OCZ SSD Test panel - Two Year Test Run Comes To An End
Two years ago there was a lot of discussion about the endurance of SSDs, failure rates and so on. OCZ at the time had to fight off a bit of a reputation and claimed that their SSDs had extremely low failure rates. At that time Guru3D proposed an idea to OCZ, give us 10 SSDs, we’ll ship them towards 10 forum users and we’ll keep track of them for two years.
The two years have now officially passed, and a lot of things happened (while nothing happened with all SSDs), OCZ is no more, they have been purchased by Toshiba. Toshiba is still selling OCZ SSDs as part of their branding. Yeah it is amazing how much has happened ever since we started this test in May 2015.
We have continued and monitored the the test panel group for two years now. And if you want to have a peek, feel free to look at the results from the test panel in theforum linkage below. All OCZ Vertex 460A SSDs are still up and running, none of them have had a failure and all of them show good to close to advertised performance values. Most panel members even have 98% lifetime left on their SSDs despite using and abusing their SSD as daily OS drive and some have written A LOT of TBs.
Personally I have to say that is has been a pleasure watching and following the test panel as it also confirms my strong believe in NAND based storage solutions being durable, long-lasting and trustworthy. Pretty much all panel members have been able to post their scores and benchmarks on a monthly basis dureing these long two years, much thanks for that. Albeit OCZ might not be around as a standalone company anymore, we do like to thanks Tobias and Bernd from (back then) OCZ for having the guts to take part of this test as really it was a ballsy thing from them to agree to. I think it is safe to state if you invest in a proper SSD, it will last you for many many years.
Have a peek at the latest updates, some panel members still need to add the final May 2017 results though.
Going strong even after two years of usage
OCZ SSD Test Panel Up and Running - following 10 people for two years - 05/08/2015 02:15 PM
A few weeks ago we asked a handful of Guru3D readers if they would like to participate in an SSD group test, for two years. The past week not five but a selection of enchanted ten Guru3D readers have ...
Would you like to join the Guru3D: OCZ SSD test team ? - 04/10/2015 06:28 PM
Would you like to be part of the Guru3D SSD test slash focus group ? Over the years we a have tested a lot of not all products from OCZ Technology. Problem is their rep got a dented with some Sandforc...
OCZ SSD prices at an all-time low - 10/17/2014 09:47 AM
OCZ is letting press know that prices for most of their consumer SSDs have dropped by up-to 30%, in an attempt to be even more competitive with parties Samsung and Micron. As a result, re-sellers are...
OCZ SSD sales increases revenue - 03/08/2011 12:02 PM
Good news for OCZ, after a dissapointing year OCZ announced record revenue for its fiscal 2011 fourth quarter of approximately $64 million, almost double as much as the $32.4 million reported in last ...
Senior Member
Posts: 7435
Joined: 2012-11-10
I use SoftPerfect RAM Disk, at the time I did a fair bit of research into the options.
SoftPerfect RAM Disk allows unlimited size + number and also has persistence options etc.
Sounds vaguely familiar. I looked into it a little and I can see why you picked it. If you do consider using Linux, I'm sure you'd like the options it has for RAM drives. You can mount them anywhere you want (so for example, I have my Chrome cache stored on a RAM drive) and you can create ones that dynamically expand. It's hard to find solid evidence of how fast they are (most disk benchmarks are bottlenecked by the CPU), but they're roughly 2GB/s per memory channel.
Its an interesting option, though atm it is still expensive and rather slow, compared to RAM drives.
Agreed, but I was thinking it might be more useful for things that might be better with non-volatile memory, such as web browser caches or maybe hibernation files. I say "better" because the sole purpose of such caches is to be able to reference them later. If you have no need to when rebooting, you don't need the cache in the first place.
FF's net cache is already on the RAM drive, I had not considered disabling before, but it probably performs better with it as I have it.
The large amounts of writes seem to come from FF's core functionality including saving a persistent record of all open windows and tabs etc
That's good you have it on the RAM drive, that certainly will help spare your SSD from such a beating. But like I said earlier, if the data isn't kept when your browsing session is over, the cache doesn't really serve a purpose. FF already retains its current activity in memory with or without the cache, so unless you find yourself re-opening FF without rebooting (whether from crashing or intentionally closing it), your RAM drive cache is really doing nothing but taking up CPU cycles and using up your RAM drive storage for other tasks. If you removed the cache entirely, that's an entire function that FF is no longer performing, without having any negative impact on your current browsing session. Does that make sense?
Just so you know, I use FF on one of my PCs and I keep the cache completely disabled. Always runs very smooth for me, even with 20 tabs open.
Senior Member
Posts: 959
Joined: 2009-10-14
Sounds vaguely familiar. I looked into it a little and I can see why you picked it. If you do consider using Linux, I'm sure you'd like the options it has for RAM drives. You can mount them anywhere you want (so for example, I have my Chrome cache stored on a RAM drive) and you can create ones that dynamically expand. It's hard to find solid evidence of how fast they are (most disk benchmarks are bottlenecked by the CPU), but they're roughly 2GB/s per memory channel.
I am getting about 2,500MB/s per channel and running triple channel. So about 7,500MB/s on the RAM Disk in HDD emulation mode.
I must say that it is Impressive that that Liquid Element PCIe SSD can nearly match my RAM Drive emulated HDD.
Curious how much one of those will cost

Agreed, but I was thinking it might be more useful for things that might be better with non-volatile memory, such as web browser caches or maybe hibernation files. I say "better" because the sole purpose of such caches is to be able to reference them later. If you have no need to when rebooting, you don't need the cache in the first place.
1st thing I did on installing the OS was doing away with the hibernation file.
The cache is an interesting point, but as my PC runs 24/7 for weeks at a time and FF still currently grows the pagefile to 10.5GB+ on about a 1.5 day basis, I restart it at that point.
FF takes Ages (3mins++) to become usable while opening/restoring windows/tabs. Just checked I have 28 windows (active tabs) for it to restore, I usually have about half that. During that 3 mins there is minima CPU and HDD usage and other browsers can open pages fast just fine. Seems FF could improve performance there a lot.
Is fine once it is up and running though.
...about a month back FF had no large memory issues or large pagefile usage. Performed Brilliantly actually, I could open a near unlimited amount of tabs with no issue - I literally opened Hundreds to see how far it would go. I was Very Impressed. Though this situation only lasted nearly 2 weeks before degrading back to normal.
That's good you have it on the RAM drive, that certainly will help spare your SSD from such a beating. But like I said earlier, if the data isn't kept when your browsing session is over, the cache doesn't really serve a purpose. FF already retains its current activity in memory with or without the cache, so unless you find yourself re-opening FF without rebooting (whether from crashing or intentionally closing it), your RAM drive cache is really doing nothing but taking up CPU cycles and using up your RAM drive storage for other tasks. If you removed the cache entirely, that's an entire function that FF is no longer performing, without having any negative impact on your current browsing session. Does that make sense?
Will do some testing without the cache if they make it's memory/pagefile usage better again soon.
Just so you know, I use FF on one of my PCs and I keep the cache completely disabled. Always runs very smooth for me, even with 20 tabs open.
I still find FF the best browser out there for a variety of reasons.
I back that up with MS Edge for compatibility with various sites, but it is APPALLING at memory usage with even low/medium amounts of tabs open lolol
The failings of various browsers is pretty embarrassing really in this day and age.
Senior Member
Posts: 959
Joined: 2009-10-14
Seems to me you're going to need to move onto DDR4 at some point to take advantage of the higher capacities. Unlike most people, it seems you actually have a reason to go this high (or higher). Though out of curiosity, what exactly are you using for a RAM drive? There are several approaches so I'm just wondering.
DDR4 will be a thing when I upgrade to my next system, waiting to see how Naples turns out atm.
I use SoftPerfect RAM Disk, at the time I did a fair bit of research into the options. A lot of the other programs have actually pretty poor performance especially when in disk emulation mode, this one is about one of the best performers (and is free for non-commercial use) without spending a fair amount of money on one of the better commercial ones. (quite a few commercial ones are also poor performers btw)
SoftPerfect RAM Disk allows unlimited size + number and also has persistence options etc.
If your system is compatible, I figure Intel's Optane would be more useful for a paging file (though I'm not entirely sure you can use it that way).
Its an interesting option, though atm it is still expensive and rather slow, compared to RAM drives.
How much do you really need FF's cache? I figure you'd be better off just disabling the cache entirely. With data writes that high, it seems to me FF is spending more time writing the data than it is reading from it, in which case you're losing performance while degrading your SSD(s) at a rapid rate.
FF's net cache is already on the RAM drive, I had not considered disabling before, but it probably performs better with it as I have it.
The large amounts of writes seem to come from FF's core functionality including saving a persistent record of all open windows and tabs etc
Linux has a lot of tight integration with many VMs, either as a host or a guest. Whether or not it is a good idea depends on a lot of things. The main reason I suggested a VM (as well as containers, such as Docker) for you was because Windows doesn't really have an elegant or efficient way to manage groups of applications, nor does it have a way to temporarily suspend an application without losing progress. For your needs, if a VM or container is involved, I suspect you'd have a better experience using a Windows host + a Windows guest. It'd offer a more seamless experience. If Linux is capable of fulfilling your workload (which I suspect it would be) you'd be better off running it as-is, without a VM or container.
Just anouther thought why you may want to consider Linux, you can use a GPU's VRAM as system RAM:
https://wiki.archlinux.org/index.php/swap-on-video-ram
Feel free to PM me if you want more info/help on this.
Might possibly give you a shout about that at some point, thanks for offering
EDIT:
SoftPerfect Ram Disk is no longer free for home users since version 4.0
Version 3.4.8 is the last version fully unlocked for home users.
I am currently on version 3.4.6 and it has been very stable for me.