crucial mx500 TBW not accurate

ciukacz

Member
hello,

i have 2x crucial mx500 ssd (same exact model), but for 1 of them TBW is not accurate.
crucial's own tool (storage executive) shows accurate TBW.
looks like data for smart 0xF6 attribute is read or interpreted incorrectly.

oh and thank you for this wonderfual software :)
regards
 

Attachments

  • HWiNFO64.DBG
    1.1 MB · Views: 2
  • report.HTM
    220.9 KB · Views: 4
  • sensors.png
    sensors.png
    91.3 KB · Views: 11
  • smart drive 0.png
    smart drive 0.png
    180.5 KB · Views: 11
  • smart drive 1.png
    smart drive 1.png
    196.5 KB · Views: 10
This looks like the drive doesn't properly report information via ATA Statistics standard and truncates the number of reported sectors written.
Try to disable "ATA Statistics Support" in HWiNFO if you'll get correct numbers then.
 
Martin said:
This looks like the drive doesn't properly report information via ATA Statistics standard and truncates the number of reported sectors written.
Try to disable "ATA Statistics Support" in HWiNFO if you'll get correct numbers then.

if i disable that both "total host writes" and "total host reads" disappear from sensors window.
why do you think the drive truncates the data when manufacturer's tool can read correct value (screenshots)?
 
The problem I described earlier is true, the difference is that HWiNFO relies primarily on ATA Statistics, which is a standard unlike SMART. And AFAIK all other tools rely on SMART.
I will add a workaround for this issue in the next Beta build.
 
thank you for the explanation.
long shot: is it explained anywhere what is the difference between "smart support" and "ata statistics support" in hwinfo's behavior?
i assume ata statistics take precedence if a counter is supported by both APIs?
 
Yes, ATA Statistics values take precedence. This is because it's completely standardized, unlike SMART.
 
Martin said:
Please try new version 5.89-3515 Beta if it reports correct numbers now.

yes, thank you kind sir, donated :)
minor nitpick: total host reads disappeared for these drives (but are still visible for nvme boot drive)
 
Thank you, also for the donation.
Try to activate ATA Statistics support again, that might show Total Host Reads.
 
is this what you needed?
 

Attachments

  • beta3515_ata_stats.png
    beta3515_ata_stats.png
    134.4 KB · Views: 5
  • beta3515_sensors.png
    beta3515_sensors.png
    25.6 KB · Views: 5
My Crucial MX500 ssd (500GB 2.5" form factor) has a problem that seems similar to the original poster's problem... maybe the same problem. I'm running HWiNFO64 v6.22-4060. Here are some values logged over the last few days:
Date​
Time​
Total Host Reads (GB)
(from HWiNFO Sensors)​
Total Host Writes (GB)
(from HWiNFO Sensors)​
S.M.A.R.T. F7
(from smartctl -a)​
S.M.A.R.T. F8
(from smartctl -a)​
02/20/2020
13:50
2,024
6,429
223,801,408
1,383,179,818
02/21/2020
05:14
2,037
6,433
223,966,265
1,388,591,917
09:50
2,047
6,435
224,017,483
1,389,029,824
14:50
2,047
6,437
224,088,982
1,389,879,812
18:31
2,047
6,437
224,115,581
1,391,279,760
02/22/2020
12:23
2,047
6,442
224,290,793
1,399,713,786
19:30
2,047
6,444
224,382,402
1,404,294,903
21:10
2,047
6,445
224,421,785
1,404,677,501
22:13
2,047
6,446
224,430,155
1,405,023,474
02/23/2020
09:02
2,047
6,448
224,518,597
1,405,960,122
10:42
224,541,107
1,406,315,952
12:28
248
6,450
224,566,093
1,406,411,752
12:55
248
6,450
224,590,612
1,406,455,265
13:48
248
6,450
224,604,985
1,406,652,719
14:42
258
6,450
224,617,092
1,406,662,503

Note that THR stopped increasing when it reached 2,047 GB two days ago. This morning THR somehow dropped to 248 GB.

The values for 10:42am are blank because I neglected to record them. Prior to that point, I'd only run the Sensors section of HWiNFO, but this morning I decided to also start running the Main and Summary windows (to see what additional SMART data it would show, and whether it has the ability to periodically log data to a file every hour or so). After that is when it started showing THR = 248 GB.

Also, how do I get HWiNFO to stop showing Main and Summary in its tray popup menu? They weren't listed in the menu before I enabled them this morning (which is when it started showing THR = 248 GB). I tried to remove them from the menu to see if that would change THR back to its earlier much higher value, but I haven't been able to remove them. Unchecking boxes in Settings, then quitting and relaunching HWiNFO, hasn't removed them.

(I've been mainly interested in trying to understand the ssd's extremely high Write Amplification Factor (WAF). WAF has been averaging about 50 during the last couple of months, after I moved frequently written temporary files from ssd to a hard drive to significantly reduce the host writing to the ssd. Host writes decreased from more than 1 MByte/second to less than 0.1 MByte/second. Before the files were moved, WAF averaged about 6 (during the ssd's first 5 months of service). My goal when reducing the host writing was to extend the ssd's expected lifetime, but it clearly backfired... the Remaining Life of the ssd has been decreasing faster. My theory is that the ssd's controller firmware runs a "static wear leveling" algorithm much too frequently, which wastes scarce NAND writes to foolishly maximize equality of NAND wear. I think it's counterproductive to maximize equality of wear before the ssd nears the end of its remaining life. Instead, an ssd should maximize its life by tolerating a lot of wear inequality for most of its life. Last night I began an experiment that may be helping to reduce WAF: the ssd is being commanded every 26 minutes to run its extended self-test, which takes about 25 minutes to complete. My hunch was that by keeping the ssd busy reading itself with selt-tests, the ssd would spend less time running its static wear leveling algorithm. A negative side-effect of running the self-tests nearly non-stop is that the ssd spends nearly all its time in the "normal" power state; without the self-tests running it had been spending more than 90% of the time in a low power state. So with the self-tests, the ssd is consuming more power and its average temperature has increased about 5 degrees C, from 35C to 40C. If running the self-tests really is helping WAF significantly, I'll need to decide whether the tradeoff of power consumption for ssd life is a good trade.)

Windows 10 Home 64. The ssd is the boot drive. The ssd firmware has always been the latest version (M3CR023). There are also two internal SATA hard drives and an external USB hard drive.

Thanks in advance for your assistance!
 
My Crucial MX500 ssd (500GB 2.5" form factor) has a problem that seems similar to the original poster's problem... maybe the same problem. I'm running HWiNFO64 v6.22-4060. Here are some values logged over the last few days:
Date​
Time​
Total Host Reads (GB)
(from HWiNFO Sensors)​
Total Host Writes (GB)
(from HWiNFO Sensors)​
S.M.A.R.T. F7
(from smartctl -a)​
S.M.A.R.T. F8
(from smartctl -a)​
02/20/2020
13:50
2,024
6,429
223,801,408
1,383,179,818
02/21/2020
05:14
2,037
6,433
223,966,265
1,388,591,917
09:50
2,047
6,435
224,017,483
1,389,029,824
14:50
2,047
6,437
224,088,982
1,389,879,812
18:31
2,047
6,437
224,115,581
1,391,279,760
02/22/2020
12:23
2,047
6,442
224,290,793
1,399,713,786
19:30
2,047
6,444
224,382,402
1,404,294,903
21:10
2,047
6,445
224,421,785
1,404,677,501
22:13
2,047
6,446
224,430,155
1,405,023,474
02/23/2020
09:02
2,047
6,448
224,518,597
1,405,960,122
10:42
224,541,107
1,406,315,952
12:28
248
6,450
224,566,093
1,406,411,752
12:55
248
6,450
224,590,612
1,406,455,265
13:48
248
6,450
224,604,985
1,406,652,719
14:42
258
6,450
224,617,092
1,406,662,503

Note that THR stopped increasing when it reached 2,047 GB two days ago. This morning THR somehow dropped to 248 GB.

The values for 10:42am are blank because I neglected to record them. Prior to that point, I'd only run the Sensors section of HWiNFO, but this morning I decided to also start running the Main and Summary windows (to see what additional SMART data it would show, and whether it has the ability to periodically log data to a file every hour or so). After that is when it started showing THR = 248 GB.

Also, how do I get HWiNFO to stop showing Main and Summary in its tray popup menu? They weren't listed in the menu before I enabled them this morning (which is when it started showing THR = 248 GB). I tried to remove them from the menu to see if that would change THR back to its earlier much higher value, but I haven't been able to remove them. Unchecking boxes in Settings, then quitting and relaunching HWiNFO, hasn't removed them.

(I've been mainly interested in trying to understand the ssd's extremely high Write Amplification Factor (WAF). WAF has been averaging about 50 during the last couple of months, after I moved frequently written temporary files from ssd to a hard drive to significantly reduce the host writing to the ssd. Host writes decreased from more than 1 MByte/second to less than 0.1 MByte/second. Before the files were moved, WAF averaged about 6 (during the ssd's first 5 months of service). My goal when reducing the host writing was to extend the ssd's expected lifetime, but it clearly backfired... the Remaining Life of the ssd has been decreasing faster. My theory is that the ssd's controller firmware runs a "static wear leveling" algorithm much too frequently, which wastes scarce NAND writes to foolishly maximize equality of NAND wear. I think it's counterproductive to maximize equality of wear before the ssd nears the end of its remaining life. Instead, an ssd should maximize its life by tolerating a lot of wear inequality for most of its life. Last night I began an experiment that may be helping to reduce WAF: the ssd is being commanded every 26 minutes to run its extended self-test, which takes about 25 minutes to complete. My hunch was that by keeping the ssd busy reading itself with selt-tests, the ssd would spend less time running its static wear leveling algorithm. A negative side-effect of running the self-tests nearly non-stop is that the ssd spends nearly all its time in the "normal" power state; without the self-tests running it had been spending more than 90% of the time in a low power state. So with the self-tests, the ssd is consuming more power and its average temperature has increased about 5 degrees C, from 35C to 40C. If running the self-tests really is helping WAF significantly, I'll need to decide whether the tradeoff of power consumption for ssd life is a good trade.)

Windows 10 Home 64. The ssd is the boot drive. The ssd firmware has always been the latest version (M3CR023). There are also two internal SATA hard drives and an external USB hard drive.

Thanks in advance for your assistance!

This looks like an overflow in reporting, please attach the HWiNFO Debug File for a detailed analysis.
I'm not sure what exactly would you like to remove from the tray menu, but I'm pretty sure it has no effect on reporting of the THR value.
 
Thank you. I have analyzed the data and it's unfortunately still the same problem. The drive firmware is truncating values reported, so if they reach a certain limit (32-bit), instead of extending to a higher nibble, they start counting from zero. This problem occurs in reporting of ATA Statistics values - this is a standard which offers plenty of room to report higher values, but the firmware doesn't use it. Only the drive vendor can fix this with a firmware update, so you (as their customer) might contact their support and report this.
 
Back
Top