Main Area and Open Discussion > Living Room
SSD's - How They Work Plus Tips
mouser:
I guess a more accurate statement would be: "All other things being equal, faster spindle speeds nearly always mean better the performance."
And then adding something about what other factors have the greatest impact on real-world performance.
If you were trying to give someone some quick rules of thumb for choosing fast drives, i still say it makes sense to tell them to look for fast drives and to not be too distracted by cache sizes, and "max" throughput rates achievable by the card. Would you agree with that? Can you formulate a simple way to tell people what to prefer in terms of density/plattercount?
Or maybe you think the best thing to do is to tell people to ignore all such stats and find a site that will compare drive performance explicitly for the drive you are considering purchasing.
ps.
As someone who has become paranoid about drive temperatures (in terms of fearing a hardware failure due to overheating) -- i tend to focus FIRST not on speed but on reported drive temperatures.. If I find a HD with great performance but which the reviews say runs hot, i run away.
Jibz:
Perhaps you are expecting too much technical accuracy from a side argument in a consumer article in a consumer magazine :-[.
Personally, I find the title more misleading -- "proper care and feeding" basically boils down to "it takes care of itself, just buy a big one".
40hz:
A quick rule of thumb without getting too technical?
General best performance = SSD
Best current $/capacity ratio = standard HD
Pick your trade-off point. (Too soon to tell if those hybrid drives will eventually hit the sweet spot.)
Stoic Joker:
that sounds like a reasonable statement to me-mouser (May 14, 2013, 12:47 PM)
--- End quote ---
Me too ... But I'm not a hardware guy either. I was under the impression that (seek time being roughly fixed) latency was reduced by the higher spindle rates...which resulted in faster access times.
-Stoic Joker (May 14, 2013, 02:40 PM)
--- End quote ---
It's not so much he's wrong about 10,000 RPM being faster than 5400RPM - depending on what is meant by "faster.". But it's a mistake to simply equate raw spindle speed and cache size with disk performance, which is what he seems to be implying.-40hz (May 14, 2013, 03:14 PM)
--- End quote ---
Seems like it's more a case of what the crew is inferring...
Partitioning, I/O distribution on the disk, I/O bus width, cluster size, filesystem(s) used, and several other factors have a much more direct and measurable effect on overall performance than just the spindle speed or cache size.-40hz (May 14, 2013, 03:14 PM)
--- End quote ---
...Yes, but these are all factors that are external to the central moving parts vs. non moving parts theme. As these are all factors that could effect either design by a users bad install, or by using a cheap MBoard.
If as mouser eludes all other possible random factors are fixed as accepted equals, and the distinction is narrowed to pro/con of moving vs. non moving designs ... In that context is makes sense. But then again I like the lower speed drives because the bearings last longer if they spin slower, and I've gotten stuck arguing with some sales-tard at BestBuy enough times that persisted in pushing the issue that the 7200+ RPM drives would be faster that I'd have to assert that it is a rather popular (miss)conception..
40hz:
...Yes, but these are all factors that are external to the central moving parts vs. non moving parts theme. As these are all factors that could effect either design by a users bad install, or by using a cheap MBoard.
If as mouser eludes all other possible random factors are fixed as accepted equals, and the distinction is narrowed to pro/con of moving vs. non moving designs ... In that context is makes sense.
-Stoic Joker (May 14, 2013, 05:52 PM)
--- End quote ---
Agree, but only up to a point since it's generally pointless to argue about specifications outside of the environment they're being applied to. Computers are systems, so there's no such thing really as something that exists in complete isolation as purely hardware or software. Every system incorporates both. And it's a balancing act as we all know too well.
But my professional activities deal mostly with "real world" performance issues. So I'm somewhat biased against absolute specs and biased towards practical suggestions and recommendations. And that usually means not reading too much into published specs or artificial test results.
;D
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version