and in which conditions.
That's why you should try things in order:
- lsusb -t (guess you may be falling back from USB3 to USB2 for some
reason, since 40-50MB/s is typical of USB2)
- hdparm -t /dev/xxxxx to see the actual reading speed of the raw
device as configured in that moment
- then add the filesystem in the middle (caching + possible issues
with ntfs that may be fuse, who knows)
- then reach the "I'm just copying" files, where we do not know what
are you reading from and if it is already cached or not
Regards and good luck.
Hello again!
A bit of spare time and here I come again: I started with lshub -t and am getting this: /: Bus 03.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/4p, 5000M /: Bus 02.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/11p, 480M |__ Port 6: Dev 3, If 0, Class=Vendor Specific Class, Driver=, 12M |__ Port 8: Dev 7, If 0, Class=Video, Driver=uvcvideo, 480M |__ Port 8: Dev 7, If 1, Class=Video, Driver=uvcvideo, 480M /: Bus 01.Port 1: Dev 1, Class=root_hub, Driver=ehci-pci/3p, 480M |__ Port 1: Dev 2, If 0, Class=Hub, Driver=hub/8p, 480M
Now when I plug in a hub on either of the 2 available ports there are connected to Bus 02 Port 1 or Port 2, which states 480M. I've read this is the standard speed for USB 2. Now I checked again the X1 carbon 3rd gen specs and it says it has 2 x USB 3.0 (1 with AOU).
I do see a 5000M port but apparently it is for internal purposes.
So already I am confused. Would you mind explaining me?
A little addition: I plugged-in the disk (again...) and got this this time: /: Bus 03.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/4p, 5000M |__ Port 2: Dev 2, If 0, Class=Mass Storage, Driver=uas, 5000M /: Bus 02.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/11p, 480M |__ Port 1: Dev 27, If 0, Class=Hub, Driver=hub/4p, 480M |__ Port 2: Dev 28, If 0, Class=Human Interface Device, Driver=usbhid, 12M |__ Port 2: Dev 28, If 1, Class=Human Interface Device, Driver=usbhid, 12M |__ Port 3: Dev 31, If 0, Class=Vendor Specific Class, Driver=xpad, 12M |__ Port 4: Dev 30, If 1, Class=Human Interface Device, Driver=usbhid, 12M |__ Port 4: Dev 30, If 0, Class=Human Interface Device, Driver=usbhid, 12M |__ Port 6: Dev 3, If 0, Class=Vendor Specific Class, Driver=, 12M |__ Port 8: Dev 7, If 0, Class=Video, Driver=uvcvideo, 480M |__ Port 8: Dev 7, If 1, Class=Video, Driver=uvcvideo, 480M /: Bus 01.Port 1: Dev 1, Class=root_hub, Driver=ehci-pci/3p, 480M |__ Port 1: Dev 2, If 0, Class=Hub, Driver=hub/8p, 480M
So it seems the disk does get to be connected "automatically" to the USB3.0 port actually.
After that I formatted the disk in NTFS and ext4 and ran the hdparm test. I get a slightly better performance result with NTFS: NTFS: /dev/sdb: Timing buffered disk reads: 1298 MB in 3.00 seconds = 432.05 MB/sec
ext4: /dev/sdb: Timing buffered disk reads: 1244 MB in 3.00 seconds = 414.36 MB/sec
Going through the graphical benchmark I can actually see "why": the read speed is making a straight line at about 450MB/s in NTFS ( https://imagebin.ca/v/5td8fZkQCGY9 ) while it's zigzagging with ext4 ( https://imagebin.ca/v/5td8voL3uiqj )
Any idea why this happens?
Thank you again.
Fred
On 3/2/21 11:27 PM, Frederic Muller wrote:
Going through the graphical benchmark I can actually see "why": the read speed is making a straight line at about 450MB/s in NTFS ( https://imagebin.ca/v/5td8fZkQCGY9 ) while it's zigzagging with ext4 ( https://imagebin.ca/v/5td8voL3uiqj )
Any idea why this happens?
By default, ext4 is formatted with lazy initialization. It's possible that it's affecting the benchmark since I assume you had just formatted it before testing.
On 3/3/21 2:43 PM, Samuel Sieb wrote:
On 3/2/21 11:27 PM, Frederic Muller wrote:
Going through the graphical benchmark I can actually see "why": the read speed is making a straight line at about 450MB/s in NTFS ( https://imagebin.ca/v/5td8fZkQCGY9 ) while it's zigzagging with ext4 ( https://imagebin.ca/v/5td8voL3uiqj )
Any idea why this happens?
By default, ext4 is formatted with lazy initialization. It's possible that it's affecting the benchmark since I assume you had just formatted it before testing.
yes I just had.
It seems that if the drive unmounted is unmounted then the zigzagging disappears. So maybe that's also just that?
Fred
-
Frederic Muller -
John Mellor -
Roberto Ragusa -
Roger Heflin -
Samuel Sieb