Which SSD last the longest? I read they only last for 3 years
>>53958842
More like 30.
>>53958842
Even if you write data all day every day, it'll last longer than your average HDD.
>>53958932
HUE HUE HUE
>>53958932
>anos
>>53958907
That's almost as old as your mom.
Almost all of the name brand ones will outlive you. Even MLC NAND ones are rated at about 3,000 write cycles before failing (ie ~768TB for a 256GB MLC NAND SSD). The only significant problem left is the controller which can have shitty wear leveling algorithms or die before the SSD's NAND wears out. Luckily most modern name brand SSDs have high quality controllers that should also effectively outlive you.
In addition to all this many SSDs will become read only should you wear them out close to the point of a catastrophic failure and give you time to transfer your data to a new SSD.
>tl;dr
You'll likely never see a modern SSD fail and you'll most likely upgrade your SSD every 5-10 years so you'll only lightly wear them out.
>>53958965
Yeah, no.
SSDs write data pretty fast, so they get to hundreds and thousands of TB in a matter of days. An SSD that's constantly writing data will fail in a few weeks
http://packet.company/blog/
>>53959044
>So, after over 7PB written the 850 Pro has finally given up the ghost. It's been an amazing run over the past 9 months, and the drive has vastly outlasted its puny 150TB warranty - by a factor of more than 46 times.
>by a factor of more than 46 times
>>53959044
Looks like I'll be buying a Samsung one then
>>53959037
Do you even own a SSD or are you still living in your parents house posting on mommies laptop.
>>53959037
>SSDs write data pretty fast, so they get to hundreds and thousands of TB in a matter of days. An SSD that's constantly writing data will fail in a few weeks
lolwut
>>53959076
You can go with a cheaper name brand like sandisk, it will still virtually outlive you. No way you're gonna write more than 50GB a day on average (~365TB in 20 years btw).
Like I said before you'll most likely upgrade in 5-10 years too. Who knows, maybe 8TB SSDs will be like $100 by then.
>>53959024
This individual knows what he's talking about. Buy any decent SSD today and it'll be the last part of your system to survive. Your data is also extremely hard to lose, as tripnigger said, when write limit is somehow reached or whatever the drive becomes ROM.
>>53959037
>Poe's law
All (except chink shit). You're going to be fucking dead before your SSD dies. If you get a 2TB SSD today you can use, die, pass it on to your grandkids, they use it, die, and pass it on to their grandkids. You're seriously underestimating how durable SSDs have become (in both nand and controller tech).
Though by the time you die, 1,000TB SSDs will become affordable.
are the ssds that are found in higher end laptops like rMBP and Surfaces book well built, and will last 5+ years?
>>53959408
>will last 5+ years?
Much longer than that.
>>53959408
Nope, almost all modern 256GB or bigger name brand SSDs found in literally everything will outlive you, aka you'll be dead and your SSD won't be.
>>53959388
>i can reach all my glorious porn to the next generations
>gives familyporn a new meaning
>>53958842
you will die before you SSD...
How are ADATA SSDs?
>>53959495
Your porn is too old and low-res.
>>53959495
>this porn isn't even VR, grandpa
>>53958842
Just keep it in the fridge and it well be fine.
Depends on storage space and controller. Old 32GB SSDs with shitty controllers will shit themselves in less than 100TB of writes. Modern ~500GB SSDs will endure more than 1PB of writes; with v-nand ~500GB SSDs capable of sustaining ~3PB of writes. Most SSD controllers also now have global wear leveling algorithms and are very high quality.
Moral of the story: don't worry about it.
>>53959515
Not name brand, controller might shit itself before the nand wears out. Get an ssd plus from sandisk if you're on a budget (like $100 for 480GB).
>>53958842
Huh thats weird because their meant to last for 30 odd years at std read/write cycles
Mine is well over 3 years
>>53959515
HAH thats my exact drive.
Controller is absolutely fine works better in windows 10 though
Anyone any experience with pciE ssd vs sata 3 ssd?
Would it be worth the extra few bucks to get a pciE instead of a sata ssd?
>>53959611
>Controller is absolutely fine works better in windows 10 though
You can't tell until it fails. I've never researched Adata so I can't conclusively tell you how shit their controllers are.
>>53959639
^ curious as well about that
>>53959639
>>53959651
I own an Intel 750 series NVME drive. It is blazing fast, doesn't boot faster than a regular SSD because it has to be initialized by the system or whatever, getting Windows to boot off of it was a bitch. Had to install the NVME drivers while reinstalling windows.
>>53959639
>Would it be worth the extra few bucks to get a pciE instead of a sata ssd?
I think it's a waste of money. Fast sata 3 SSDs like the 850 pro can write data at a staggering ~500MB/s. That means ~30GB/minute or ~1TB/30 minutes. Anything faster than this seems pointless and overkill.
>>53959720
So installing your os on it is problematic? I've heard that alright.
What motherboard you got?
>>53959757
Installing not so much of an issue, it's more getting it to boot. Took lots of playing with bios, Gigabyte Z170X Gaming 7
>>53959812
>>53959749
Thanks goys.
>>53959024
So long as you power it up more than once a year....
>>53959812
Though I should note, I have no issues now.
Probably not worth the money, but I like owning it.
>>53958907
Got a 4 year old Intel SSD and it's reporting bad sectors. Also, was never used much.
>>53959884
>intel
>not an 850 evo
>>53959945
>evo
>not pro
>>53959854
>So long as you power it up more than once a year....
This is a meme, SSDs have data retention of like 5+ years if left in storage at room temperature. Extremely worn out SSDs left in extremely hot temperatures (40°C+) will however lose data in as little as days-weeks. Of course for the almost all user this will never happen.
moral of the story: don't leave your SSD in the Sahara dessert after using it in a data center.
>>53959037
Sure, if youre going to be writing that amout of data to them. What do you do that constantly writes data to the disk?
>>53960325
Compiling code? Java is shit I'm that sense as creates lots of crap especially that god awful web logic server
>>53960713
Are you legitimately mentally disabled?
>>53958842
>I read they only last for 3 years
What idiot told you that?
>>53958842
i have an OCZ agility 4, not the fastest but has lasted me 4 years and going strong
good names right now are samsung & intel from what i hear.
also look into m.2 if you have a compatible mobo
A good sad under 50 bucks? I need it for a Facebook machine for my cousin
>>53960713
>compiling code
>worried about writes
Epic lol
>>53960843
Just get an HGST HDD you cuck. 50 bucks is enough for a 1TB HGST.
>>53960843
Just buy the cheapest one you see, fuck it.
just bought a kingston savage 240gb. does it come with all the cables / adapters or is it only for the ones with the 'upgrade kits' only?
>>53958932
>5 anuses
wut
>>53959578
>>53959641
>>53959611
why would the controller die before the mem?
When I got my SSD in 2009, /g/ told me it would be dead in 2 years. I have had two stand HDDs fail in that time. The SSD is still going strong as the primary drive.
>>53961872
is it this one?
>>53961892
No but it was a Corsair.
>>53958932
I have a 950 Pro, it's a nice drive :)
>>53959020
>implying OP is old enough to be 16
Noice bait, show me some incestous mom-son porn in POV please.
>>53960843
Don't bother buying cheap SSD, Get something like Samsung 850 or Intel. Most cheapo SSDs have shit controllers that will give up if you stare at them.
Is the intel 535 480GB a good SSD?
>>53962080
I have two 530s (pretty sure it's the same but only 180 GB) and they've been rock solid in a raid0 for a few years now
>>53961846
Because it's shit. This is mostly a problem on chink shit though. Bad thermal regulation, shitty wear leveling algorithms, ect are all factors that can compromise the integrity of the data on a SSD.
>>53962123
Wear leveling involves the NAND not the controller
>>53962038
Can vouch for this had two machines fail after a year only to realise that they had shitty kingston 16gb ssd's installed, no fucking warning either just died also they do funky ass shit when they fail making it look like the ram is faulty or motherboard on the way out rather than boot disk error during post from a regular hdd can be confusing as fuck
>>53962210
The controller is in charge of writing data to nand and obviously in charge of wear leveling.
>>53962244
You wear level NAND, you don't wear level the controller. Shitty wear leveling algorithms means the NAND is wear leveled badly. Has nothing to do with "controller lifetime" or whatever is being talked about here,.
>>53961952
You also seem to be really new here
>>53962277
You don't think it's a good drive?
>>53961824
American education system? Anos is the plural from ano in Spanish and it comes from the Latin anno
>>53962319
That's años dude. Anos means anuses.
>>53958965
My old Maxtor 273MB hard disk is still running fine after over 25 years.
Hell, my IBM Type 0665 46MB hard disk is still running fine and dandy with only something like 4 bad sectors and it's nearly 33 years old now.
>>53962319
Nice try but the word Anno is from a Game
I bought mine in 2011 and it still works.
SSDs aren't that good desu, recovering data from them is almost impossible
>>53963176
>storing data on SSD
Now why would you do that? Why would you do that?
>>53963180
because they're for that
>>53963176
If by impossible you mean much easier and more reliable than a HDD then yes,.
>>53963196
So you store your family photos on an ssd?
>>53958842
I have a 120 gb ocz vertex 3
did a fall for the ssd meme to quickly? did I fuck up I feel like i fucked up
>>53962319
It's annus in latin, and means year.
>>53959984
What about us where 40C is our room temperature?
>>53963227
I install windows and have documents etc
>>53963227
What are you getting at here?
>>53959037
Yeah, it's too bad we all run manifold data servers. You fucking imbecile.
>>53960763
Thats because Intel and Micron makes the best.
>tfw
>nas full of SSD
>versus load as fuck HDD
>>53963298
he's implying that SSDs should be used not for storage but for fast access.
>>53958842
>The SSD Endurance Experiment represents the longest test TR has ever conducted. It's been a lot of work, but the results have also been gratifying. Over the past 18 months, we've watched modern SSDs easily write far more data than most consumers will ever need. Errors didn't strike the Samsung 840 Series until after 300TB of writes, and it took over 700TB to induce the first failures. The fact that the 840 Pro exceeded 2.4PB is nothing short of amazing, even if that achievement is also kind of academic.
http://techreport.com/review/27909/the-ssd-endurance-experiment-theyre-all-dead/4
>4 8 and 16tb SSD's when?
They are already getting cheap enough a 2tb ssd is only 800
>>53958842
>3 years
lel
Good one. more like 8 more years.
Based micron / crucial
>>53963534
what the fuck, 5 months old and lost 2% already?
I think you got a broken one mate
>>53963534
>98%
How...?
>>53958842
>2012 840 Basic 120
Works fine.
>2014 840 Pro
Same speed as 850 but 22 TBW
>2015 850 Pro
Literally perfect
Should I get a 2nd 850 Pro for RAID? On Ivy, no no fast m2 and no nvme / lains for Intel PCIe
My SSD outlasted the company it was made by.
>>53963596
>>53963632
on the old frimware i had on it i think it was MU01 it fucked up my drive I upgraded a few weeks later and it stopped degrading.
and its based in writes i did more writes than >>53963632 because my wife doesn't know how to write to her HDD and just deletes and writes shit and copies it over to her HDD.
I keep telling her but she never gets it.
>>53963294
And why would you want to recover them?
according to this thing, this samsung 840 pro only used 9% of its wear level count and still has 100% of reserved blocks and has over 40tb of write after ~4 years
used as a scratch disk mainly in a workstation
i'm guessing 950 series and higher can last even longer
>>53963869
important documents? work related stuff
>>53963986
The company is totally cannibalized by toshiba now. It was never a budget brand either, they regularly put out top performing ssds at proportional costs. Their revodrive 3 from 2011 still has the best seq read speed of any ssd on the market even 5 years later.
>>53964100
nevermind i dun goofed
Had a OCZ Onyx 64gb since 2010 and its still going strong
Not sure how to test it for bad sectors in Ganoo Lonux though
>>53959125
>>53959264
>>53959339
>>53960325
>>53963307
Post that I quoted said "Even if you write data all day every day". In that scenario it'll be dead in weeks.
Not my fault he gave an unrealistic scenario. Learn to read, Idiots.
>>53963854
Hah, women!
I have one from 2012 with full disk encryption and it works fine.
>>53962339
>Maxtor
Holy shit, somebody else with a Maxtor. And they were the ones with really bad HDDs.
Mine still works.
>>53958842
What about Z400S by Sandisk ?
>>53959515
I have an adata xpg900 240gb, it's been serving me for one year and the "life left" report 98%
SSDs die because their controller chips die.
The memory is the most durable part of the entire SSD.
They die randomly at a pretty low rate but if they survive the controller exploding then they should last 20-30 years.
>>53964097
If they're so important why don't you upload them to Dropbox?
I got an 850 evo 1TB, unless I write over 150 GB a day I don't have to worry about it dying on me within 5 years
I probably write on average like 5GB a day so the lifespan is more likely in the 60-70 years
eight fiddy pro should be pretty good. Or any legit SLC enterprise level drive (probably fucking expensive tho).
TBW for my 256GB 850 pro is 0.99TB after almost a year.
So it would take over 100 years to reach 150TBW with my usage...
And the warranty is 10 years or 150TBW (whichever comes first). Not too worried desu.
>>53958932
>Duracao
>120 meses
>120 minutes
The fuck?
>>53959024
>You'll likely never see a modern SSD fail
Unless of course you get a firmware error, a unlucky power outage, or a failure of the power control system on the motherboard which influences the SSD and fucks it up sourcing either from the mobo itself, the PSU failing in some way, or other components connected to the mobo influencing failure elsewhere.
>>53959984
>This is a meme, SSDs have data retention of like 5+ years if left in storage at room temperature.
Academic papers pls.
The magnetic dissipation of particles on the HDD platter way outlive the SSD under the same conditions in a full drive plastic encasing, and i have seen no proof whatsoever anywhere that the transistor held charges can last that long under continuous non-operation of the drive.
Ergo, academic papers proving your words pls.
>>53964961
which is why i have shit that just has to go fast on my ssd, like OS and apps
and important shit like porn on regular hdds
you should read this
http techreport com/review/27909/the-ssd-endurance-experiment-theyre-all-dead
>>53965032
Why would you put your more important files on the less reliable drive?
>>53965032
Precisely.
>>53965052
say that to my face
>>53965052
The only important files for that anon are his archives of media/documents/work shit.
So long as those are stored on an HDD, they are safe since the only way an HDD can get fucked is vibration based. Even if it has mechanical failure, you can still use special hardware to retain the data and clone it onto a new drive.
As far as system files go if they die with the SSD, no sweat. You didn't really lose anything, you just have to go through a headache to re-install shit again.
>>53965088
>>53965107
SSDs are more reliable than HDDs.
>>53964961
All of this can happen to a HDD.
>>53965115
An HDD doesn't store data in the form of electrical charges anon. It operates by flipping magnetic particles.
If the same things that can happen to an SSD could happen on an HDD, it wouldn't be called an SSD anymore.
>>53965133
HDDs can have firmware errors
Power outages can affect HDDs and corrupt data
Etc
>>53965143
We are talking about instant failure here anon, not long-term corruption of data which can affect even tape drives. Also data retrieval, since an SSD operation failure will render it impossible in most cases, while an HDD won't so long as platters are intact.
>>53965188
>HDD head needs to seek to a location on a platter
>Power failure
>Your data is corrupt
>gg
I've been hammering this OEM version of an 830 for a while now, up to 60TB writes. It's only down to 93% health.
>>53965326
>HDD head needs to seek to a location on a platter
>power failure
>HDD head can't move anymore
>sky magic corrupts your data
>>53965341
It needed to seek for a write.
Data doesn't get written and your power is corrupt.
>>53965350
Power > data
>>53965350
It needed to seek for a power.
Data doesn't get powered and your write is corrupt.
>>53965359
You're just memeing now so we can officially declare this over and that you were wrong.
>>53965188
>instant failure
And we're back to SSDs being more reliable than HDDs.
>>53962319
Años is spanish for years
Anos is spanish for annuses
>>53965396
How exactly?
>>53965407
Instant predictable failure is much better than stretched out unpredictable failure.
>>53965428
How is instant failure predictable?
>>53965443
In the case of SSDs you can predict when the drive is going to fail due to its write limit being reached. You can't do this with HDDs.
Obviously you can't predict things like power failure killing a drive, but this applies to HDDs too.
>>53963381
>2tb ssd is only 800
Honestly tempted
>ITT
>>53965456
>Obviously you can't predict things like power failure killing a drive, but this applies to HDDs too.
Except it's irrelevant if a power failure disables an HDD because you can still retrieve already existing data from the platters. Which is the topic-at-hand.
>>53965498
You can remove data from the NAND.
Why would a power failure fry the NAND?
>>53965508
>>53965498
>having such a shitty PSU that it blows your drives
>>53965508
>You can remove data from the NAND
That's assuming the drive can be powered on. And with a 50-50 chance.
And also assuming the power outage didn't affect the transistor charges that form the data.
Do you know how data is retrieved from NAND?
>>53965545
>having such a shitty PSU that it blows your drives
Any component on a motherboard can affect the entire system of power distribution anon, not just the PSU.
>>53965456
>he thinks the nand is the issue
lewl m8
Controller on HDD Dies -> hdd needs a new controller but platters are fine (the shit holding the data)
Controller on SSD dies -> Nand gets rekt, and since the nand is holding the data, data is gone
>>53965574
>>53965578
If it's that bad you don't even need to replace the controller. You can just put the NAND in a drive with a working controller.
Do you think a dead controller kills the NAND with it or what? This thread is bongos.
>>53965603
>he thinks it doesent
c u c k
u c k c
c k c u
k c u c
>>53965574
Even tho the drives get power directly from the PSU?
I've had my 840 evo for almost 3 years and it's still 100%, and the 840 evo is that flawed Samsung model.
Modern SSDs have been stress tested for hundreds of terabytes to even petabytes of writes, I worked out my average usage on a yearly basis and it would take me 20 years to reach what my SSD has been tested for, and it's already much cheaper to buy a 256GB SSD now than it was when I bought it.
>>53965481
http://techreport.com/review/27909/the-ssd-endurance-experiment-theyre-all-dead/3
Lesson 1: I can see why people shill for samsung
Lesson 2: HDD's rarely last over 10 years if they're used everyday. If you wrote 500GB EVERY DAY, that's 1825TB over 10 years, still not enough to kill the samsung. (Not to mention if you're writing 500 gigs a day, you're not looking for consumer tech).
Lesson 3: SSD's often fail gracefully, you can read and copy the data off an SSD even after it has run out of write cycles. If a HDD fails, it's often a complete failure, requiring a professional repair to recover the data.
Evo 850 pro, absolutely no doubt.
>>53959037
Is this bait? Fuck off kid
>>53958842
Using ocz vertex 2 from 2010.
>>53964932
months, bitch
>>53959884
>old ssd tech
Irrelevant
840 pro past 3 years no problem here
>>53959515
I bought USB 3.0 drive from them once. Controller died halfway h2testw.
>>53965987
Nobody said at max speeds, you ASSume as such.
Fuck
>>53959515
bought the 480GB version for $110 a few months ago on newegg.
it's fine. i use it for my vidya gaymes.
>>53965722
I'm not against SSDs (got one myself and I'm happy with it), the point I tried to make (vaguely) was that everyone posting "I've had X for X years and it works great" are retards.
And honestly your link is also anecdotal evidence from what I can gather, as they only tested one drive of each model. But at least they tested more than one drive.
But I also know that there have been large scale statistics done on SSDs (mainly malfunction rates over several years) and only the first generation drives had issues, and later generations last many many years (except for a few models with abnormally high failure rates).
>>53959037
We've assuming OP has a vaguely average use for the SSDs. Yes, you could probably kill one fairly fast if you're writing *massive* quantities of data to it constantly, but
> Why
dope senpai
>>53959020
>29 year old mom
>>53966452
pic related.
my 120GB intel takes a lickin' and keeps on kickin'
>>53966628
>>53966651
my head when i accidentally quoted him instead of myself one post above
ssd jokes aren't funny and they never were.
>HDD
> guaranteed 10 year life span
>SSD
>make live for 6-8 years or it may randomly fail and die on you
Ssds are still a meme
>>53966695
>SSD meme meme
>>53966695
moving parts fail more often than non moving parts.
you wouldn't know this since you don't drive and never leave the house.
Will leaving free space on a SSD have the same, or nearly the same, effect as over provisioning with use of regular TRIM?
My SSDs have 10% OP base, and I set them to 15%, but I typically leave 40~50% of the drive free anyways. Would it make more sense to just increase the OP value more, and not worry about the usage of space on the drive after that?
>>53966404
I'm guessing streaming Twitch and downloading a lot doesn't help with my SSD life.
>>53966769
it makes more sense to use 80-90% of the drive space.
>>53966721
SSD have moving parts
electrons
>>53966817
SSDs decrease in performance as space is used up though, don't they? That's where my question was stemming from.
>>53966837
in theory, yes. in reality, not really.
slam that bitch full.
>>53966814
You didnt move user files and page file to a HDD?
>>53966830
srsly, nigga?
>>53966837
no... where the hell did you get that misinfo from?
>>53966858
>no... where the hell did you get that misinfo from?
>http://www.sandisk.com/assets/docs/WP004_OverProvisioning_WhyHow_FINAL.pdf
>http://www.seagate.com/tech-insights/ssd-over-provisioning-benefits-master-ti/
>http://www.samsung.com/global/business/semiconductor/minisite/SSD/global/html/whitepaper/whitepaper05.html
Any number of forums, and people I've talked to as well. Is all of this false, am I reading it incorrectly, or what?
>>53966858
>You didnt move user files and page file to a HDD?
I only have a laptop. Must be nice having the best of both worlds in a rig with SSDs and TB HDDs.
>>53966857
As in the differences are marginal, or non-existent?
>>53966872
It's mainly for getting longer life. The increased speed isn't really noticeable.
>>53966429
>480GB
>110$
What the fuck? Where I'm at they cost about 180 -.200$! Motehrfucker. I'll just fucking buy one anyway, fuck it!
>>53966907
i have 1.5 GB free on my 120GB intel drive because downloads and i haven't removed shit to another drive, so this is just what you want to know.
running crystal disk mark, set to 1GiB.
One of the best purchases I ever made.
Have shifted GBs of data over the past few months and have owned it since April 2013.
Beautiful.
>>53967022
>>53967036
Interesting, nice to know. I think I'll just bump both of mine up to 20% OP, and then proceed to not care about total space used as well. Appreciate the heads up.
>>53967090
no prob bob.
when it comes to ssd's read speed is all i care about so if i'm getting full speed with the disk full i don't give a fuck about write speeds.
>>53963713
>>53967090
I don't have the 840 Basic in this PC but its slower anyway
>Mfw Evofags enjoy slow speeds
Sandisk sd7sb7s-512g-1122 is one of the best.
Last 20 years if you only move 300 GB per month.
>>53962239
i'm running 2 shitty SSDs that were $40 for 120GB and you would have to be an idiot for not backing up your shit at least monthly.
250 GB Intel SSD +320GB & 1TB WDBlue here.
Symlinks are man's best friend.
Put documents and shit on my old 320 GB drive that I got in 2008 and is still kicking, use TB for media, and when I need a game to load fast and not stutter I got a shell extension to make symbolic links easily.
>>53968147
>having data important enough to back up
your /b/ and /s/ folders aren't that special, snowflake.
>>53968624
Interesting, but SSDs arent that expensive anymore.
>>53968783
I have other data too, for examply my hand collected music and movie folders, these were thousands of hours work. Plus I couldn't download so fast as I'm now on 3 mbit.
>>53958842
are external SSDs a thing yet? like external hard drives?
>>53959024
what about SSDs on macbooks? i have this fear that i'm gonna lose my shit but i olan on using my 2015 macbook pro for at least 5 years
>>53968909
for what purpose?
you can have a sata cable that comes out of your computer, hook it to a ssd. wow it's external now!
>>53968960
to carry it around
>>53968974
you can do that with a HDD external.
or a usb drive.
or a ssd.
>>53968944
Depends on SSD controller quality. It's probably good so don't worry about it. ~90% of SSDs out there will outlive you.
>>53958842
I have one of those shitty OCZs and its still kicking for almost 4 years.
>>53969072
just binged it, uses a samsung pcie controller
>>53969349
Then you'll be fine, don't worry about it.
Are sandforce controllers still good?
A 512gb 950 pro will outlast you and your grandchildren as well. Same with the 1tb 850 pro. Either one of those will last an eternity.
Crucial MX200 has a Marvell controller, while BX100 and BX200 have some Silicon Motion. Are those Silicon Motion controllers bad?
>>53969919
no.
>>53969958
Why are they cheaper than MX200 then?
>>53958932
>non-consistent units
>>53969969
people making a dollar an hour in china?
>>53970116
Um, what? Are you saying the cheaper drivers aren't made with little yellow hands?
ey
>>53959037
nope, running a server farm that does this all day on consumer SSD's for work.
you're just dumb.
>1-2TB still too fucking expensive to buy, have to make due with 500GB models.
>Pay attention to SSD announcements, companies still announcing and pushing out 120GB-1TB drives
>Mfw
The professional world is already sporting gigantic 16TB 2.5" drives and we're still getting these sub 1TB little shits, that we've had for quite some time now.
This SSD size increase seems to take forever, even though the tech to make them a lot bigger is already out there.
Can't wait till I can lump all of my data to an affordable 8TB SSD
Unfortunately at the rate we're going, we'll see affordable sizes like that somewhere in 2021 or so.
>>53958842
Samsung 850 Pro comes with a 10-year warranty.
I've used mine for about 2 years and I still have 99% of the P/E cycles left.
You can write petabytes to that shit and it won't wear out on you.
Linux now bans some drives with broken TRIM:
https://github.com/torvalds/linux/blob/v4.5/drivers/ata/libata-core.c#L4223
List includes Crucial MX100 and Samsung 8xx.
Why can't manufacturers design good controllers?
>>53970302
It doesn't ban them, it just disables queued TRIM. And it has done this forever.
>Why can't manufacturers design good controllers?
Because their shit is closed source.
>>53970280
Another factor is that most consumers just aren't demanding increasing storage space like they used to back between 2000-2007. There was a HUGE storage capacity war during that period that just petered out because of streaming and cloud storage.
If the vast majority of consumers don't care about giant drives but do want snappy performance, why would manufacturers bend over backwards to push capacity out to the market?
>>53970331
Whoops, sorry for the bad wording.
>Because their shit is closed source.
But one happy merchant told me proprietary is better!
>>53970280
But why would you put all data on an SSD, when they have limited write cycles? SSDs aren't good for data, that you don't have to access fast, like downloaded movies, which is what you need terabytes for.
>>53962347
I miss that game
>>53959037
this is actually closer to reality than everyone wants to believe.
if you assume you are writing 500mb/sec to an ssd nonstop, you would hit 700tb which is the limit on most drives in about 17 days. and those limits are hard set unlike a spinning disk.
now naturally, no one is constantly writing 500mb/sec worth of data to a drive nonstop
>>53970331
>Because they are closed source
There are people that actually believe this.
>>53970331
So what exactly does that mean as far as performance/long term reliability?
>>53970438
If they opened up their firmware design, the Linux driver team would be all over it.
Don't you think I'd rather fix the bugs in my shitty firmware rather than having to work around them from within the kernel?
I wish I could fix the stupid bugs in my nvidia drivers and my display's firmware, too. Sadly, I can't touch all of these.
Closed source software is literally the worst software on my PC, because it's all the software that I can't fix myself and just have to live with the bugs.
>>53970390
wait, I was referring to 1602. Anyone know if they emulate these games now? I would love to play this again. Just wondering if we'll have A.I. before windows game emulation
>>53970481
Don't act like you would actually be able to do anything about it. You would sit around waiting/hoping that someone else would find and fix the problems.
Why didn't you fix the bash bug 20 fucking years ago when it was introduced?
>>53970480
>So what exactly does that mean as far as performance/long term reliability?
It means all TRIM commands must be synchronous, which can reduce performance if you're mounting the fs with the ‘discard’ option. (Which TRIMs every removed file instantly)
The best approach is to disable ‘discard’ and schedule fstrim to run regularly (e.g. cronjob)
>long term reliability
no difference
>>53970507
Thank you for the explanation, I greatly appreciate it.
>>53970504
>Don't act like you would actually be able to do anything about it. You would sit around waiting/hoping that someone else would find and fix the problems.
I've contributed many thousands of lines of code to open source projects throughout my past few years of using Linux, including some fixes to kernel drivers. (For example, I've fixed bugs in the open source components of the nvidia kernel driver)
I also do my best to debug bugs in black boxes and report all of my findings upstream. I've managed to get at least one major bug fix into the nvidia 364.12 drivers by annoying them with new information for long enough.
I've actually reverse engineered closed source program using gdb, ltrace etc. and patched their binaries by hand (or used LD_PRELOAD to fix their API usage) in order to work around other bugs I've encountered.
I'm no stranger to fixing bugs. I just await the day companies realize I'll fix their bugs for free if they just let me. Fucking assholes.
It's SO much easier to fix bugs in open source software than in closed source software where I first have to stumble around in the dark trying to figure out what the hell is going on.
>Why didn't you fix the bash bug 20 fucking years ago when it was introduced?
What bash bug? I legitimately don't know what you're talking about.
>>53964816
samsung is love
>>53970280
give us MLC already SSD companies Reee!
>>53970581
>I've managed to get at least one major bug fix into the nvidia 364.12 drivers by annoying them with new information for long enough.
And right now, I'm bothering them with bugs in their vulkan driver implementation + how to reproduce.
I've narrowed down which specific API calls return the wrong values, under which circmstances, and it would probably be a trivial matter of fixing it myself - but I can't since it's a proprietary black box.
So my only hope is again to post on their forums and away the day they finally get competent enough to write decent drivers on their own, which no other company seems to have managed so far...
>>53970280
Well years ago there were only 60-250GB SSDs, the size doubled.
>>53970636
Sadly it they wont get any cheaper soon and we wont see cheap $/GB storage because they are that good in comparison with HDDs.
Unless somehow HDDs magically get cheaper.
>>53970581
Fix the file chooser bug
>>53970636
2011 we already had 512GB SDDs, Crucial M4 for example.
2013 Same company released it's first 1TB SSD which sold for 600$
The SSD size growth hit a standstill in the consumer market after that.
>>53970385
SSD drives are small and dead silent, that's the main reason. Bigger memory capacity mean bigger write capacity like hundreds of terabytes or petabytes.
If only there were affordable and reliable SSDs with like 4-6-8 TB capacity, then you wouldn't worry about keeping all your cache, swap and other stuff on it, as well as movies, games, music, anything - at home you will never exceed the amount of writing cycles on a huge SSD drive no matter what you put there.
>>53970747
You mean the gtk file chooser not displaying thumbnails? I already have a local fix for that.
The problem with GNOME stuff is not that people don't fix their bugs, it's that Red Hat is too arrogant to accept patches.
(Red Hat is basically cancer)
>>53970777
I been thinking for a while, do SSDs suffer from coil whine?
>>53970777
>SSD drives are small and dead silent
Funny, I actually built a semipassive build and the SSD was the loudest thing in the system. All that coil whine.
find / was like listening to a song.
>>53970280
There's a reason the professional world has much larger capacity, it's because they need it. Consumers don't. Storage needed in computers aren't really increasing. Some programs like games are getting bigger but internet speeds are getting faster meaning that it's easy to stream all media you need and just uninstall games you aren't playing and re-download them within an hour or less if you need them again.
If you really need 8TB drives you'd be better off stop waiting for such drives to reach consumers because there is no demand that will drive them to market any time soon, and instead just buy an enterprise product.
PNY 120gb
Speeds fine
Do wish I sprung for the 240gb but at the time the sale was only for the 120
>>53970802
You're doing something wrong or the SSD is defective. It's not normal for SSDs to make any sound at all.
I don't even see how there can be any coil whine, why would there be coils in the SSD and even if there is I imagine almost no current passes through them as SSDs hardly use any power.
>>53970848
>stream media
I hope you like your pixelated forced 720p streams with ads even when you pay premium to not have ads.Not to mention shitty subtile support
>>53970952
>You're doing something wrong or the SSD is defective. It's not normal for SSDs to make any sound at all.
That's what they say about every component in the system. Until the coil whine gets 'em.
>>53970783
> already have a local fix for that.
Can you link the patch?
>>53970973
How do you know it's the SSD?
Coil whine is usually caused by the power section of the motherboard or the PSU.
>>53970962
The quality of most streaming sites are more than good enough, and there are no ads and subtitles just work. Of course you have to pay, most people don't have a problem with that. It's clear you are not a majority that affect market trends in a significant way.
And even torrenting can count as streaming, I know I don't wait until a torrent has finished downloading before I start it, I start playing at like 5% downloaded and deletes the torrent when it is done (and the movie is usually still playing).
>>53970984
gtk+2 https://gist.github.com/ahodesuka/01213036b58e510dc074
gtk+3 https://gist.github.com/wfr/55d99f680fea8f815d17
>>53971017
I'm reasonably sure it's not the PSU. It might be the motherboard / chipset though.
>>53971079
>almost complete
Post a screenshot for 3.
>>53971057
>good enough
No,they arent,they drop bitrate and distort constantly(it isnt my connection,datacenter is about 15 miles from here)
Source:I had hulu,netflix,crunchyroll
They cannot compete vs traditional piracy.
>5%
Yeah no,thats bullshit.Even if you grab the first and last peice,you're not going to be watching without constant distortion as you are missing the data,IF your player can play incomplete files(vlc)
So no.
