Can Someone PLEASE Summarize What SSD Benchmark Numbers MEAN?

Ok, It's clear not all SSD's are created equal. But I'm having a hard time digesting what the benchmark numbers actually mean for real-world utilization.

Do I really need to care about 4k read/writes? Do sequential read and writes even matter? I mean maybe for that occasional huge ISO file you want to move or copy? But what about installation of programs, windows boot times, file copies, extracting or compressing data (like a zip file). What about virus scans? And what is queue depth and how is it really used and does it matter? Can someone please explain what each of these things mean itemized? I don't mean technically but I mean in a real world sense. It would really be helpful:

Sequential read =
Sequential write =
512k read =
512k write =
4k read =
4k write =
4k-QD32 read =
4k-QD32 write =
4k-64thrd read =
4k-64thrd write =

Right now I'm debating between the Kingston V-series SSDNow 64GB and Vertex 2 60GB. It seems that the Vertex 2 should be a no brainer but based on this article at TechSpot originally posted by Phil, it seems the Kingston competes with the Vertex 2 and Agility 2 in read and file copy performance. What that article DOESN'T cover though is any amount of write performance like OS and application installation.

I plan on buying two SSD's, one for my laptop and one for my desktop. The Kingston can be had for $95 after rebate and Vertex 2 for $125 after rebate, so in essence for me there's a $60 difference, which is pretty substantial. But if I later realize performance is below my Seagate Momenuts 7200 hard drive by any significant amount in any respect, I won't be too happy.

My other hangup is power consumption. From what I've read the OCZ's sip power even while reading/writing, but I can't find any article that shows actual power consumption of the Kingston drive, but specs show it has a 5.2W active power consumption compared with 2.0W with the OCZ. Significant difference.

My biggest complaint is that as expensive as SSD's are there is very little useful information out there, especially from manufacturers.

 

From what I gather, in general, 4K or random R/W speeds are the most important values for SSDs (since all have similar access times). These speeds will be used for OS loading and give you the general snappiness that SSDs are regarded for.

Sequential speeds are used mainly for file transfers and unless you copy big files such as movies daily, are more of a marketing tool. Also, speeds will have a diminishing return of effect, so just because something is 120MB/s vs. 80MB/s, doesn't mean it will be 50% faster.

From what I've seen, Samsung/Indilinx/Toshiba are fast enough (substantially over HDD yet not noticeably faster in most day to day uses when compared to the next group) while consuming very little power and being the most economical. The Intel/Sandforce/Marvell (Micron C300) are the fastest of the fast but consume a bit more power and are a fair step up in cost.

So if you're on a budget, the former three are excellent choices, but if you want the best of the best, no doubt the latter three will be what you want.

 

Thanks. According to the specs though, the Kingston (Toshiba controller?) states 5.2W active power consumption. The Vertex 2 (Sandforce) says 2.0W active power. So seems opposite what you're saying.

 

I'm basing my results on what I've seen personally and what I've read from users (mostly from this forum, specifically the long SSD threads) with the above aforementioned SSDs. Specs are garbage for the most part IMO since they are notoriously inaccurate.

 

after all of the recent benchmarking work done someone finally asks 'what does it mean'.

 

What do SSD Benchmark numbers mean?

Summary: Nothing.

How I 'test' a new component is not in a vacuum (nor with a random program that posts only the 'best' scores) - I test it by comparing it to a fully configured and running system.

None of the numbers that benchmarks throw at us compares to how an SSD 'feels' in comparison to a mechanical HD.

My Inferno launches PS CS5 in 6-7 seconds with the registry tweaks applied. My Momentus XT launches the same program in 3-4 seconds. This is with an identical install (C: ) between the machines in question.

Then again, the Inferno scans 2.5 Million items with MSE in around 90 minutes - the XT takes a little longer (around 4-6 hours) for 6.5 Million Items.

I'm sure people can summarize what the numbers mean in benchmarks - but even summarized, defined and categorized, they still will have no correlation to the real world. Unless, the 'real world' for your specific computer usage is running these same benchmarks, over and over again.

 

That's the problem. You shouldn't have to spend $500 on an SSD to test it yourself only to find out it doesn't perform as expected and you can't return for a refund.

My point is that there needs to be better established numbers to determine what really works for your needs.

 

htwingnut,

I agree 100% with you - but the problem is the reviewer's that focus on the simple and very narrowly focused benchmarks - or the simple and very narrowly focused 'real world' scenarios that they then extrapolate to 'everyones' real world scenario.

I've stated this a few times: currently, the Inferno (and I'm assuming any SandForce based SSD) is the best in terms of 'feeling' that you got an upgrade over a 7200 RPM based system. Whether it is worth the $$$$$ premium is up to each of us to decide.

Where there is no comparision is the 'robustness' of the SSD's: I wouldn't use a mechanical HD like how I use the Inferno based system as I do.

 

Well I ended up ordering a 60GB Vertex 2 for my Sager laptop and a 64GB Kingston V-series for my desktop. In both cases I will have a secondary HDD for storage and games, and will probably put a couple games on the SSD that would benefit from it. To supplement, a Momentus XT in my laptop and WD Caviar Blue in my dekstop of which I already own. Hopefully that should suffice.

 

I don't know what the numbers mean but I can tell you this: I have an Alienware m15x at work that I use for Excel, powerpoint and doing Chemical Engineering material and energy balances and I just installed a 250 gig OCZ Vertex in it. At home I have a second M15x with a Corsair force 240 gig and they both perform exactly the same in the real world. No noticable difference to me. The OCZ seems like it may be a little snappier loading my game levels sometimes but thats it.

 

If that is an OCZ Vertex 2 that is essentially the same as the Corsair Force. Both Sandforce controllers.

 

Regular Vertex but they still feal the same to me. I don't really do anything more intensive than open and close files and play some games though.

 

That original Vertex is an Indilinx I think. Should be similar performance to Intel, so still a great drive.

I was going to buy a 256GB SSD for my laptop, then I thought better of the $600 investment and go and buy the 60GB with a Momentus XT for considerable cost savings and storage I need, along with significant performance boost. Then I went and bought a Kingston 64GB for my desktop and one for my netbook. Still ahead by like $200!