 |
 |
 |
|
 |
You are not logged in.
Just a small sample from the article:
Flash SSDs Don’t Improve Your Notebook Battery Runtime – they Reduce It
Flash-based solid state drives (SSDs) are considered to be the future of performance hard drives, and everyone seems to be jumping on the bandwagon. We are no exception, as we have been publishing many articles on flash-based SSDs during the last few months, emphasizing the performance gains and the potential power savings brought by flash memory. And there is nothing wrong with this, since SLC flash SSDs easily outperform conventional hard drives today (SLC = single level cell). However, we have discovered that the power savings aren’t there: in fact, battery runtimes actually decrease if you use a flash SSD.
Could Tom’s Hardware be Wrong?
No, our results are definitely correct. We’ve looked at almost a dozen different flash SSDs from seven vendors over the last few months, and measured acceptable or sometimes even disappointing power requirements with most flash SSDs. In an effort to determine the actual impact on notebook systems, we took four SSDs that we had available in our test lab, and ran a series of Mobilemark benchmark runs on a Dell Latitude D630 notebook. We found runtime differences of up to one hour (!) when using a flash SSD compared to a high-performance 7,200 RPM 2.5” notebook hard drive.
Offline
I posted this in general... Sorry moderators, i didnt know better ... sigh!
Offline
very interesting!
and a bit sad.
Offline
If that test was made under constantly doing read/writes then yes... it may be true. However for a "normal use" a drive will be idling by most of time. Then SSD should consume less energy.
Offline
SSD is still a developing technology. It may be a few years yet before its full benefits are realized.
Offline
One testing flaw pointed out in some of the comments to that article is that Tom's simply ran the benchmarks continously until the battery ran down. What's not taken into account is how much *work* is performed by the system with the SDD vs. the HDD for the respective battery runtimes. Since SDD is faster, the CPU doesn't have to wait to process the next task as it would for the slower HDD. If there are many of these idle CPU periods, then obviously less overall system power is consumed in a given time interval. If the SDD version performed more work on a single battery charge (because it's faster) than did the HDD version, then the SSD version would be more productive and power efficient even though it may not have lasted as long based on clock time.
Another inconsistency that was pointed out has to do with the chart on page 14 of the article. It shows the Sandisk SSD consuming 1.0 watts of power under LOAD. Compare this to the Hitachi HDD which consumes 1.1 watts of power when IDLE. Under load, the SDD uses LESS power than the HDD does when idle. Does it make sense then that the SSD would be LESS energy efficient? The article explains it thusly: "However, it has become clear that idle and maximum power do not provide the full picture when we talk about flash SSDs." So they are acknowledging they don't have the full picture when confronted with this anomaly, yet they don't bother to find out what those unaccounted factors may be in their testing methodology or how these possible oversights/flaws may negate the validity of their findings.
It would have been more "real world" to see a comparison of two side-by-side users, one with SSD and the other with HDD, launching the same applications at the same time, accessing the same web pages at the same time, editing the same documents at the same time, listening to the same music files at the same, watching the same videos at the same time, etc. until the respective batteries ran down. Swap the batteries and repeat this test to offset any variations in individual batteries. This way, both systems will be performing the same amount of work in the amount of time, and the first system to conk out will be the one that's less energy efficient.
Last edited by artrafael (2008-07-02 5:01:40 am)
Offline
artrafael wrote:
Another inconsistency that was pointed out has to do with the chart on page 14 of the article. It shows the Sandisk SSD consuming 1.0 watts of power under LOAD. Compare this to the Hitachi HDD which consumes 1.1 watts of power when IDLE. Under load, the SDD uses LESS power than the HDD does when idle. Does it make sense then that the SSD would be LESS energy efficient? The article explains it thusly: "However, it has become clear that idle and maximum power do not provide the full picture when we talk about flash SSDs." So they are acknowledging they don't have the full picture when confronted with this anomaly, yet they don't bother to find out what those unaccounted factors may be or how they may negate the validity of their findings.
Typical journalism focused on making grandiose claims just to generate extra web traffic. Most people will read headers only and make wrong assumptions. But "sensation" is already running wild and how we may know if there wasn't some other interests doing so?
Offline
artrafael wrote:
One testing flaw pointed out in some of the comments to that article is that Tom's simply ran the benchmarks continously until the battery ran down. What's not taken into account is how much *work* is performed by the system with the SDD vs. the HDD for the respective battery runtimes. Since SDD is faster, the CPU doesn't have to wait to process the next task as it would for the slower HDD. If there are many of these idle CPU periods, then obviously less overall system power is consumed in a given time interval. If the SDD version performed more work on a single battery charge (because it's faster) than did the HDD version, then the SSD version would be more productive and power efficient even though it may not have lasted as long based on clock time.
Another inconsistency that was pointed out has to do with the chart on page 14 of the article. It shows the Sandisk SSD consuming 1.0 watts of power under LOAD. Compare this to the Hitachi HDD which consumes 1.1 watts of power when IDLE. Under load, the SDD uses LESS power than the HDD does when idle. Does it make sense then that the SSD would be LESS energy efficient? The article explains it thusly: "However, it has become clear that idle and maximum power do not provide the full picture when we talk about flash SSDs." So they are acknowledging they don't have the full picture when confronted with this anomaly, yet they don't bother to find out what those unaccounted factors may be in their testing methodology or how these possible oversights/flaws may negate the validity of their findings.
It would have been more "real world" to see a comparison of two side-by-side users, one with SSD and the other with HDD, launching the same applications at the same time, accessing the same web pages at the same time, editing the same documents at the same time, listening to the same music files at the same, watching the same videos at the same time, etc. until the respective batteries ran down. Swap the batteries and repeat this test to offset any variations in individual batteries. This way, both systems will be performing the same amount of work in the amount of time, and the first system to conk out will be the one that's less energy efficient.
I agree. Their method is flawed (although their conclusion might not be, they just need to re-run the test, correctly this time). Hard drives are a major bottleneck in a system. If the CPU and other hardware is waiting for the hard drive, they will be idle, and will consume less power, resulting in longer battery life. SSDs are much less of a bottleneck. That means less idle time, more power consumption by other hardware, and a lower battery life.
I'm surprised THG couldn't see this for themselves. It's a horribly flawed method. But I don't think they were out to perform a thorough unbiased test. They went into this wanting to prove SSD was inferior contrary to what most believed (for sensationalism), and that's exactly what they did. THG fails for the very same reason they have for all these years. Lack of neutrality. They got tunnel vision here.
Offline
Well, I won't be visiting Tom's Hardware any time soon. They've lost me (as if it matters).
Offline
See this for a convincing rebuttal: http://blog.laptopmag.com/web-surfing-t … #more-2307
Offline
I stopped visiting Toms early this year, their testing methods went Hollywood, as others above pointed out quite clearly, including damnable quotations.
Offline
I love seeing EEEuser mentioned in my Tech News/Blogs. Makes me proud to be an EEE owner.
Offline
Well, speaking as an electronics engineer for the past 35 years I smell a hidden agenda or incompetence here. Look, one watt of power dissipation in an SD card package would raise that package to burning hot in a matter of seconds and likely would soften and melt the SD card. Last night I did an install of Ubuntu to a 4 gig SDHC card which subjected that card to 45 minutes of the heaviest activity it is likely to ever see. At the end of the process the SDHC card was at the same temperature as the inside of the Eee so the average power consumption of that SDHC card during the install of Ubuntu was FAR less than one watt. If I had to estimate I'd say no more than 100 mW could have been dissipated in that package for that long and have it come out not even perceptibly warmer than ambient.
So there's something fishy about this argument.
Offline
To be fair, they were not benchtesting SD cards, but HD replacment SSDs - These use a lot more power than normal flash cards as they are highly multiplexed/parallelized in order to achieve HD-level speeds.
SSDs will use more power when being hammered compared to a disk drive, esp. if the HD is writing a linear stream rather than random-access.
To write a single bit, Flash requires a great deal more power than a conventional HD, so the findings are feasible. Flash should still be more efficient overall since it has near-0 power draw when idle, and while HD's idle power draw is very low, it's still significantly more than 0 
Offline
Tom's Hardware just published a new article that acknowledged the procedural flaws in their first set of tests. In addition to offering an apology, they ran new tests that showed one SSD device outperforming the HDDs. However, in general, since results are dependent on the type of usage employed, there's no clear winner in an SSD vs. HDD face-off. Details here: http://www.tomshardware.com/reviews/ssd … ,1968.html
Offline
The OCZ flash drive seems interesting,but when I went to their forum,immediately I noticed an unusual amount of people complaining about failing drives; one of them the day after the OCZ's Sata2 SSD came out!
Basically the review just praises the (still working) OCZ drive as number one, the MTRON for it's speed, and the Sandisk SSD5000 for it's low power consumption.
The drive that stood out for me, but which wasn't really documented is the Supertalent Master drive MX!
Sure it's not as fast as the other SSD's and uses more power then the Sandisk SSD, but the price is several tens to hundreds of dollars cheaper then an MTRON; and it also performs faster then a HDD.
I think price / performance the Supertalent isn't a 'Superbad' buy! Especially if their drive failure rate is lower then OCZ SSD's.
Last edited by ProDigit (2008-07-19 8:00:48 am)
Offline
This is all very interesting, but for some reason, the manufacturers - both Asus and the competition, don't seem to be embracing SSDs. Most of the new models in this space sport traditional HDDs, without even an SSD option. There must be a reason for this - anyone know or care to speculate what it is?
Offline
grizdog wrote:
but for some reason, the manufacturers - both Asus and the competition, don't seem to be embracing SSDs. Most of the new models in this space sport traditional HDDs, without even an SSD option.
Err... Asus Eee PC 900/901/1000, Acer Aspire One, upcoming Dell E, Thinkpad X300, MacBook Air (option), Voodoo ENVY 133 (option), as well as a number of lesser known brands -- all have SSD storage.
There must be a reason for this - anyone know or care to speculate what it is?
HDD currently costs less per unit of storage than SSD, so notebooks that employ them can hit a lower price point along with higher storage capacity. If you're advertising a loss-leader notebook in the Sunday newspaper supplements, you're not going to be pushing SSD-equipped models. That's also the reason only high-end models (such as the Thinkpad, MacBook Air and Voodoo ENVY) offer SSD for storage and why inexpensive netbooks that do employ SSD (such as the EeePC, Aspire One and Dell E) only offer them in such small capacities.
I don't think Lenovo would risk the reputation of its button-down Thinkpad line by offering SSD storage if the technology is suspect. I think they, and the other top tier manufacturers, would have done a lot more testing than Tom's Hardware before choosing to incorporate SSD in their top-end products.
Last edited by artrafael (2008-07-22 9:09:40 pm)
Offline
Yes, good point - I didn't really make my point very well. With the early Asus models, it was all SSDs. Now you have a mix in the Asus models, and no SSD option on some like the Gigabyte, etc. So I should have said that I saw some movement away from SSDs. And it may be that part of my question was answered by the CEO of Sandisk, viz.:
http://news.cnet.com/8301-13924_3-99963 … 7-1_3-0-20
Offline
