Test 2: Streaming Reads
We made a statement in the initial article that turned out to be wrong. We said that hard drives would consume close to idle power for sequential read operations. This is true for many 3.5” hard drives, where mechanical activity requires more power than in 2.5” notebook drives. Unfortunately, we did not take the time to repeat that analysis with 2.5” drives. In fact, notebook drives oftentimes reach peak power when they have to provide high bandwidth data streams.
Let’s look at the benchmark results for our streaming read IOMeter pattern, which applies 100% sequential reads at 64-KB, 128-KB and 256-KB block size in an equal proportion.
All of the Flash SSDs with the exception of SanDisks’s slower, but more efficient, SSD5000 outperform the mechanical hard drives when it comes to delivering throughput. More I/Os per second translate into more blocks of 64, 128 or 256 KB information per second, which again translate into higher data transfer rate. The results are very different for writes, but since read activity typically dominates over writes in desktop environments, we do not dig into it at this time. The Samsung HDD, which is one of the best performers when it comes to throughput, shows the best result for the mechanical hard drives. And again, OCZ is the one to beat everyone else in performance. Time to look at the power consumption for streaming reads:
Clearly, high bandwidth streams cause all the hard drives to reach their peak power, although many Flash SSDs utilize more power for sequential access. Only the new Seagate Momentus 5400.5 can beat the Crucial and Mtron Flash SSDs. The Super Talent MasterDrive does well, but the real winner is the new OCZ Flash SSD again. It only requires 0.6 W at maximum sequential throughput, and it delivers the highest bandwidth by doing so. Remember that the SanDisk performance is disappointing, which reflects on the performance per watt results.
As expected by many of our readers, a Flash SSD provides better performance per watt when operating at maximum sequential throughput. However, the only drive that actually showed mind-blowing results is the OCZ SATA II 2.5” SSD, which we did not have when we compared SSDs for our initial article.
- Storage,
- Business Storage,
- ssd ,
- hard ,
- drive
- 1 / 2
- Next
-
LOL
Flash SSDs do not inherently contribute to increasing battery life and better efficiency comes with the appropriate Flash SSD used for a specific application. "Flash SSD" is not a qualifier for efficiency or performance.
This is much better than the first article now. Specially the DVD playback is interesting.
The 64GB OCZ seems to be an amazingly exceptional drive.
Hahaha. Once again contradicting yourself as Xeus pointed out.
Kudos to Tom's for having the decency to admit their mistakes and revise the article. The results, while not extraordinarily different than the previous article (discounting newcomers) provided a much more complete picture of the advantages and disadvantages of SSDs.
Articles like this one are a necessary step to regaining some of the lost credibility from previous articles like in this scenario. (Though it would be better if the articles didn't need revision) In short, it is encouraging to see that Tom's is listening to the readers. IMHO the real enthusiasts are more concerned with well thought out procedures, and accuracy/completeness/comprehensiveness of results than reading the "We posted first" articles that are all to easy to find on the internet.
iLOL
I like when people talk about the power inefficiency of hard-drives,yet don't blame things like CPU's,or discrete GPU's in their laptops.
However,I would like to see this as a high power part in desktops sooner or later,but with of course more power and(hopefully) stellar reliability.
And I mean for under 300 dollars.
iLOL
I like when people talk about the power inefficiency of hard-drives,yet don't blame things like CPU's,or discrete GPU's in their laptops.
However,I would like to see this as a high power part in desktops sooner or later,but with of course more power and(hopefully) stellar reliability.
And I mean for under 300 dollars.
Is measuring only wattage correct? What about the amps? I'm no electrical/mechanical engineer but I have always thought you need to measure both wattage and amperage to get the actual power consumption?
Dragunover - no one is refuting the fact that the CPU/GPU will be power hungry... that's not what is under discussion here. i'm interested by the fact you seem to be asking for SSDs that consume more power...?
Ziria, Wattage (P) is calculated by multiplying Voltage (V) by Amps (A).
Hence, the only important number is Wattage.
This article demonstrates that too often people generalize from specific test data.
Whether SDD or HDD, how you use your electronics device will influence how it much power it consumes. To conclude that one drive uses less power because a specific test does is an inaccurate conclusion - the testor can only state that it uses less power under those given circumstances and that is most likely will hold true under similar circumstances. (For example: this test is for random read/write operations, so the results most likely will be true for most applications that follow that pattern of disk usage.)
My point? Many times reviewers do not have the time to evaluate a product in depth, and thus they should not make broad generalizations or exaggerated claims.
Hopefully Tom's will keep up with in-depth reviews for a variety of users - road warriors (office apps/web/email), multimedia entertainment (DVD/video streaming apps/music playing), gamers, and "typical home users" - and creating good summary tables of their findings.
Myself, I find the IOs/Watt information especially valuable. I'll use that to find the IOs/Watt/$ when evaluating my future purchases. I'm kinda curious how the SSDs stack up against each other in that category (and I'm sure the HDDs are still waaaaaaaaaaayyyyyyy ahead of the SSDs right now).
to me teh argument isnt so much a "efficency" issue , to mee flash SSd's represent a future in computing where we don't have to replace , storage devices every 5-6 years (this is assuming you are getting agood high quality hard drive ) old mechanical hard drives (the current standard) fail after about 5-6 years to date the longest livinf hard drive i have is at 6 and i'll be amazed it it last half another year./ Also note you can cut this hard drive life expectancy in half if you smoke around your computer or you ahve a dusty home like many peopel in more rural areas. This hard drive death is soemthign that can't be stoped as the metal paltes require mechanical parts to spin them , and like any motor the parts will wear outa nd eventually stop working all together. Now fast forward to when they have SDD's that range in a better size (like at elast 250 gb) i can see these drive replacing even desk top hard drives , simpply for teh reason ... They HAVE NO MECHANICAL PARTS THAT WEAR OUT !
Not true. Flash SSD's suffer the same flaw as your typical Flash "thumb drive"... The recording medium fails after X number of writes. They've added technology in the controllers to extend the drive's life by spreading the writes around the drive's landscape, but the drive WILL fail eventually... even if you treat it with kid gloves.
Wow. Pretty rare these days to see a big publication go out and admit a mistake and even make a big public apology. kudos to you guys.
hard disks have been here for ages.. give ssds some more time and only then we can actually evaluatethe differences..
Wow, I have to admit, I did not expect Tom's to admit their errors in the last article. I am glad they did though, and this one looks to be much more along the lines of what we would expect to see. Props for having the gut to admit your mistake Patrick and Aaron.
All that really matters to me in regards to these articles, is that i'm not going to notice any difference in battery life while using my lapop. I could however expect to enjoy the speed if i got a good SSD.
Nice job Tom's. I thoroughly enjoyed this article a lot more than the previous one. It seems like the last one way maybe rushed. Either way, thanks for keeping things legit and up to (at least my idea of) Tom's quality standards. Thanks again!
so where are the benchmarks for OCZ Core Raid-0?
It's pretty simple. Solid state storage power can be turned off while in use in between read and writes. Hard drives have to continue spinning, or their access times would be seconds. I think the future is pretty clear, and there is no "hoax." Early hard drives were sub-par as well. If you release garbage stories like this, you'll hurt the SSD market and it will take longer for fast efficient SSDs to come into play. This is more like a retraction to an advertisement trying to slow the SSD market than anything. Methinks Tom's Hardware has been bought and paid for by the very companies it reviews, even going so far as to post articles and videos about integrity among video game reviews, to paint themselves in a "holy" light. Tom's, you have become an absolute farce. Leave it to California business to be corrupt and try to favor certain companies over others. We'd be morons to believe that it doesn't happen in the geek sector. Everyone has a price.