Do all SSDs have the same performance (practically)?

[LTG]

Has anyone ever experienced a difference in speeds in real life usage between two different SSDs?

File copy performance I can see being different, but all this random IOPs stuff - has anyone really noticed a difference outside of a benchmark?

 

[LokutusofBorg]

I've personally only used a Vertex 2, but it was so fast compared to mechanical hard drives, that I am sure if I'd used another SSD I wouldn't have known the difference. Look at the benchmark charts in any of the reviews that include a mechanical hard drive and you'll see how far in the dust SSDs leave HDs.

When you've been driving 25MPH and you switch to 55MPH are you going to notice the difference? Yes. If you then switch to 60MPH are you going to notice? Not really.

 

[razel]

For 95% of people, yes. For the remaining 5% it's only during specific, particular times when you are pushing alot of data from one drive to another primarily because that is the only time I can think of where we even come close to reaching a particular SSD's speed limit which is what benchmarks try to figure out.

However if your system is fast enough you might be able to tell a difference in latency between drives. On HDDs with daily computer usage, the HDD is the overall bottleneck. With SSDs, integrated video cards on Win7 systems with Aero or latency in your chipset becomes the bottleneck. I fully understand that Aero effects complete at the same time. The only difference between a slow video card and a faster one is more frames that are seen. It's the shared memory nature of integrated graphics that increases chipset latency in a heavily multi-threaded Win7 universe. If you're lucky enough to have a speedy CPU you might be able to shift the bottleneck back to the SSD and that's where some very lucky few of that 5% maybe able to tell a difference.

In my experience, when I had a 3.6 GHZ OC'd laptop, I had the 1st generation Kingston back in 09. It's write latency was worse than most. When the Intel G1s (king of latency at the time, now it's a dying king ) came down in price I upgraded and noticed things felt smoother, but not necessarily fast enough that it'll save me even a second. I would have been happy with either.

File copy performance is a great marketing tool. Those numbers look great on flash ads and it's easy for anyone to replicate... just copy a a file. However when you copy a file what you're reading from has to be fast enough to keep up with the SSD and that's the only time when I ever make use of just how fast SSDs can be. During a SSD to SSD clone. 17GB wrote in less than 1 1/2 minutes. However, that is something that we do not do everyday.

Random IOPs is also marketing. 10,000 IOPs is an impressive number but means nothing to the average user since they don't even come close to making that many IO requests. If you were using it for a server, then it'll matter, however at that point if you are a server admin in this economy, you don't toy around in a production environment and there are far better, easier, established ways to improve performance.

 

[LTG]

nice answer, thanks-

 

[George Powell]

I've got 2 SSDs myself, the first is an OCZ Vertex 60GB (The original Indilinx one) which I use in my HTPC (Core i3 2100). The other is an Intel X25-M G2 160GB which is in my main desktop (Core i7 860). I've also added an OCZ Vertex 3 120GB to a friends PC (Core i7 920). From my personal experience I am not able to perceive any substantial difference between the 3 different drives. Admittedly all 3 drives are in quite different setups so I can't really compare them directly and I've not run any benchmarks on them at all.

 

[imaheadcase]

Pretty much what razel said.

If you work with a certain app you could see a big difference. But for %95 all of them are pretty equal.

 

[Eug]

This is my slow 128 GB SSDNow V Series, booting up a MacBook Pro. The SSD is a 2 year old entry level model, and it does not support SATA 3. The laptop is a 2 year old Core 2 Duo 2.26 GHz, with 2 GB RAM.

http://vimeo.com/27951421

 

[randomlinh]

i'm going to go with reliability should probably be looked at first. though, my friend just had a brand new intel 510 die on him =/ but thanks eug, now only if they still had some of those clearanced somewhere.. i could just raid 0 em, heh

 

[Eug]

i'm going to go with reliability should probably be looked at first. though, my friend just had a brand new intel 510 die on him =/ but thanks eug, now only if they still had some of those clearanced somewhere.. i could just raid 0 em, heh

Does not compute.

The good thing about the V Series and the V+100 is that they do heavy duty garbage collection, so I suppose they'd work well in RAID. (TRIM isn't supported in RAID configs.) I don't know if the same is true for the V100 though.

---

As you can see from the video in my previous post, my 2.26 GHz Core 2 Duo MacBook Pro absolutely flies with the "slow" V Series SSD installed.

I put the similar ballpark speed V100 SSD in my 1.3 GHz dual-core Pentium SU4100 laptop, and it makes a huge difference. However, I wouldn't say it "flies". App loading is relatively quick, but the machine doesn't feel anywhere near as fast as the MacBook Pro. The CPU limitation of the Pentium SU4100 really becomes noticeable. Still, the perceived speed is pretty decent with the SSD installed.

To put it another way, my C2D MacBook Pro feels faster than a desktop with 7200 rpm drive, but the Pentium SU4100 laptop feels slower, albeit with faster app loading.

 

[groberts101]

Look at it like this.. if lowend entry level SSD's were just as fast as the top line models in actual usage?.. this place would be full of gloaters who saved a fortune over all the e-PEEN's who only concerned themselves with benchmark specs before buying. lol

Not the case at all and is EXACTLY like speculating how a 60GB drive "feels" compared to a 240GB drive. Is the casual observers/users perception that both are SSD's?.. well of course. Are they just as comparable in moderate workflows/multitasking?.. ummm.. no.

Speed cost's money.. and you are only given what you pay for in this market.

 

[Eug]

In many moderate workflows and most multitasking I'd say a lower end SSD is fantastic. Where the lower end SSDs tend to fall down more is big sequential file transfers, but the key here is that even an entry level SSD will still have a seek time of say 0.1-0.2 ms. This is an order of magnitude better than platter drives, which is why even entry level SSDs feel lightning fast.

Hell, I just installed an entry level SSD in one machine... from 2009... and the machine just flew. Soooo fast. I then returned the drive however, NOT because of the speed, but because of the power usage. I got a new one with better power characteristics.

So yeah, if you're paying a lot extra to get a "500 MB/s" drive for stuff like business multitasking, then I'd say you're wasting your money.

IOW, for my laptop use, I'd much rather spend more money on things like low power utilization than things like raw transfer speeds.

 

[CFP]

I'm not sure to believe myself, but I could swear that my PC with C300 64 as boot drive w/ Win 7 feels slightly snappier than X25-M G2 160 w/ Win 7.

 

[Eug]

I'm not sure to believe myself, but I could swear that my PC with C300 64 as boot drive w/ Win 7 feels slightly snappier than X25-M G2 160 w/ Win 7.

Well, it could be true, but it's not like you're like saying "OMFG it's the second coming of SSD!!1", which is what some of the posters were getting at.

 

[groberts101]

I hear ya from the mainstream user's standpoint... but from the perspective of having raids on top of raids with raidcards and all.. I pay money to recoup time whenever I use my system. Time is far more important to me than a little more money/energy saved.

And I have in fact tested enough SSD's to see and "feel" the distinct difference between 30-240GB SSD's. Capacity = extra channels and better I/C config's for much stronger small file performance where the OS likes to run.

Once you start experiencing seemless and efficient transfer's between at least 2 SSD's(OS) and 4 drive HDD array's?(maybe 500MB/s transfer's with major low end grunt for the OS.. then you start raising the bar to what you think is fast enough anymore. It just happens naturally and what one man thinks is fast?... is another man's "backup machine" or "kids computer".

While the main benefit of reduced latency is similar enough between the smallest to highest capacity SSD?.. the actual speed and loading capabilities go up quite noticably. Don't need to be comparing to raid speeds to see it either as one is just clearly better than the other, is all.

Then there's the whole capacity = write stamina thing which only becomes obvios over a little time/usage.

Then again.. if you hardly use or task the system? Like you said.. can be wasted money with diminishing returns.

 

[Eug]

Well, "raids on top of raids" is hardly relevant to most people here concerned with "moderate workflows/multitasking".

 

[groberts101]

maybe so.. but the point was that my single Vertex 3 240GB(Sandforce) is noticably faster than any of my 30GB Vertex(Indilinx) drives.

Don't even need to be a raided gear-headed speed freak to "see" it and even my wife saw/see's the difference and made me put a larger-faster 120GB drive in here laptop for that very reason. And she doesn't task a system nearly as hard as I do(which was my wasted arguement against buying another. lol

In the end.. the percieved comparisons between capacities and controller types may be slightly different for some folks?.. but speed = time and bigger drives give you more back. Speed cost's money.. how fast you wanna' go?

Then there's.. what's your time worth?

Anywho.. to each their own but I would never recommend buying an SSD below 100GB's unless you don't plan to push lots of data across it and just use it lightly without a gazillion u-tube video's streaming into the Windows buffer and filling your drive up too quickly. I started out that way(30GB Vertex) and it wasn't nearly as pretty as using the 240GB Vertex 3 drive. That's the thing about SSD's.. all controllers.. and all capacities.. are NOT created equal. Still don't believe me?.. go buy a $70 dollar SSD and then compare it to a $200 SSD. Then to a $500 one.

and that's the question I answered based on my experience so far.

 

[Eug]

64 GB is perfectly fine for some people. 96 GB is also perfectly fine for many more.

BTW, I bought an $80 64 GB entry level SSD, a $150 prior generation entry level 128 GB SSD, then bought a $190 current generation middle-of-the-road 128 GB SSD.

In real world usage they're pretty much identical in speed. It's just that 128 GB is bigger than 64 GB. Most important are the power characteristics IMO, for everyday type usage in a laptop.

Ironically, in some benchmarks, the $190 one is slowest of the three, but like I said I'm more interested in power characteristics than benchmarketing.

 

[gramboh]

I've been thinking about this recently, I want to get an SSD in my next build (probably wait for Ivy Bridge) and am willing to spend $250ish. If I were going to get something today, I'd be choosing between current generation SATA3 120GB drives for $225-275ish and previous generation 240GB drives (Vertex 2 for example) for around the same price, or a bit less. I think the space is more important to me than the extra speed, I'd like to fit my entire Steam game collection on the SSD and not have to manage with linking/junction points.

We will see what happens to prices in the next 8 months.