SSD vs HDD Torture Test

[Hans Power]

I just recently cloned my OS from a partition on my 1TB WD Black to a new SSD - a Crucial M500 120GB drive. The overall performance increase was definately noticable at first but not as much of a jaw dropper as I expected. However after restarting the OS a few times and resetting the virtual memory options it went a good deal faster. I discovered that the main improvement was happening with multitasking. Also I installed some games where a lot of streaming goes on and that was definately a gamechanger! Sacred 2 Gold for example used to crash a lot on my old drive (although it was always properly defragmented and doesn't have faulty sectors) and that suddenly runs completely stable after installing it on the SSD. Pretty cool so far. Also opening the gamespecific 3D options in my Nvidia drivers took forever and now it's just there instantly. And last but not least, browsing the internet with Firefox became a LOT snappier. Now I really feel my 3,2MB/s Internet connection. And concidering that my WD Black is among the fastest consumer drives and the M500 is one of the lower end SSDs the improvement is pretty amazing overall.

[bracka]

I think it would be great to add one more dataset to the test. It's cloning the data to a new HDD and then testing. 
 

It's clear that SSD is magnitudes faster in random small sector operations than HDD. However, there have been performance improvements in HDD performance as well. I think such data point would be great addition to this test.

I have experienced solid performance boost after upgrading my data drive from 5 year old 1TB HDD to 3TB WD RED. Will look for benchmarks once I get home.

[Blingdaddy1]

I don't understand why people are saying this is an unfair test, it kind of isn't. If your going from an old hard drive to an SSD this is probably a great video. 

[CGurrell]

Good luck, the readycache drives are extremely poor and are infamous for having issues that completely prevent them from working. How those things are still on the market is beyond me. Perhaps do better research next time you buy something, quite literally one quick googling could have saved you

Yeah I figured that out tbh I didn't have much cash to spend at the time of purchase, and the closest 64GB SSD (Big enough to hold OS, sorta) was around £70 at the time, and this was £30 and would decrease boot times etc.

[ManWithBeard1990]

I don't think that's being entirely fair to the HDD. Because one test crashed, the thing was probably on its last legs anyway.

 

While I did like this video, I think a best case scenario HDD versus a worst case scenario SSD would be much more interesting consumer advice though. You know, like dual short-stroked HDD's or even SSHD's in RAID0 (like I'm running in my own system) versus one SSD, for example. Or a short-stroked hard drive versus an SSD without TRIM support. That sort of thing.

[StingBull]

Some people already said this in the comments but I think the test would be fairer if the HDD was also new. New thing beats old thing, that's not really surprising. If you tested that HDD against a brand new HDD, the new HDD would have been quicker.

 

 If you decide to redo the test have a new HDD and an SSD, copy the OS to both of them.

[Aniallation]

Long story short on SSD durability: Laptop on a subwoofer

 

https://www.youtube.com/watch?v=LMzWvNXa7VM

[shananigans89]

great video! i'd like to see something continuing from this, for example: i use a program called Boostspeed 6 by AusLogics. its like a total pc cleaning/repair program. i don't use it religiously and i don't use all the features and programs it has, but i just use it now and then to clean up junk files and other stuff. also, it has what it calls "Live Protect" where i have it running at startup, just minimized to the icon tray, and supposedly it optimizes CPU  and Memory usage. 

i think a good follow up video would be to use something like Boostspeed 6 or any other program you guys would recommend (i've also read a little about a really good program called System Mechanic), run the program and see what improvements each drive gets. show us if programs like these are worth it. 

 

great vids LTT, keep up the good work!

[IdeaStormer]

Not the wear and tear but the messed up filesystem an the like. That would still be wear and tear but virtual.

You can directly compare a ssd and HDD they are designed to preform the same general task. The ssd is a stop gap solution for a quicker hdd until a solution like holographic storage is found then the sad will die out nearly over night. Those in the industry and enterprise have basically known that since the ssds incarnation.

I don't know why you insist on comparing a floppy to a HDD an ssd it makes zero sense.

 

 

There was no proof of corruption on the HDD (aka messed up drive). The HDD still works and if it was corrupt it would not be doing much work and blue screening right and left. One crash, yea that is still no proof of a corrupt drive. The HDD is still working and due to its age is one great drive and a test of its longevity, not that that was the test done.

 

A Floppy compared to a HDD is as valid as a HDD compared to an SSD, both are used for storage and can do the same things, again not as fast due to being different technologies. An SSD is not a stop gap but just the latest storage tech on the market that is faster than the previous generation tech (HDD's).

 

Basically,

 

Floppy < CD < DVD < USB Drives < HDD's < SSD's < Future Even Faster Tech

Edited June 27, 2014 by IdeaStormer

[Syntaxvgm]

Hard drives actually slow quite a lot over time mechanically. For a fair test it should have been a new drive as the SSD was, or an ssd as old as the hdd. Results still would have been similar, but not the same.

Also, what about trim? 

[The Benjamins]

To provide a fair comparison? The idea of this video was to show a true fair comparison, yet he used a very old HDD with a pretty much brand new SSD. Little bit unfair in my eyes so this would provide a level playing field for both devices.

It would also be quite useful to run these drives for YEARS to see how long it would take for each to eventually die

http://techreport.com/review/26523/the-ssd-endurance-experiment-casualties-on-the-way-to-a-petabyte

 

here is a good article that shows SSD lifespan, and the performance changes over time.

spoiler, not performance changes over its lifespan and can last over 800 TB of writes. my SSD for 2 years has 5 TB of writes.

[Tech Addict]

It's interesting to see just how much of a performance increase you get with a simple switch in storage technology. I have an SSD for my OS and main programs, and my HDD for more large-data type stuff, such as games I don't play too often. Oddly enough, I get a 30s boot, so I was quite bummed with that, since it didn't seem like there was much of an improvement, but the in-OS experience is nice, where a specific area I've noticed a big improvement in is the transferring of large files.

[bryannahangman]

Sorry I have a noob question and sorry if its been answered. What is the cache stuff for hard drives? Like 64mb cache on most hard drives. I would look for answers on Google but I trust Linus Tech Tip forum members more

[The Benjamins]

Sorry I have a noob question and sorry if its been answered. What is the cache stuff for hard drives? Like 64mb cache on most hard drives. I would look for answers on Google but I trust Linus Tech Tip forum members more

16-64mb

I know the samsung evo SSD have 512 mb

[Celmor]

I would have liked to see more tests in general and more tests that reflect the boosts you get working with the SSD and not things you could just click a button and go make a coffee while the computer works, just things that happen more often than rebooting or uninstalling with a desktop PC.
Tests I would have liked to see: starting 10 programs per BAT file.
Using the Windows Search in an area that isn't indexed (of SSD/HD).
Installing smaller Games with Steam.
I know that you won't get that big of a difference with the SSD/HD in the last to scenarios but it would be interesting how much the difference would be.

[rashdanml]

Worst case scenario tests are absolutely pointless - they tell you exactly what you already know before you do the test. 

There's a reason why scientific tests are repeatable; it eliminates all of the random variations from test to test. Worst case scenario tests aren't repeatable, and therefore not scientific, as every use case is totally different.

 

e: Also, this is far from a "torture" test.

[TheProfosist]

To provide a fair comparison? The idea of this video was to show a true fair comparison, yet he used a very old HDD with a pretty much brand new SSD. Little bit unfair in my eyes so this would provide a level playing field for both devices.

It would also be quite useful to run these drives for YEARS to see how long it would take for each to eventually die

It still would matter very little because the thing tripping up the HDD is not its age but the age and messiness of the data on it. Also consumer HDD design hasnt really progressed all that much for at least three years now, ever since we got 1TB per platter data density tech.

Other place have already done this sort of thing. It takes a rather expensive setup and it outside the bounds of what LTT seems to do.

I don't think that's being entirely fair to the HDD. Because one test crashed, the thing was probably on its last legs anyway.

 

While I did like this video, I think a best case scenario HDD versus a worst case scenario SSD would be much more interesting consumer advice though. You know, like dual short-stroked HDD's or even SSHD's in RAID0 (like I'm running in my own system) versus one SSD, for example. Or a short-stroked hard drive versus an SSD without TRIM support. That sort of thing.

Even if you cloned it it would only make a small difference as this is literally playing on the weakness's of a HDD and how it works.

I have a RAID 10 array of single platter 1TB per platter drives. Short stroking them would indeed increase speed but not by some huge number and your then decreasing the price to performance ratio the HDD, which is the one thing it has over the SSD. Current HDDs will never reach a SSD in terms of performance. I still use my Intel X25-V 40gb SSD in one of my systems and even that feels far peppier than any HDD currently.

Some people already said this in the comments but I think the test would be fairer if the HDD was also new. New thing beats old thing, that's not really surprising. If you tested that HDD against a brand new HDD, the new HDD would have been quicker.

 

 If you decide to redo the test have a new HDD and an SSD, copy the OS to both of them.

Thing is the SSD would only crush the HDD by every so slightly less, so it wouldnt really matter.

*snip*

 

There was no proof of corruption on the HDD (aka messed up drive). The HDD still works and if it was corrupt it would not be doing much work and blue screening right and left. One crash, yea that is still no proof of a corrupt drive. The HDD is still working and due to its age is one great drive and a test of its longevity, not that that was the test done.

 

A Floppy compared to a HDD is as valid as a HDD compared to an SSD, both are used for storage and can do the same things, again not as fast due to being different technologies. An SSD is not a stop gap but just the latest storage tech on the market that is faster than the previous generation tech (HDD's).

 

Basically,

 

Floppy < CD < DVD < USB Drives < HDD's < SSD's < Future Even Faster Tech

sorry when I said corruption what was talking about was all the stuff that had happened to it versus a clean install of windows. Call it a dirty install if you want.

I dont think a floppy being included is valid because it hasnt been used as a real form to run a current/modern OS off of in well years and years and thats whats being tested here. CD's and DVD's could never have OS's installed to them. There are however thing like linux live CD's. the performance of discs is actually downright horrible. You could do a OS install to a flash drive (not that this is something someone regularly does)or run a live CD. You will however fine that its has zero random performance minus the newer ones using SSD controllers. Also the endurance of the drives is piss poor.

[IdeaStormer]

sorry when I said corruption what was talking about was all the stuff that had happened to it versus a clean install of windows. Call it a dirty install if you want.

I dont think a floppy being included is valid because it hasnt been used as a real form to run a current/modern OS off of in well years and years and thats whats being tested here. CD's and DVD's could never have OS's installed to them. There are however thing like linux live CD's. the performance of discs is actually downright horrible. You could do a OS install to a flash drive (not that this is something someone regularly does)or run a live CD. You will however fine that its has zero random performance minus the newer ones using SSD controllers. Also the endurance of the drives is piss poor.

 

Well if Floppies are in valid then how can any tech introduced in 1956 be a valid comparison to an SSD? Yea HDD's were introduced in 1956! Freaking old crap. You can boot DOS off of a floppy just fine and some Linux's, same for CD's and DVD's so your lack of usage does not mean the rest of the world does not nor has not or could not. Flash drives are used to boot computers everyday! So again just because you haven't used it does not rule it out. Lastly you are extremely wrong on SSD endurance, how many tests do you need to validate their longevity? I'm sure you've heard of all the ongoing testing on many sites for this if not you should look into it, its happening and going on if not being proven day in day out.

[Nick Paviour]

So imaging the HDD to the SSD also copied the wear and tear of the HDD? Come on...

 

This was my thoughts exactly through the video, surely the SSD hasn't physically been through so much, therefore it would perform better?

Can anyone explain if I'm wrong. 

Edited June 28, 2014 by Nick Paviour

[TheProfosist]

Well if Floppies are in valid then how can any tech introduced in 1956 be a valid comparison to an SSD? Yea HDD's were introduced in 1956! Freaking old crap. You can boot DOS off of a floppy just fine and some Linux's, same for CD's and DVD's so your lack of usage does not mean the rest of the world does not nor has not or could not. Flash drives are used to boot computers everyday! So again just because you haven't used it does not rule it out. Lastly you are extremely wrong on SSD endurance, how many tests do you need to validate their longevity? I'm sure you've heard of all the ongoing testing on many sites for this if not you should look into it, its happening and going on if not being proven day in day out.

Yes but can you do a full windows 7 installation to a floppy. You should also take into consideration how HDDs and ssds are mainly used now days compared to floppies. It seems that you only skimmed my post instead of reading it. I did cover CDs and DVDs and their possible use cases and why it doesn't apply in this instance. Also by your standards we would have to install DOS on everything and bench it to be more thorough. I also said you could do a full install to a flash drive but it is not all that common and the endurance of the flash is the main issue along side the utterly horrible random performance. I actually use all of those things and more which is why I know about them. You need a lot of tiricks in your bag when your diagnosing computers and the more you have the better. What for SSD longevity yes there is one site that set something decent up and has been going on for quite some time. It been mentioned a number of times in this thread. Care to share another that's been properly done as well as ones for HDDs that are not just data pulled from a enterprise environment.

[TheProfosist]

This was my thoughts exactly through the video, surely the SSD hasn't physically been through so much, therefore it would perform better?

Can anyone explain if I'm wrong.

They never said really I assume its one they use on their bench rigs or the like. Either way with garbage collection and trim off the ssd would still easily beat the HDD.

[Bosh]

Stealing @Fate 's image, this is what I thought at looking at the video

So imaging the HDD to the SSD also copied the wear and tear of the HDD? Come on...

...

Did you watch the intro of the video?

They clearly say that this isn't the usual test, but a real life scenario where you have an HDD and you buy an SSD to clone the HDD, something very common these days.

And why isn't the comparison fair? People don't care about the theoretical limits of untypical scenarios, the comparison is fair as it goes along with the typical use case.

Try to be more open minded man, the video is useful and not an HDD-geek's fantasy, move on.

[rashdanml]

And why isn't the comparison fair? People don't care about the theoretical limits of untypical scenarios, the comparison is fair as it goes along with the typical use case.

Yet the test tells us nothing that we don't already know. 

HDDs are slower, period. (New vs. New)

HDDs are always dramatically slower (old HDD vs. any SSD) due to fragmentation and seek times. 

 

The theoretical best case scenarios are the most reliable to make the purchasing decision. A realistic, 5-years-down-the-line test is pointless, as people already know that replacing an HDD with an SSD makes the system several times faster, regardless of the amount of clutter on the HDD. 

 

Not to mention, a "torture" test is an endurance test. Nothing in this video tested endurance scientifically. 

 

It doesn't matter if the above scenario is common or not - the video tells us absolutely nothing that we don't already know. 

[TheProfosist]

Yet the test tells us nothing that we don't already know.

HDDs are slower, period. (New vs. New)

HDDs are always dramatically slower (old HDD vs. any SSD) due to fragmentation and seek times.

The theoretical best case scenarios are the most reliable to make the purchasing decision. A realistic, 5-years-down-the-line test is pointless, as people already know that replacing an HDD with an SSD makes the system several times faster, regardless of the amount of clutter on the HDD.

Not to mention, a "torture" test is an endurance test. Nothing in this video tested endurance scientifically.

It doesn't matter if the above scenario is common or not - the video tells us absolutely nothing that we don't already know.

What there are trying to show is that a ssd has even more of a HDD the more mucked up the HDD is. As slick said its just something they thought up after finding his drive.

[rashdanml]

What there are trying to show is that a ssd has even more of a HDD the more mucked up the HDD is. As slick said its just something they thought up after finding his drive.

... which everyone already knew.