I have (and LOVE) my Intel 750 400GB PCIe. I use it as my boot OS drive on Win10 64bit for the past 8 or so months, I was very surprised to see that the "estimated life remaining" / wear leveling SMART count has already dropped to 88%. I also noticed that it has quickly dropped from 90% to 88% in the past 3 weeks with no major increase in writes on my end.
I think my write workload is pretty light: aside from win10 VM being on the drive I also left 20GB unformated / free (ie over provisioning)
I have tons of experience and usage of various SSD (most intel), ie several 24 x SSD rackmount servers of all types of intel SSD models (some enterprise, some consumer) so i have a decent feel for how wear level endurance decreases over time, and in fact this 750 pcie has dropped the fastest of any SSD ive ever seen.
According to my calculations (from the SMART data host bytes written) ive only written about 8 TB of data which isnt a ton for 8-10 months of boot drive type workload. CrystalMark Disk Info also shows 7900GB written / 12.8TB read
Can anyone shed some light on this? or maybe my calculations of data written is incorrect?
PS / FYI: i did install / update to the intel supplied nvme drivers as soon as win10 was installed
We understand the wear out of the drive has decreased a lot in not even the first year, this is not normal, but, we would like to verify some things first:
-What are the specs of your system? (Include motherboard BIOS version)
-We understand this is your boot drive, what is the workload you put on this drive?
We will be waiting for your response.
Thanks for the quick reply!
My workload is such that I use the PC daily, it's a bit hard to describe but I do a lot of web browsing (Google Chrome) and run several network management tools as I specialize in network security and network IT As well as do some web oriented programming (php , Java, sql). I also have Adobe Photoshop in premiere Pro but don't use them very often although I have from time to time for small jobs. Being that I am in IT, I'm aware of what type of activities would write to a disk/SSD, I often look at resource manager disk tab as well to see which processes might be reading or writing and rarely see anything unusual ( however I've only been using windows 10 for about a month and all my prior use was on Windows 7 both boot on the 750 pcie) . I barely do any gaming although I have played Battlefield 4 online a few times on this system.
More so than my workload wouldn't the SMART data be a bit more relevant as it shows the write amplification as well as the bytes written and p/e data (or maybe that data could be in accurate?) ?
I do also have 2x Intel sata SSDs in raid 0 as well as 2x 2 TB Hitachi hdds in raid 1, and in addition a majority of my data is located on a raid 6 server ( so a lot of my data is accessed through a network share, i.e. My local "my documents" folder is practically empty ). For example: of the 350 GB formatted space on my boot 750 SSD (400gb version), I have about 202 GB free.
The motherboard is an Asus rampage IV (4) black edition running bios 0701 (the only later bios is a beta 0801 that addresses using 128 GB of memory according to the change log) . My CPU is an i7 4930K - 6 core and I have 32 GB of RAM.
I did do a few benchmarks on the 750 SSD when I first got it ( bc it's such a cool, fast drive! but during those benchmarks it was a secondary, non boot drive as I was just testing it) I do have those results saved. Current benchmarks show the drive still performing very well although some numbers are a bit down but that maybe because it's actually the boot drive under these current benchmarks versus the originals benchmarks. None of the benchmarks I've ever ran were endurance type tests nor heavy write wise.
I can provide any additional info you may need. Thank you!
EDIT (july 23 2016, 922pm): My wear leveling has now fallen to 87 (from 88) in the past day or 2 (I def. haven't done anything write intensive recently that i can recall except normal daily use), im attaching an updated exported CSV from Intel Toolbox (also the one from 2 days ago when WL was at 89) . BTW, i forgot to mention earlier i do have screen shots of the SMART "details" data from intel toolbox going all the way back to November 2015 when i first got my 400gb 750 pcie (i took screen shots bc i noticed the wear leveling was falling pretty quickly, so about every 3 weeks i would save/log screen shots of all the smart data, i can post those if that would help)
A quick/recent timeline from that prior data:
June 25 2016: Nand Bytes Written: 1,746,519 Host Bytes Written: 184,899 (didnt record CM DiskInfo) Wear Leveling: 91 (raw: 579839918100)
July 15 2016: Nand Bytes Written: 2,624,832 Host Bytes Written: 219,295 (Host Reads:12422 gb Host Writes: 7071 gb) Wear Leveling: 89 (raw: 811771822103)
July 22 2016: Nand Bytes Written: 3,013,440 Host Bytes Written: 223,459 (Host Reads:12894 gb Host Writes: 7152 gb) Wear Leveling: 87 (raw: 919149346839)
(note: the Host Reads/Writes data in ( ) is from CrystalMark Disk Info 7.0.0 the wear leveling number and Nand bytes written/reads is from intel toolbox, both apps are taking their data from the 750's SMART stats)
BTW: I keep my PC on 24/7 from day 1 (always have on all my PCs, and ofcourse on all servers)
EDIT 2 (july 24 8:32pm cst)- Overnight, last night, i kept note of how many GB my PC writes to the 750pcie SSD (boot) while im not using it (i did this via CrystalMark DIskInfo reading the SMART data from the 750pcie for Host Writes). While my PC was not being used for 13 hours ( i did have several google chrome windows and tabs open as well as a few other small apps running, all across a few windows 10 "virtual desktops" [the alt-tab kind of virtual desktop, not real "VMware type" virtual machines]) 4GB of data was written to the 750 SSD. I find this to be somewhat normal for a boot drive with several apps running idle though..
Also one option i though of is: being that i have a separate raid0 array of 2 x sata intel SSD drives, i could move my Windows 10 "Virtual memory paging file" to that location if you all think that would help. I guess the real question here is: Should my wear level be at 87 after 7-8 TB of data has been written to the drive (according to SMART stats)? That would predict i can only write ~70 TB total before the wear level is at 0 and my 750 PCIe locks itself in Read Only Mode until power off (at which point it will brick itself if i recall correctly).
Thanks for replying back and the information provided. We highly recommend you to please get in contact with our support department in order to further assist you with this situation.
You can contact them over the phone, chat or e-mail in the following http://www.intel.com/content/www/us/en/support/contact-support.html# @16 link.
Please let us know once you have contact with them.
Ok, I have opened a Intel Service request support ticket. My ticket / reference number is: 01553540 I have cited this thread in the ticket request as it contains alot of info related to my issue.
Btw, My SMART Wear Leveling Count has now slipped to 86 in the past day or 2 (whereas my Total Host writes for this drive has only increased about 75-100 gb from 87 to 86)
also when i bench mark the 750 pcie, the results are quite a bit slower than my orginal benchmarks i recorded from day 1. This could be bc the 750 is now a boot dirve vs a 2nd drive during benchmarks, but i did take care to ensure that very little disk IO was occurring when i ran these recent benchmarks. ( i also moved my win 10 VM page file over to a separate intel SSD drive i have connected) However my VM page file rarely would grow beyond 3gb in size (its set to use between 1024mb and 24gb in windows VM settings, so a 3gb page file is small / pretty un-used - so i dont think that the page file was causing alot of write activity when it was on the 750, however i can't be sure of this)
(i have a feeling a hardware/SSD swap and alot of disk cloning / disk imaging and testing is going to be in my future )
thanks for all your quick help with this issue.
Thanks for letting us know about this, we are sure our support team will provide you with further assistance on this and will take care of it.
Please let us know if any other help is required from us here.
Hi, It was a long process, but just this past weekend I was able to get an RMA'd SSD 750 that i was able to swap with my bad/problem 750 SSD and they charged my CC (750$ + 25$ fee, although they waved the 25$ fee bc it took a few more days than normal for them to get back to me -- I will ofcourse get teh 750$ CC charge refunded once they receive my bad 750 via return shipping).
However, I'm pretty disappointed to see, that this NEW RMAed 750 has already dropped from 100 Wear Leveling count to now 97 in 7-10 days of use.. again same system (win10) not much write activity (624 GB Host Write , 760 Host Read .... keep in mind about 280GB of those writes came from me running the Intel SSD toolbox's Full Diagnostic Scan)...so we are talking 3% of 127 TB should be about 4TB of writes, and ive gone though that in 400-600GB of writes ((
For Erase cycles im showing Min: 2 , Max: 104, Avg: 49
I do know the unit they sent me Had firmware from 2 x updates ago (i did update its firmware, before i did antying so that i matched the bad 750 i was sending back).
This is pretty disappointing... additionally i will be calling the RMA team tomorow, as they received my UPS returned RMA card on 9/12/16 (this past monday) yet they have not refunded the 750$ to my credit card yet (i checked a few hours ago). and i need them to do that before i hit the end of my CC cycle or i'll be stuck paying the 750$ , only to get it refunded next month.
This is sad because i love LOVE intel products (this is first real issue ive had whtih any intel products).. I just 2 weeks ago ordered 24 x 1 TB Intel Enterprise SSDs for a client server build im doing (i had free reign on the parts and easily could have gone with HGST or Samsung Enterprise SSDs for thsi build, but went with intel bc of my history with them)..
Maybe i need to look at getting a P3700 PCIe enterprise...people are selling those used for ~ 200-400$ now
Attached is my ssd toolbox CSV export from thsi NEW 750 pcie @ 97 life remaining.
Im also attaching a ssd toolbox CSV export from 5 days ago when the WE was at 98 .
We are concerned that the new SSD you received is already showing 97% wear out in a week. This is definitely not normal and as far as we can tell, you use this drive as a regular use, correct?
You need to keep in mind, this is a consumer drive and not enterprise, but taking a look to the logs, it does not seem to be overwritten.
We strongly recommend to check this with the support department and check for further options.
I have been trying since you made this post to open a new support case at
but i keep getting this whenever i hit submit (after filling out all the fields and typing up my long explination and uploading 2 x attachments (ive tried to submit 6 or 7 times now, both early and late yesterday and today as well:System Unavailable
Intel Customer Support Service Request Management is unavailable. We are actively working to resolve the system issue.
We apologize for any inconvenience this may cause and appreciate your patience.
- Intel Customer Support Team
I really love Intel products, but this past month dealing with Intel support has been an absolute NIGHTMARE.
We do apologize for all these situations, we understand how frustrating this could be and specially since you need to get in contact with our support team.
We will forward all this feedback you have provided and at the same time we suggest to open a ticket over our chat support so they can work on this request, http://intelsupportchat.force.com/icslivechat/ics_tech_ssd_ww_english_Chat here is the link, the support will be open from 7am to 5pm PST (Monday to Friday)
Please let us know.
Ok i had a chat with the live chat. Im not so sure it went too well though. I was told that the person on my case couldn't be located and the live support tech was going to email that department and see if he could take over the case. I said great, and could he please update my case with this info, he said yes, but that has yet to be done (nearly 4 hours ago).
Since i couldn't make a new case (intel web server error), i had to add comments to my old, closed case, i just want to point out in that old case, the LAST reply i got from an Intel rep was 8/8/16, since then i have made 8 posts / replies that have either been ignored, or simply not replied to. (the most recent was a few minutes ago updating that the wear level on the "new" 750 ssd pcie has now fallen to 96).
At this rate this ssd will last me less than 1 year. My hardware cant be to blame as the SMART stats are totally independent of my hardware and those show ONLY 706 GB of writes. It also shows Min erase cycles: 2 Max erase cycles : 134 avg erase cycles: 64.
this drive is warrantied/ rated for 70GB of writes a day for 5 years. Im nowhere near that... This 2nd RMA drive sent to me (sent with the 1st revision firmware btw) is worse than the drive i sent in for RMA...
I told the live support tech (because he seemed to suggest there is no solution), that perhaps some tech at Intel who knows what they are doing should take a new 750 pcie, perform some kind of real writes to the drive for say 5 days (such that the tech can confirm that the Wear Leveling count is working as it should on that specific drive), and then send me that drive as an RMA (even if it gets to me with a Wear level below 100 ) and if that drive still does the same thing, than ill just be stuck with it (and sell it on ebay and get a samsung pcie or a Intel enterprise pcie)
Thanks for replying back.
We will be checking with the department in charge so they can take actions about your situation. Please expect an answer soon from us.
We would like to confirm that it seems our support department is handling your case now and they should be providing you with further steps or solutions.
In case you need anything else please let us know.