Are SSDs Really More Power Efficient?

Bakasama writes "Tom's Hardware compared the power performance of several available SSD cards with a Rotating HDD that was chosen specifically for its poor power efficiency. The results seem to fly in the face of current wisdom. '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.'"

Comments

More power but only while being hammered?

[gmack]

So if your not a laptop user and aren't currently benchmarking your drive how long will it last?

What is the power usage for real world office/ web browsing type use?

Re:More power but only while being hammered?

[Anonymous Coward]

Do you only talk in questions?
Ever consider a career in politics?

Re:More power but only while being hammered?

[neokushan]

Shouldn't he be a philosopher?

Re:More power but only while being hammered?

[PinkPanther]

Shouldn't philosophy be him?

Re:More power but only while being hammered?

[morgan_greywolf]

Shouldn't you guys stop answering questions with a question?

Re:More power but only while being hammered?

[autocracy]

You should not be.

Re:More power but only while being hammered?

[thePowerOfGrayskull]

Why do that?

Re:More power but only while being hammered?

[Anonymous Coward]

Were they playing Whose Line Is It Anyway?

Re:More power but only while being hammered?

[ralphgra]

More tests have to be done. I have one of the earliest hard-drive MP3 players, the PhotoTainer. It has a 20GB drive, but also a CompactFlash slot. I did a timing test to see how long the battery would last running songs off both storage media. The battery lasted about 1 hour longer when using the CompactFlash memory card than using the hard drive.

Re:More power but only while being hammered?

[Anonymous Coward]

Why was this thread modded off topic?

Re:More power but only while being hammered?

[HardCase]

So if your not a laptop user and aren't currently benchmarking your drive how long will it last?

What is the power usage for real world office/ web browsing type use?

That's the question that the test didn't answer. The true measure of SSD versus mechanical hard drive would be to normalize the power usage by analyzing the difference in total number of benchmark runs per drive. I'll bet that the SSDs completed more benchmarks than the hard drive. And if that's the case, then there should be some additional hardware overhead that would account for additional battery power drain.

Maybe a better analysis would be to determine the number of iterations per charge. But the best analysis would be to measure the actual power consumption by the SSDs and hard drive. With the sort of low budget experiment that Tom's did, there's really no way to effectively do that. But a little bit of statistical analysis, with repeated tests and normalization would give an accurate picture of the relative power use between devices.

It's pretty clear that the experiment that Tom's did is flawed, even if the article promises that it can't possibly be wrong (hubris, anyone?)

Re:More power but only while being hammered?

[Yvan256]

Stop talking like that, you should.

Re:More power but only while being hammered?

[cjsm]

It has a 20GB drive, but also a CompactFlash slot. I did a timing test to see how long the battery would last running songs off both storage media. The battery lasted about 1 hour longer when using the CompactFlash memory card than using the hard drive.

Well, the problem with that is you aren't writing any data; which, due to how flash writes, takes quite a bit more power then reading.

Re:More power but only while being hammered?

[default luser]

Yes, this is correct, flash has terrible power consumption while writing. TWO REASONS:

1. When you re-write any data word inside a block, you have to re-write the entire block. This can get power-intensive if you are doing random writes, and even sequential writes can eat a lot of energy (yeah, show me a flash controller that can detect, cache and optimize every sequential write perfectly).

2. Write voltages on flash are much higher than the read voltages, because the write voltage forces electrons through a thin insulator (quantum tunneling).

With an almost pure read-only role, like an mp3 player, flash will use much less power than a mechanical drive.

Re:More power but only while being hammered?

[TClevenger]

From TFA: "While the total battery runtime certainly depends on the workload -- we used Mobilemark 07 -- the minimum and maximum power consumption measurements prove that Crucial's statements of low power consumption are in fact wrong: 1.6 W idle power is more than any 2.5" notebook hard drive requires."

Obviously that cannot be!

[Anonymous Coward]

Impossible! Those results are obviously wrong!! Now go back and do the experiment. Keep doing different experiments until we get the desired results!

How dear you try and endanger my stock portfolio?!

Re:Obviously that cannot be!

[cyfer2000]

Are you serious, dear?

Re:Obviously that cannot be!

[Anonymous Coward]

If I was Control your stocks would've already tanked

If you were Control your stocks would've already tanked

Neither of our stocks have tanked, so obviously I'm not from Control

Re:Obviously that cannot be!

[hclewk]

Unfortunately, the article comes to miserably faulty conclusions, and nobody seems to have noticed it. One thing everybody needs to note is this: the test that Tom's ran was designed to run a program to completion over and over again. All that Tom's came up with, thanks to that test, is that the computer with the SSD ran out of power faster. What they SHOULD have noticed, and what any sane and logical tester would have done, is counted the number of times the program RAN. A system with an SSD will run the program a significantly higher number of times because of its increased speed and lack of moving parts. With the higher number of runs, the CPU STAYS MORE ACTIVE, sucking more power out of the system and thus causing it to power down earlier. The only reason the HDD system died later is because the CPU idles waiting for it to retrieve data. The SSD may have chewed up your battery faster, but it did, comparatively, run through that program a BUNCH more times. I am willing to bet the work per watt was much better out of the SSD system than the HDD - and once again, Tom's Hardware's scientific testing system proves itself anything but.

Re:Obviously that cannot be!

[WK2]

When I read the summary, I thought, "Surely Tom's Hardware is doing something wrong. It is Tom's Hardware, after all." But I wasn't sure what. Thanks for pointing it out, so I don't have to RTFA.

Re:Obviously that cannot be!

[SoupIsGoodFood_42]

Interesting, I didn't notice that. But one thing I did notice is that it didn't seem to take into account real world use, things such as at what point does the frequency of the extra power needed to spool up a HDD outweigh any power savings -- laptops are turned on and off a lot more than a desktop, and if the responsiveness is there, then people are going to put their computer to sleep more often to save power. None of that was taken into consideration. So technically, he may be right in some circumstances, but that might not be the way most people use them in the real world.

Re:Obviously that cannot be!

[Anonymous Coward]

It'd be nice if you credited me when you took my comment straight from Engadget ;)

-Matt (original writer)

Re:Obviously that cannot be!

[immcintosh]

I'm not so sure your point is quite that significant. There is a graph, towards the end, that seems to be multiplying throughput by total battery life, giving a sort of "megabyte-minutes" rating for the different drives. In terms of simple hard drive throughput, this seems to indicate that the work per watt of the traditional drive was still superior (albeit by a small margin over the most expensive SDD), despite your complaint. But obviously it's not quite that simple--no real life usage would cause non-stop disk access like that.

The claim that the CPU stays more active with the faster drive, while technically true, is a little misleading and not nearly as clear cut as you're making it out to be. The only time the CPU would really be more active with a faster drive is under circumstances where it would be waiting for some kind of blocking I/O from the drive, which in my experience (at least under mundane use) isn't all that much. Most of the time you're much more likely to be dealing with system RAM than hard drive storage during program use (unless you run out of memory and start swapping things out, but then you've got other problems).

In short, while you point out perhaps an interesting oversight, I don't think it is quite as serious as you make it out to be.

Re:Obviously that cannot be!

[Carewolf]

Ehm no. Current generations of SSD are slower than ordinary hard-disk for regular use. SSDs are only faster than HDD when testing random access. So if you are testing; let us say MP3 playing, then the regular hard-disk would be both faster and more energy efficient, not to say cheaper.

Re:Obviously that cannot be!

[Anonymous Coward]

I don't know if you made it as far as page 14, performance x battery runtime index:

http://www.tomshardware.com/reviews/ssd-hdd-battery,1955-13.html

Does that address your "miserable failure" conclusion, or am I missing something?

Re:Obviously that cannot be!

[necrodeep]

The only 'Hoax' here is that this was a "scientific" study which Toms Hardware could not be wrong (they even state that they can't be wrong in the article). Talk about a load of bull. I couldn't have put it better myself.

You need to measure - WATTS/PERFORMANCE not TIME/DEPLETION.

Unfortunately they have no way of showing if it was the SSD drive or the CPU that was eating all that extra power with the method they used. I would call that decidedly un-scientific.

Good call hclewk!

Re:Obviously that cannot be!

[Kuvter]

"Dear", I don't think that word means what you think it means.

Re:Obviously that cannot be!

[limaxray]

As I read the Tom's Hardware article the other day I had this same thought. My problem with the test was measuring power consumption by battery life which just screams poor proxy measurement. I think a much, much better method would have been to use an ammeter inline with the drives' power supply. This way, they could have measured and calculated the actual power consumption of the different drives over the course of the test.

I tend to believe you are correct though; a processor requesting data on a SSD is going to have to idle for a shorter period of time since the latencies are considerably lower compared to a mechanical drive. This point may be partially concealed in the performance tests due to the lower throughput but I'd think a lower seek time would have a much greater effect on the amount of time a processor idles compared to a higher throughput. Basically the processor is going to idle during the extra several milliseconds that a mechanical drive is searching for a block of data, but it's not going to idle during the extra fraction of a nano second between each byte received on a SSD.

As the technology improves though, I see this becoming a non-issue.

Re:Obviously that cannot be!

[Anonymous Coward]

I would have thought hte easiest way to test thias is to have a laptop with SSD and spinning platter options. Record testing script of several hours worth of normal off power activity.
Fully charge battery, conected SSD drive, run script over and over until laptop finally suspends.
Swap to spinning platter and retest.

One of the things I found could add 15-20 mintutes more runtime to my tablet was to not have a swap partition, but performance was soo poor I decided If I really needed more than 2.5 hrs runtime, I'd better bring the power pack or buy another battery.

The best Laptop I have is a 5yr old Toshiba Portege 1.2GHz CPU with a long life battery and 2GB RAM. Weighs almost nothing, and the battery can last 6 hrs or more depending on wireless connectivity. However it takes nearly 10-15 minutes to boot from cold to usable, but luckily I can come back from suspend in about 3 minutes.
This works fine for me as my laptop is really just a terminal back to the servers I run. Only ocasionally will I actually use the local applications for anything.

Re:Obviously that cannot be!

[Anonymous Coward]

Let's save some money and skip this experiment stuff and say we did

Re:Obviously that cannot be!

[Anonymous Coward]

Don't bullshit a bullshitter sir.

Re:Obviously that cannot be!

[Anonymous Coward]

actually, i believe many noticed the inconsistency in the test, hclewk, but all faulty procedures aside, the article still brings up a point for consideration:

* why are the tested SSDs using so much power at idle?

i went on and dug up some more information on some of the tested models, and it turns out that aside from any on-board ASICs those SSDs may host, the fastest of them also feature DRAM cache, some as big as 16MB. now, those caches seem to be running off fat capacitors (i know at least one caching SSD that does it), and capacitors are not 100% efficient either, adding to that DRAM's raw power draw.

now, that said, some of the lower-end SSDs (esp. PATA ones) seem to have much (and i mean *much*) better power draw at idle. for instance the newest SSD by Intel does ~1.5 mW at idle (~300mW active, IIRC) - this is a far call from the 1W range recorded by Tom.

so, the point about not-quite-efficient ASICs/extra circuitry hosted by the SSDs is not quite without merit either.

Re:Obviously that cannot be!

[RealGrouchy]

The more important question is: Is there any truth to the commonly held notion that SSD have fewer moving parts than HDD?

This will require vigorous scientific testing...

- RG>

Re:Obviously that cannot be!

[ralphbecket]

You're not a climate scientist, are you? :-)

Re:Obviously that cannot be!

[Anonymous Coward]

That is very true, what they actually ought to do is use an ammeter and measure the current while the different drives are idling and again while they are reading. That way it would actually measure the power consumption of the drive. Though after measuring the values they would need to factor in that SSDs are also faster so they will spend less time active, thereby reducing power consumption.

Re:Obviously that cannot be!

[Anonymous Coward]

It wasn't even funny the first time

Re:Obviously that cannot be!

[Anonymous Coward]

link please

Re:Obviously that cannot be!

[Anonymous Coward]

He'd have to remove that bothersome 'science' thingy and wear a robe with a hood.

Yes, SSDs are more efficient

[damburger]

If you want to kill everyone on a particular planet, vapourising its surface is way more efficient than trying to blow apart the planet.

Re:Yes, SSDs are more efficient

[j-pimp]

If you want to kill everyone on a particular planet, vapourising its surface is way more efficient than trying to blow apart the planet.

But its not as pretty.

Re:Yes, SSDs are more efficient

[Trespass]

If you want to kill everyone on a particular planet, vapourising its surface is way more efficient than trying to blow apart the planet.

But its not as pretty.

It depends where you put the cameras.

Re:Yes, SSDs are more efficient

[BoberFett]

Why would you need cameras? Just re-render the scene for the director's cut.

Re:Yes, SSDs are more efficient

[servognome]

Planet shoots first?

Re:Yes, SSDs are more efficient

[OrangeTide]

If you want to kill everyone on a particular planet, vapourising its surface is way more efficient than trying to blow apart the planet.

Now you're just being lazy.

Still too new

[ArcherB]

I believe that much of the problem is that SSD's are still a new technology compared to rotating disks. Right now, engineers are more concerned with increasing capacity and just making the damn things work. These are much more important than efficiency. As time goes on and the technology gets more mature, efficiency will get more attention from engineers.

Re:Still too new

[drinkypoo]

The problem is apparently that they're using lots of flash without power-saving (presumably cheaper than the other kind) and they're also not turning the currently-unused chips off. I can't speak for anyone else but I spend the vast majority of my time accessing a minority of the data on my hard disk. The technology already exists to reorder disk blocks based on usage and it could easily be adapted to reduce the number of flash chips which need to be activated in an SSD. Perhaps the time to first access is consider to be too large to implement this sort of thing without on-chip power-saving features, although I doubt it would be more than a few small fractions of a second.

Re:Still too new

[initialE]

If this was true then SSDs are already condemned to failure. A more holistic approach is required in order to get people to want the thing in the first place. Which would then lead to more money for research. And then to faster access times, lower power levels and manufacturing costs.

Re:Still too new

[Ilgaz]

Call me some guy mystified by brands but after using Seagate cheap stuff (SATA) and SCSI for years and never seen these things actually fail (besides stupid filesystems), I wait for Seagate, Fujitsu, Hitachi like known brands to ship their SSD rather than being abused by some memory vendor who has no clue about the hard disks to buy some overpriced flash memory fantasy.
I also don't know the actual reliability of SSD too. What about journaling? Can it handle? A journal is still needed on SSD drive, what if kernel fails or OS filesystem layer goes nuts? A journal will be in same area of disk and will be written over and over millions of times.
I could never buy the "speed" claims of SSD not just because I use very fast SCSI stuff but I actually see the horrible performance of them in my smart phone, HD Camera. It is like performance suicide if someone dares to put a very complex applications to "memory card" instead of phones built in memory.
They are trying to ship it before it is a technology fit to general use. Much like some video sites existed while everyone had to struggle with 56K modem.

Re:Still too new

[pipatron]

I could never buy the "speed" claims of SSD not just because I use very fast SCSI stuff but I actually see the horrible performance of them in my smart phone, HD Camera.

Connect your very fast SCSI drive to your phone and see if it's still as fast.

Re:Still too new

[MindStalker]

You realize that the in phone memory is probably flash based too?

Re:Still too new

[Gewalt]

So what they are saying then is that this brand new technology is not as refined yet as the one that's been around for almost 30 years? Shocking!

Re:Still too new

[TubeSteak]

The technology already exists to reorder disk blocks based on usage and it could easily be adapted to reduce the number of flash chips which need to be activated in an SSD.

Uhhh... wouldn't that defeat the purpose of using wear leveling algorithms?

A more relevant technique would be to avoid turning on flash chips that do not have data you're accessing... but that is not as easy as it sounds when your data is (purposely) fragmented all over the place.

Re:Still too new

[Darkness404]

A more holistic approach is required in order to get people to want the thing in the first place.

There already is enough want for a SSD. The downside is most people want more then the pathetic few GBs of storage they have to offer. So when they make an affordable, say 500 GB SSD, more people will buy them leading to more SSD makers leading to faster ones and lower powered ones along with more reliable and more space.

Re:Still too new

[Anonymous Coward]

I could never buy the "speed" claims of SSD not just because I use very fast SCSI stuff but I actually see the horrible performance of them in my smart phone, HD Camera. It is like performance suicide if someone dares to put a very complex applications to "memory card" instead of phones built in memory.

And what do you think that phone's built-in memory is, if not flash? Maybe a very little bit of it is actual battery-backed RAM... But I'd be willing to bet that most of it is flash of some kind.

There's quite a bit of difference between, say, an SD card and a Sony Memory Stick. There's also plenty of difference between an SD card and these flash-based drives.

I mean... What you're saying is like claiming regular old spinning-disk drives will never have good performance because your old 5.25" floppy was slow.

Re:Still too new

[emj]

Uhhh... wouldn't that defeat the purpose of using wear leveling algorithms?

No not if you activate only those chips that you need to execute that read-write-write operation. But then again I'm sure if it really was as simple as switching some thing on and off then we would have good results by now..

btw, look at the IO ops per second graph, it's interesting to note that not all SSDs are better than the disk. Though the best SSD beats disks with three orders of magnitude in webserver load..

Re:Still too new

[kesuki]

it's different when you're just reading data, VS writing data, and these benchmarks compared power usage while writing gobs and gobs of data.

flash memory chips can use 5X as much power to 'write' data as they do to 'read' data, oh and hey if you're reading the same old data over and over, why not just have it in ram, and just not spin up a drive or read flash memory at all?

oh and wear leveling can be designed around keeping as few chips powered up as possible, you just need to reserve a bit of flash memory to cache a wear leveling pattern, so it knows when to power up and down which chips... trying it's best to keep data sets on the fewest chips, again this is a bot more complex than basic wear leveling, so it's not tro be expected on early SSDs...

Re:Still too new

[GleeBot]

Did you know that all memory cards aren't built the same? The cheapest cards often use very slow grades of memory (for obvious reasons).

I use SanDisk Ultra and Extreme CompactFlash cards in my DSLR, and right now it's limited by the speed of the camera's processor, not the memory cards. I can read/write data off them at something like 20 MB/s using my computer's card reader, and these aren't even the fastest cards on the market (they're older generation versions).

Re:Still too new

[somersault]

A more holistic approach is required in order to get people to want the thing in the first place

Yeah, people are never drawn in by flashy gimmicks, are they?

Re:Still too new

[rolfwind]

I believe that much of the problem is that SSD's are still a new technology compared to rotating disks. Right now, engineers are more concerned with increasing capacity and just making the damn things work.

Perhaps the problem is that rotating disks are too mature - I have not seen a capacity increase in a year and a half beyond 1TB (though I may have missed an announcement).

Re:Still too new

[Mr. Slippery]

it's different when you're just reading data, VS writing data, and these benchmarks compared power usage while writing gobs and gobs of data.

Exactly. To determine if people will see battery power saving benefits from using flash vs. hard drives, you need tests that reflect the usage pattern of notebook computers on battery - occasional reads, rare writes, and a lot of idle time. (If your notebook is actually a desktop replacement, then I presume it's plugged in.)

It's the idle time that make a big difference: flash doesn't consume power when idle, whereas with a hard drive you need to either keep that sucker spinning (costing power) or suffer a big performance hit by starting it up for each operation. It's not clear that their benchmarks reflect this usage pattern.

I put a small (4 GB) SSD in my old Vaio SRX77 and get about an extra half hour to 45 minutes of battery life out of it now. Of course, I lot several gigs of storage on the deal, but that's fine - I took a junk machine and for about $70 turned it into something roughly in the class of a Asus Eee on the cheap.

Re:Still too new

[Darinbob]

Of course, you have to take into account that these SSD drives do have low power modes. Don't know if the systems take advantage of that or use it aggressively. Ie, they should go to low power mode immediately after using the device. The SSD drives need to do this without waiting for the operating system to tell them to go idle.

Of course, none of this would seem to matter in the Tom's Hardware tests where they used the drives continuously rather than simulate real world usage.

Re:Still too new

[Anonymous Coward]

I am not positive, but I thought they try to spread the data out across the flash chips due to the fact that they wear out over time. They need to spread out the writing over different chips so the drive as a whole ages evenly.

Can anyone confirm my thinking or am I mistaken?

Re:Still too new

[kbielefe]

I maintain a flash device driver for an embedded system as part of my job, and I have to say you have an interesting mix of misconceptions and valid points.

First of all, I'm wondering what you think your phone's built in memory is. If it's not flash, I feel sorry for you if your battery dies. Second, the nature of flash memory makes it highly desirable to implement some form of journaling. To change one byte of a file, you have to erase a sector, then write it back with the one byte changed. It is much easier to just write it to a new location then invalidate the first copy. I don't make memory cards or SSDs, but it is my understanding that most memory cards do this behind the scenes, as the world's most popular OS is too dumb to handle flash memory correctly. The OS thinks it is writing to the same physical location, but behind the scenes it's changing on every write.

You made a good point about the millions of writes, though. I'd love an SSD for my root partition that hardly ever gets written to, but I think it would wear out too quickly for /home unless it was severely oversized, and I'd probably want to get a lot of RAM to use for /tmp instead. A lot of applications with autosave and disk caching would need to be reworked I think in order for ubiquitous flash drives to be feasible.

Re:Still too new

[BoberFett]

I believe WD has 640GB on two platters, meaning they should be able to do a ~1.28TB drive anytime.

I'm fairly sure about that anyway. I think most current 500GB drives are 2 platter, with the 1TB drives being 4.

Re:Still too new

[Mista2]

Or have two drive, small super efficient 4-8 GB micro-SD flash for OS and swap, less efficient but larger drive for user data with aggressive powersaving turned on (suspend drive when not in use etc)

Re:Still too new

[LoRdTAW]

You also have to take into consideration that a phone electronics are optimized for low power consumption and very small form factor. Its apples to oranges. A micro SD card has a painfully slow data transfer rate of about ~900kB/sec read and about 600-700kB/sec write due to its low speed serial data interface. You cant expect wide and/or a fast data bus in such a small and low power storage device. USB thumb drives can afford the nearly 10MB/sec thanks to 100-200mA@5V power and 480Mb/sec data interface. Compact flash uses a 16 bit wide ATA bus that enables 10+ MB sec in CF cards.

Now imagine an array of flash chips optimized for high speed transfer and a few chips all making up a wider data bus that ties into a SATA interface. Thats along the lines of a SSD. You can get high speed, it just needs to be engineered to do so.

But one thing that concerns me as well is not just speed but read/write life span. An SSD with a busy swap file on it can be beaten to death pretty fast regardless of wear algorithms.

Re:Still too new

[Eil]

Wow, um, you definitely could use some edumacation on how SSDs and hard drives in general work. Here, let me help.

after using Seagate cheap stuff (SATA) and SCSI for years and never seen these things actually fail (besides stupid filesystems)

Then you don't handle very many hard drives. I work in a data center and around 5% of the disks we buy fail within a few months. We don't have one brand in particular that we use because they seem to all fail at about the same rate. (Seagate does have the best warranty and RMA program, though.)

I wait for Seagate, Fujitsu, Hitachi like known brands to ship their SSD rather than being abused by some memory vendor who has no clue about the hard disks to buy some overpriced flash memory fantasy.

Who do you think Seagate, Western Digital, and Fujitsu will buy their memory from? Those companies manufacture hard disks, not memory chips. They have huge investments in the production of mechanical drives. It's possible that some of them might set up memory fabs at some point, but that could still be a long way off because mechanical drives are not going to be completely obsolete for quite some time. In the short term (and possibly long term), they're going to be outsourcing flash chips for their SSDs from lots of companies you've never heard of before.

I also don't know the actual reliability of SSD too. What about journaling? Can it handle? A journal is still needed on SSD drive, what if kernel fails or OS filesystem layer goes nuts? A journal will be in same area of disk and will be written over and over millions of times.

It's too early to tell what the long-term reliability of current SSDs will be, but it's dead-certain that they will improve regardless. (Keep in mind that all early hard disks came with defects on them from the _factory_ and users were expected to format around them, so SSDs already have a good head start in terms of reliability.) SSDs were just introduced and haven't had much real-world testing yet. However, all of the manufacturers have been marketing them as replacements for mechanical hard disks, so clearly they expect the lifespan of an SSD to come close to that of a mechanical disk.

Journaling isn't a concern, because all of these drives implement wear leveling to lengthen the life of the drive.

I could never buy the "speed" claims of SSD

You don't have to buy anyone's claims, look at the numbers yourself. The read speed of SSDs beats the pants off mechanical disks. And I believe they've caught up on write speeds already.

not just because I use very fast SCSI stuff but I actually see the horrible performance of them in my smart phone, HD Camera. It is like performance suicide if someone dares to put a very complex applications to "memory card" instead of phones built in memory.

Err, yeah, I'm pretty sure your bottleneck is going to be the smart phone with it's 200MHz CPU and 0.9MB/sec max transfer rate.

They are trying to ship it before it is a technology fit to general use. Much like some video sites existed while everyone had to struggle with 56K modem.

The problem is that they're hidously expensive. SSDs are just fine for general use. There are people using them that don't have a problem with them. Yes, there are some drawbacks because they're an early technology, but it will get better with time. Remember the first LCD monitors? People said those would never catch on. They were analog only, had a very narrow viewing angle, displayed washed-out colors and horrible refresh rates. They were also hideously expensive but eventually these problems were fixed and price came down until they were comparable with CRTs. In just a couple years, you won't be able to buy a CRT monitor at an affordable price because LCDs have made them obsolete in just about every field.

The same thing will happen with mechanical drives and SSDs.

Re:Still too new

[drinkypoo]

Uhhh... wouldn't that defeat the purpose of using wear leveling algorithms?

No.

Longer answer: if a seldom-written file is fragmented enough to be written across two chips, then you have to turn them both on to read it. So it's desirable to prevent this from happening as much as possible.

Ideally your filesystem minimizes this tendency as well. There are some filesystems meant for use on flash these days.

Re:Still too new

[Larsrc]

Why would you want a SSD vendor to know anything about hard disks? Do you care whether your car salesman knows anything about boats?

Comparing SCSI to handheld devices is like comparing apples and handgrenades. Guy at my former workplace did extensive testing of high-performance (RAID, 15K, SCSI) HD vs. SSDs on Lucene searches and found speedups of a factor 2-5. That's where the SSDs really shine right now, and it's an area where there's a lot of focus on energy.

I *am* rather shocked at the high idle power usage. That needs some fixin'.

Re:Still too new

[TClevenger]

From TFA: "While the total battery runtime certainly depends on the workload -- we used Mobilemark 07 -- the minimum and maximum power consumption measurements prove that Crucial's statements of low power consumption are in fact wrong: 1.6 W idle power is more than any 2.5" notebook hard drive requires."

running a synthetic benchmark 100% of the time...

[Anonymous Coward]

I wish they would measure power usage under conditions that many notebook computer users actually use them in, which does not include running synthetic benchmarks on their computer 100% of the time it's running. Of course, if you keep the machine writing to the ssd constantly then it's not going to show power savings. But how many mobile users' usage patterns include constant reading from and writing to disk?

Sweet

[Anonymous Coward]

Now I'm even happier that I paid the extra $1300 for the SSD option in my MacBook Air.

That way it will run out of battery sooner, leaving me free to use a real computer.

Re:Sweet

[Anonymous Coward]

I am sure you can get one of those old fashioned hard drives that consume less power, hold much, much more for much less than $1300. It even writes faster and no special software required.

People are rushing to this new technology, I thank them for their beta testing efforts and bringing the costs down. I will probably adopt when:

• the price drops to near hard drive levels
  
• write performance gets better
  
• drivers and kinks worked out
  
• density/capacity is that of hard drives or better
  
• power savings are real (new)
  

Re:Sweet

[Kuvter]

Now I'm even happier that I paid the extra $1300 for the SSD option in my MacBook Air. That way it will run out of battery sooner, leaving me free to use a real computer.

I thought you were going to say, "..more free time to have a real life."

Not so good benchmark

[marc.andrysco]

I believe this comment from the article could explain some of this away.

'There could be a systematic error in the benchmarks shown: if the flash based "disks" are faster then the whole system CPU/MEM/Chipset would draw much more power with flash "disks" compared with conventional disks - just because the benchmarks could run more often in the same time.

Maybe one should compare something like playing video from disk where it is assured that the systems do precisely the same work?'

Re:Not so good benchmark

[CastrTroy]

Not only that, but compare the power consumption when the disk is completely idle. Or at least when the computer isn't requesting any information from the disk. Most of the time, if your computer has enough RAM, it will access the drive quite infrequently, especially in many cases where power drain would be of concern, such as in those UMPCs.

Re:Not so good benchmark

[Anonymous Coward]

Isn't it obvious that you run the bench marks a specific number of times, then include enough idle time so that all configurations do the same work in over the same time. This isn't rocket science.

Re:Not so good benchmark

[MBCook]

That's an excellent point. I was wondering if the metric should simply be different altogether.

What if we used MB/Watt or some such? The "disks" are supposed to be really good at some things (random reads) but don't hold too much advantage over others (long continuous reads). So how many WattHours does it take to load a continuous 50 GB file? How about a random 50 GB of data off a 128 GB disk? How does that compare to the same measurements with a standard magnetic disk? How does power consumption change between reading, writing, and mixed disk loads? Writing flash takes far more power than reading, doesn't it? Yet on a physical disk it's not that different in power requirements.

There are other things too. Operating systems still aren't designed around these things, they are designed for physical rotating disks. Do these things even have native controllers (designed for this purpose) yet, or are they still using modified rotating disk controllers like the first models used? As time goes on, better power saving features will show up, especially as the OS cooperates to tell this disk more information about what's going on. A well managed flash drive may be able to shut off large chunks of it's self and only wake up the bits that actually need reading/writing. That would help quite a bit, I'm sure.

PS: First time I've been to Tom's Hardware in 6 months to a year. Nice to see they found a way to make it uglier. Used to be a nice site. I especially like the "you must login to see the printer friendly version" trick.

Re:Not so good benchmark

[marc.andrysco]

Yeah, another post mentioned that a SSD's power consumption during load is higher than another magnetic disk on idle... here it is:

"The SanDisk SDD drive at LOAD requires 1.0 mW while Hitachi HDD requires 1.1 mW at IDLE"

Of course, benchmarks are always a better indicator. Also, to be clear, as mentioned in the article (which I'm sure nobody else read), the test was performed by repeatedly running a benchmark on the system until it ran out of battery, so the test with the SSD is likely to have run the test more often.

Re:Not so good benchmark

[drinkypoo]

if your computer has enough RAM, it will access the drive quite infrequently, especially in many cases where power drain would be of concern, such as in those UMPCs.

...which tend to have reduced memory.

I really haven't found this to be true any more. My computer hits the disk pretty much all the time (for logging if nothing else) and I've got 2GB RAM. (Currently 1100M allocated in the user space - and all I'm doing is running my fancy desktop and firefox. Linux has truly reached parity with OSX... Firefox is using 128MB with a dozen extensions) And most windows users seem to have a ton of little system tray gadgets they think they need, as well.

I do remember back when I had practically nothing running, and the disk would spin down a lot. Boy, was that annoying.

Then again, my quadro-equipped 17" widescreen laptop gets maybe an hour and a half at best. I want a smaller, more power-efficient laptop to serve as my mobile system (this one then becomes a TV and media server, I think) but haven't got around to that one yet.

Re:Not so good benchmark

[kisielk]

I hate to be pedantic, but 1.0 mW is not higher than 1.1 mW ...

Re:Not so good benchmark

[morgan_greywolf]

Not only that, but compare the power consumption when the disk is completely idle. Or at least when the computer isn't requesting any information from the disk.

Somebody correct me if I'm wrong, but isn't it true that with HDDs, the system can auto-power them down when they're idle, but with SSDs, this can't be done as easily due to the nature of SSDs?

Re:Not so good benchmark

[peragrin]

Of course if the SSD is running under load the entire time it may actually run more often than a regular HD.

from real world benchmarks SSD's only gain is from random reads. writes, sequential reads, etc all prove that SSD's are only as good as a regular spinning disk if not far worse.

It is a trade off. both have advantages and disadvantages. Maybe in 5 more years SSD's will ultimately win but for now it can go either way.

Re:Not so good benchmark

[CastrTroy]

I think that as people request more power efficiency from their laptops, that we will be able to turn off logging, and any other unnecessary writing to the disk. I already boot up my laptop with the "noatime" option so that it doesn't cause an extra write to the disk every time I read a file. I think there's a lot we could do in order to cut down on the number of reads and writes we do. And since spin-up, and spin-down isn't a problem with SSD, we should definitely be trying to cut down on how much we use the drive.

Re:Not so good benchmark

[afidel]

Actually two out of four SSD's reviewed used MORE power when idle than the 7200rpm drive, that's just stupid. The Sandisk used about 60% less, so there's a reason to go with the name brand in this case, they do a heck of a lot more R&D than simply throwing some components together off the shelf.

Re:Not so good benchmark

[FrankieBaby1986]

If running the system until the battery failed was the measurement, well.. thats just stooopid. Who says the battery's lifespan is identical the second time? Where these tests done one right after the other? Because if the first test was run on a 'cool' battery and the second one run on a battery that was just finished recharging five minutes ago, there could be a large margin of error. Only using the internal power usage meter in watts could be even remotely accurate.

Re:Not so good benchmark

[PacketShaper]

Not sure why this has not been mentioned already, but it seems the whole experiment is flawed.

Why would they measure whole-system power draw to test the disk power draw? Why not simply put a power meter in-line with the power connector of the HDD?

Re:Not so good benchmark

[b0bby]

PS: First time I've been to Tom's Hardware in 6 months to a year. Nice to see they found a way to make it uglier. Used to be a nice site. I especially like the "you must login to see the printer friendly version" trick.

I don't go there at all anymore; I've got my ipcop box running URL Filter with the Squidguard blacklist, and whatever tracking junk they run at Tom's just leaves me hanging with a 1x1 gif. It's not even worth my while to add them to the whitelist.

Re:Not so good benchmark

[K. S. Kyosuke]

A well managed flash drive may be able to shut off large chunks of it's self and only wake up the bits that actually need reading/writing. That would help quite a bit, I'm sure.

A very reasonable idea. A friend of mine who happens to be working on Linux power saving features once noticed that the disk controller of X60 in fact consumes even more then an idle disk - at least if I understand it correctly. As the flash does not need to spin up, if you can wake up the controller quickly enough and turn it off again, it might be reasonable even to turn the whole thing off for an extended period of time, including the controller, especially if you have enough RAM (i.e., more disk cache).

Re:Not so good benchmark

[PReDiToR]

My EEE 4G has 4GB SSD, 512MB RAM.
I'd like to think that it isn't suffering from reduced memory, it has the default Xandros distro on it which seems to be a fully fledged Linux system once you get it out of "easy mode".
My other laptop (a tired old IBM T22) has 256MB in and runs openSUSE 11 quite happily.
Is this "reduced memory" a Windows user term? I got Puppy Linux running on a Libretto that has 16MB RAM, and that I could probably say is a little on the small side.

Re:Not so good benchmark

[sexconker]

It's Tom's Hardware, what do you expect?

Re:Not so good benchmark

[TooMuchToDo]

You just need to be able to power the controller and the SSD back up fast enough for the user not to notice.

Re:Not so good benchmark

[Lodragandraoidh]

My MP3/MPEG/RADIO player is a Sansa (SanDisk product), as well as the Micro SD card in it. SanDisk gear is surprisingly high quality.

I can see where they would have an edge in solid state memory hardware.

Re:Not so good benchmark

[PitaBred]

SSD's only use power when they're being read from or written to, so there's no need to separately power them down. As a side-effect of them being non-volatile flash memory, they don't need a "running" charge like a hard drive needs to keep spinning.

Re:Not so good benchmark

[Eil]

The phrase "real world benchmarks" is an oxymoron. There's no such thing as a real-world benchmark. Either you're simulating regular usage of something or you're actually using it. If you're actually using it, you can't use any specific numbers collected because they're going to change the next time you use it because you likely won't use it exactly the same way as before.

Re:Not so good benchmark

[Eil]

Operating systems still aren't designed around these things, they are designed for physical rotating disks.

How exactly are current OSes designed for physical rotating disks?

You wouldn't design an OS around the hard disk any more than you would design one around the keyboard or power supply. As far as any OS is concerned, /dev/sda is just a block device, no matter whether it's a single-spindle hard disk, solid state disk, or a 42-disk hardware RAID 5. If the SSD needs any special treatment (such as wear-leveling), that function should be abstracted away and implemented in the drive itself, not the OS.

Re:Not so good benchmark

[Anonymous Coward]

Surprisingly? Why? SanDisk has been the uncontested leader of flash based storage for years on end. If you ever get near a professional photographer, check out his CF stash. I'll bet you dollars to dimes they will all -- no exception -- be SanDisk Ultra-III or IV. I know mine are, as are all of my colleague's.

Re:Not so good benchmark

[Anonymous Coward]

Not only that, but compare the power consumption when the disk is completely idle.

while your at it, compare the consumption when the computer is off, that will give you most accurate results..

Re:Not so good benchmark

[virtual_mps]

Operating systems still aren't designed around these things, they are designed for physical rotating disks.

How exactly are current OSes designed for physical rotating disks?

You wouldn't design an OS around the hard disk any more than you would design one around the keyboard or power supply.

You should do more research before spouting off. Almost all current filesystems (generally regarded as an OS function) are designed with the assumptions that seeks are expensive, full block writes are no more expensive than single bit writes, and sequential writes are cheaper than random writes. The layout of the data and metadata on disk, as well as OS buffer behavior (e.g., write combining) is designed to optimize for those assumptions, which are untrue for SSDs. (In practice the assumptions are actually somewhat true because current SSDs pretend to be hard disks in order to be compatible with the assumptions of existing OSs--but if the OSs were smarter SSDs could expose an interface more suitable for their unique characteristics.)

Re:Not so good benchmark

[Thundersnatch]

A lot of those assumptions made by the OS about disks are also incorrect for large RAID arrays or SANs as well, not just SSDs. The biggest problem I think would be read-ahead caching based on sequential block ID, which often actually becomes a performance-killing random read on a SAN or even local storage system with volume management or snapshots enabled.

It would be nice if the interface for all storage was just an array of blocks, as the GP suggests, but filesystems really do need to know about the physical characteristics of the storage to be efficient. Which is why NetApp filers, for example, often perform better in NAS mode than iSCSI or FC SAN mode (depending on the filesystem and application). NetApp's WAFL and OS know a lot about the hardware they're running on. Your FC-mounted LUN formatted with EXT3 or NTFS doesn't know jack about the characteristics of all the components inside the NetApp box.

This knowledge of the hardware is why ZFS is killer, by the way, and is not the "massive layering violation" that the Linux kernel guys seem to think it is.

Desktop Computers / Servers

[sleekware]

I would use one of these SSDs on a Desktop Computer or a Server, because in those two situations I am looking for performance, rather than portability that I would want in a laptop.

Been there, seen that, got the t-shirt

[Plammox]

...it's like LCD TVs, people also thought they consumed less power than conventional CRTs. Personally, I can warm my hands if I stick the palms up in front of my 32" LCD which chugs away at 152W when fully "lit" (powersave mode off).

A 32" Philips HD ready CRT was around 100-110 W at the time I looked.
However, this is highly dependent on brand as well.

Re:Been there, seen that, got the t-shirt

[afidel]

Wow, that seems really inefficient, my 42" is only 169W max (typical is more like 110W). I can't find the power usage numbers for what I believe is the only 42" CRT produced, the Mitsubishi MegaView Pro 42, but it weighs in at 200lbs!

Re:Been there, seen that, got the t-shirt

[moosesocks]

Hm. I didn't believe you at first, but it does appear to be true that big LCD panels are hideously inefficient.

Although a 19" LCD will easily draw far less power than a 19" CRT (about 1/3), the same doesn't appear to hold true for big TVs.

Does anybody know why this is? It seems extremely counterintuitive.

Re:Been there, seen that, got the t-shirt

[Bazer]

A 32" LCD may need more power than a 32" CRT but my 19" LCD monitor runs on 30W compared to 120W my old 19" CRT had used (TCO'99).

It's not that simple(tm).

Re:Been there, seen that, got the t-shirt

[clone53421]

I'm guessing that backlight power is probably close to the square of the size (it's pretty closely linked to area). I don't know about CRTs, but they've got the electron beam so maybe it's closer to a linear function of size... that'd explain why small CRTs are less efficient but large CRTs are more, at least.

I'm not an expert, though, so this isn't gospel. It's just what made sense to me in a rare moment of inspiration...

Re:Been there, seen that, got the t-shirt

[GleeBot]

...it's like LCD TVs, people also thought they consumed less power than conventional CRTs. Personally, I can warm my hands if I stick the palms up in front of my 32" LCD which chugs away at 152W when fully "lit" (powersave mode off).

Conservation of energy still applies here. It's not that LCD technology is necessarily less efficient (compare a scanning electron beam exciting phosphors to cold cathode fluorescent), it's that people have demanded (and gotten) much brighter screens, sometimes by a factor of 5-10 or more.

Obviously, if you're putting more light out of the screen, you're going to need to pull more energy out of the wall. There's no free lunch. If you care about saving power, turn down the brightness. Otherwise, don't sweat it.

(Incidentally, my old CRTs got quite hot, so there was plenty of wasted energy coming out of those, too.)

Re:Been there, seen that, got the t-shirt

[GleeBot]

Here's a handy rule of thumb when trying to decide questions like this: Energy in = energy out (conservation of energy).

Hand-waving a percentage of that for inefficiency losses, you can clearly see that, for a given level of brightness, the CRT has to be a quadratic relation, too.

If you try to scan a larger screen with the same total power, you're going to have to do it faster, imparting less energy per frame to any given phosphor on the CRT. In other words, the screen will be dimmer. You're spreading the same amount of energy per second (power) over a larger area. To make up for that, you need to make the beam stronger in proportion to the gain in area.

Of course, this assumes you're not overengineering at the low end. Phosphors can only be excited so far, so you could use the same power beam in a smaller set, and just waste the excess.

Re:Been there, seen that, got the t-shirt

[clone53421]

Yes, but you also have to consider the energy to run the magnets directing the beam. Since the deflection of the beam is only increased linearly, I don't think it will be a quadratic increase like the power to the electron gun. I don't know what percentage of the power is drawn by the electromagnets, but it's probably a good chunk of the total, so it should make the overall efficiency significantly better than quadratic.

Re:Been there, seen that, got the t-shirt

[toddestan]

If I had to guess, it's because people like their TVs insanely bright compared to their computer monitors. Turn down the brightness a bit, and the energy useage drops by a lot (assuming the brightness control actually dims the bulb, not all screens do it that way). TVs also have to deal with powering the speakers and the tuner, which could account for 30-40W by themselves.

Good article.. BUT...

[jonnythan]

The thing is that Tom measured battery life with the hard drives constantly working.

He glosses over this with the following statement:

Keep in mind that this benchmark keeps the system busy in several disciplines, and the results would of course be different if we measured the runtime in idle. However, this wouldnâ(TM)t quite reflect user behavior, as no one turns the notebook on just to wait around until the battery runs empty.

No, Tom, no one turns on their computer and simply waits for the battery to die. However, no one turns on their computer and has their hard drive constantly thrashing either.

Typical usage patterns include document editing, movie watching, music listening, etc, which involve very, very small amounts of hard drive access.

Use a better battery benchmark that leaves the hard drive idle most of the time, then come back and let us know how these drives fare.

Re:Good article.. BUT...

[Anonymous Coward]

Tom = Patrick Schmid and Achim Roos

Re:Good article.. BUT...

[Anonymous Coward]

However, no one turns on their computer and has their hard drive constantly thrashing either.

Unless they are using Vista.

Re:Good article.. BUT...

[neokushan]

I think the article is still somewhat Valid, though, even if it does gloss over some of the details a bit.
The short end of it is that a SSD will use less power when idle, but more power at full load than a typical hard drive will, which may be a real factor when deciding which to purchase.

Personally, though, I don't see the real benefits to them in 99% of situations. Their performance is only marginally better than standard HDD's and, as you pointed out, MOST people wont regularly thrash the HDD anyway. Plus, they're expensive - you could spend a little extra on a bigger and/or additional batter and some more RAM, get a massive regular hard drive and probably still come to less than half the price of a 32Gb SSD.

Hopefully when the technology matures a bit, they'll be to HDD's what the common DVD was to CD's.

Re:Good article.. BUT...

[initdeep]

funny
i bought my Dell xps m1330 with 3gb of ram, 64gb SSD and all of the other bells and whistles available at the time for lessthan $2k.

doesn't seem all that expensive to me.

Re:Good article.. BUT...

[Anonymous Coward]

According to TFA, a 64Gb SSD costs about $1300 on it's own.

Re:Good article.. BUT...

[mrv00t]

Typical usage patterns include ... movie watching, ... which involve very, very small amounts of hard drive access.

Apparently your pr0n collection is not as big as mine.

Re:Good article.. BUT...

[TheRaven64]

It's not that SSDs use less power when idle, it's that they can transition between idle and non-idle modes much more quickly. Spinning up a hard disk takes several seconds and uses more power than leaving it spinning. Turning an SSD on is just a matter of putting power into the flash chips. This means that you can turn of an SSD as soon as you stop writing, while a mechanical disk can only be powered down when it's not likely to be used for a few minutes or more (and, even then, degrades the user experience by making them wait for it to spin up again). If you are hitting the disk for a second once every ten seconds, then an SSD needs power for six seconds per minute, while a hard disk needs power for 60 seconds per minute. Looking at the disk activity graph on my laptop, this is about the usage pattern that I see most of the time.

Re:Good article.. BUT...

[Anonymous Coward]

One of the places where SSDs are expected to help the most is in web server environments where the faster access speeds would help reduce the overhead of server-side scripting. So, I think it's actually very, very valid to measure the power consumption at full capacity, since that's a potential large-scale use for the devices.

Re:Good article.. BUT...

[FireFury03]

I was wondering why these SSDs draw any power at all while idling - I guess they have to leave the controller powered up, but some of the figures quoted were around 2W when idle - you can get x86 processors that run on less than that FFS!

Re:Good article.. BUT...

[DFJA]

>However, no one turns on their computer and has their hard drive constantly thrashing either.

Yes they do- they're known as Vista users.

Re:Good article.. BUT...

[initdeep]

no.
http://www.crucial.com/store/listmodule/SOLID%20STATE%20DRIVE/~Solid%20State%20Drive~/list.html

and crucial isnt exactly a market leader in pricing.

and currently, the 64gb drive is a $500 upgrade at Dell on the m1330.

so you can get a T7250 proc, 3gb ram, 64gb SSD, LED display, DVD burner, 8400M GPU, wireless N, Bluetooth built in, 2MP camera, Fingerprint reader, etc all for less than $2k currently.
Thats without any special deal other than going to Dell's site.
Now through in a coupon or two and you can gain a fairly steep discount.

Re:Good article.. BUT...

[Lumpy]

However, no one turns on their computer and has their hard drive constantly thrashing either.

you haven't seen a 4 year old windows machine boot have you. Last one I fixed for a client the hard drive thrashed for 1 hour after boot and NOTHING was touched. There was so much spyware and other crap loading it was insane.

and that seems to be the norm for most consumers PC's... full of crap and constant disk thrashing for most of the time.

Re:Good article.. BUT...

[Thaelon]

That has nothing to do with the age of the machine or the operating system on it.

I have a windows XP machine at home that was up to 60 days of up time before the power went out. No perceivable slowdown in any application.

I would liken your example to running a car for four years and never changing the oil or other fluids, then complaining when it doesn't run well.

And one hour boot times are NOT the norm. At that point most uneducated users will think it's "worn out" and simply go buy a new one. Of course, most computers have very few parts that actually wear out, but the point is, your example is a lot less realistic than it first appears.

Re:Good article.. BUT...

[vrdlbrnft]

Hmm... I just got my mom a Dell with such a SSD.
32 GB is more than than she needs for Word/Excel and this is a one-save-per-minute load.
But she likes it, really, because:

• XP starts up faster as on a Desktop HDD
• Wakeup from hibernation is about 6 sec
• during work, you hear -- nothing.

Re:Good article.. BUT...

[Anonymous Coward]

A Prius works the same way basically.

At low power (speed) it does indeed save gas. I rented one for a 14 hour trip, got a great deal saved $130 in gas money. But.. what if I had driven 85 the entire time? That little engine would have put out maybe 15mpg.

Its more efficient for the majority of people, and given the usage context - ultra portables - I'd wager SSD provides more battery life.

Re:Good article.. BUT...

[mad_ian]

Actually, My girlfriend is terrible about opening her laptop, and just leaving it open on the table, letting the battery run down completely because she has the power-save and auto-sleep functions completely turned off "because they annoyed her"

~DW

Re:Good article.. BUT...

[Chrononium]

This is an interesting point, as if the SSD controller, if it is a repurposed HD controller, or OS is expecting to be driving a HD, then it might preemptively ask the controller to spin up the HD, just from the standpoint of power savings. It's not impossible for the OS to observe the long-term behavior of the current application set (or at least the active application) and it might be able to ask the disk to spin up periodically just in case. This is potentially a good thing for HDs, but a bad thing for SSDs. It could end up taking away a big part of the power advantage an SSD should offer.

Re:Good article.. BUT...

[Pulzar]

and currently, the 64gb drive is a $500 upgrade at Dell on the m1330.

And a 80GB "regular" HD is $39. $500 vs. $39 does seem very expensive to me.

Re:Good article.. BUT...

[Anonymous Coward]

It's actually very realistic. Most people HAVE that problem that are not computer gu-ru's. And the "its worn out buy a new one" is only a very tiny minorty in the rich demographic. Most people cant afford to dump $350.00 into a new pce every 3-4 years. Poor people outnumber you rich people 100 to 1. (make $80,000 a year? you're rich, get over it.)

I see it regularly as well. In fact I have a 2 week old Vista laptop in the shop right now that does that. (512meg ram 1.8ghz single core new Dell laptop)

Lumpy is right on. He must deal with small businesses and home pc's.

Re:Good article.. BUT...

[dbIII]

However, no one turns on their computer and has their hard drive constantly thrashing either.

32 bit Vista where the memory ceiling is almost on the floor so you need that virtual memory. Then again that's not a good way to run a laptop - you would want something that can give you enough real memory to run what you need or just use the thing as a typewriter.

Re:Good article.. BUT...

[Carewolf]

No neither a harddisk nor a SSD use any power at all when powered down. Idling means turned on but not reading, so the harddisk is still spinning, it is just not moving the heads.

Re:running a synthetic benchmark 100% of the time.

[Rakshasa Taisab]

Exactly... The thing about spinning platters is that it takes energy to start up _and_ keep it spinning. So obviously doing read/write 100% of the time would bias towards the conventional hard drives.

Hell, even read/write 10% of the time is too much for normal usage.

Author = Clueless

[Rhys]

He makes the claim in the comments about the article that "well who just watches dvds? You have to keep the system busy and test that!" That's about as valid as setting the machine not to sleep and seeing how long it can idle there.

On an ultraportable especially, you aren't going to be churning the CPU with a benchmarking program. You won't be rendering animation frames. Mostly you'll be in a web browser or text processing program, waiting on the user to type something. With occasional spurts of OS and program start/stop. Good gosh it sounds like a MIX of tasks, rather than either extreme.

Re:Author = Clueless

[noidentity]

He makes the claim in the comments about the article that "well who just watches dvds? You have to keep the system busy and test that!" That's about as valid as setting the machine not to sleep and seeing how long it can idle there.

I run disk benchmarks all day, you insensitive clod!

Re:Author = Clueless

[Chrononium]

At Apple (a few years ago), we would test typical portable battery life by scripting a set of tasks for the computer, both in OS X and in Windows. This way, we would be trying to simulate this common mix of tasks and obtain more realistic battery lifetimes (and comparison between similar Windows laptops and our own). Naturally, it wasn't always the case that our benchmarks were the ones put up on the web.

Stuff like this benchmark is really just an extreme corner case. As an engineer who relies on lots of hardware to help perform long and complex simulations, I know something about thrashing a system to death. And yet, I would never, ever, ever run such a continuous thrasher on my laptop (at least without plugging it into the wall). Their scenario seems (to me) as extremely unrealistic and may qualify as FUD.

Swap

[zippthorne]

Why are we still using swap? We can put, cheaply, more than 4 gigs of ram in a machine. With the differences in SSD, and the concerns about power efficiency, it really makes no sense to me that machines are still being designed to page out memory.

Re:Swap

[SDF-7]

Because in general folks know how to do it (so you're not adding code complexity in most cases) -- and it seamlessly handles the odd folks who *do* eat all of their RAM in a workload and end up needing it. (And since the folks with big working sets relatively to current states also tend to be the folks paying more money... they do get listened to -- and these are the same folks that would require a perfect patch in 24 hours when they're unhappy, so you're much better off having a little planning pre-release than trying to crowbar this in post).

The thing to do is to make the virtual memory subsystem as efficient as possible about handling swap statistics just in case you need it so that the folks who really don't need it aren't aggressively impacted.

(Note: I am a virtual memory subsystem kernel engineer -- but not on Windows. I make no claim about how efficient or inefficient Windows is at doing this as a result. I would seriously expect that since they're designing the core kernel to operate from laptops up through Windows Server Whatever --
they have to accommodate cases beyond the 4Gb in your laptop should be enough to keep everything in core, though. )

Re:Swap

[peragrin]

well In linux and OS X you can actually disable swap. Windows however relies on it heavily, and with Vista under 4 gigs of ram you need swap space to store the entire OS.

Re:Swap

[afidel]

All those chips for RAM require power too...

Re:Swap

[Ilgaz]

Used/tried a Quad G5 with 16 gig memory installed (mine has 4,5 GB).

Even on Leopard, that mainframe like monster will still create a 64MB swap file amusing us. I also saw 6 swapfiles has been created on my 4,5 GB RAM installed Tiger (OS X 10.4, doesn't have auto cleaning) after days of heavy load uptime.

Is it a strategic thing that OS kernel does? I remember reading some real weird stuff about Microsoft's tactic to swap the applications to disk even while they didn't have to but I can't find the file now. You seem to blame application developers which I can understand, does it have something to do with "pageins" and "faults" I see generally caused by badly written software?

Re:Swap

[Anonymous Coward]

With XP at least you can disable swap and in my experience it improves the performance since XP appears to swap stuff out needlessly. I set my mum's computer up this way and it hasn't caused any problems, windows will throw up an error message to let you know if it runs out of memory and this very rarely happens with my mum's usage which is mostly just light web and email on 512MB RAM.

Re:Swap

[FrankieBaby1986]

swapping is often done on a modern system in order to increase performance by increasing cache space. By swapping out applications that are almost never being used, like say, the quicktime tray or something, there is more ram available to cache program data for applications that you are using much more heavily.

Re:Swap

[K. S. Kyosuke]

Is it a strategic thing that OS kernel does? I remember reading some real weird stuff about Microsoft's tactic to swap the applications to disk even while they didn't have to but I can't find the file now.

No matter what amount of memory you have, it might be often more useful to swap out the application you have not touched for the last two hours and use the memory pages for the disk cache instead so that the apps you are using right now do not have to hit the disk constantly. It is quite possible that on Mac OS X, you can tune it somewhere (at least on Linux you can).

Re:Swap

[SDF-7]

Most Enterprise-class Unix OSs by default want swap "just in case". The defaults are that each new virtual object created (malloc / anon mmap, etc.) reserves the equivalent amount of swap so that if the system ever does get under pressure -- the OS will have to figure out what disk block/FS range/whatever it will use to page the memory object, but it will be guaranteed that it will find the appropriate resources available. As such - the OS is unlikely to ever really need that actual disk space (the performance would really stink if you're under that much memory pressure), but it does want to have plenty of room for swap-backed virtual objects... folks don't like their malloc() to fail when they don't think they're out of resources.

When you don't do things this way (by changing the defaults, etc.... and Linux I know is more flexible about this, though I don't know what's the urrent default state) you run the risk of needing swap when the system is already under pressure. Then either you fail paging out (which means that you're pretty well stuck... if you are desperately low on memory and can't make more -- you're headed for a perceived hang) or you start forcing memory objects to go away [typically by killing processes... and woe unto the poor process picked to die].

I wouldn't be at all surprised if Leopard is setting things up for this state (swap reservation requirements). As cheap as memory is, on a per Mb or GB basis... disks are still a whole lot cheaper... so you might as well try to plan for both safety from an application standpoint (i.e. not random death) and a large virtual address space consumption.

Re:Swap

[ErikZ]

My last game install was over 30GB. Age of Conan.

Re:Swap

[zippthorne]

And? Unless you've got 32GB of RAM, I really don't see the point of loading the entire game into memory. What's the difference, really, between loading 2 Gig into RAM + 28 gig into swap, and just loading 1-2 GB at a time, only into memory.

A game is an especially bad example as opposed to scientific applications. A game doesn't need to be particularly accurate, but it does need to have good performance. It should just drop stuff from memory when it needs it for more important stuff. No need to page-out, because the "stuff that changes" (e.g. position, inventory, game-state flags) doesn't take up nearly as much space as the "stuff that gets referred to a lot" (e.g. assets)

If the game needs to refer to a bunch of stuff, it should be aware of the available memory, and load versions of those assets with smaller memory footprints to make sure it doesn't go over.

It shouldn't make any difference whether you load assets from swap, or you load them from the files directly. Assuming that the install is on the same disk as the swap.

Engineering is a trade-off

[Anonymous Coward]

I have an old flat-panel display (NEC-Mitsubishi NXM76LCD) that takes in 25W at full brightness. It's far less than what current LCD models draw, but it doesn't scream about 2ms timing, either. (Find any benchmark, and a manufacturer will take it to irrelevance. Do you need 2ms if you're not refreshing at 500 Hz?)

Re:Engineering is a trade-off

[sexconker]

That's 2ms grey-to-grey (or worse...).

At best you're looking at 4ms (white-to-black) which is 250 Hz.

Re:Engineering is a trade-off

[vux984]

That's 2ms grey-to-grey (or worse...).

At best you're looking at 4ms (white-to-black) which is 250 Hz.

Black to white is actually generally faster than grey-to-grey, not slower. That's why the whole grey-to-grey benchmarks started showing up. Because screens that could go black-to-white in 4ms were easily available, but they still had grey-to-grey times of 32ms for some level transitions.

Further, most 'fast' screens use 'overdrive' which actually overshoots the target destination color (because larger transitions are faster than smaller ones), and then brings it back down to the target color. (leading to 'sparkle' when whatching movies etc because a pixel on a small transition from 'almost black' to 'just a little bit less almost black' shoots through medium-grey to get there.

Re:Engineering is a trade-off

[sexconker]

Yes to the sparkling and such, but as far as I know grey-to-grey is still very fast, and doesn't have the control issues that other positions have.
(Getting to 50% quickly helps you get anywhere else quickly. Getting to 12% quickly isn't as helpful.)

I believe the old standards for measuring were white-to-black (and back).
Basically, | - |
Grey-to-grey would be / - /

A few years ago (3? 4?) when the first "fast" displays came out, the bottom line was that the screens weren't faster, the measurement was changed. We're still stuck with the less-than-honest measurement, which was my original point.

Re:Engineering is a trade-off

[vux984]

Grey-to-grey would be / - /

I think grey to grey is rgb [x,x,x] to rgb[y,y,y] where x,y are in {0..255} and both are somewhere in the middle-ish... like x might be 155 and y might be 174. grey-to-grey is the time to transition from one grey to another grey. Not grey-black-and-back-to-grey.

I don't think there is an established standard or even de facto standard for which grey-to-grey transition they meausure, or how, and I cynically think vendors choose the intensities of the 2 levels for optimal measurements.

They really -should- report average & worst case g2g transition. Because if they can get worst case g2g transition to be faster than a 120Hz refresh rate (or about ~8ms) then that is all anyone needs. (Hell, even 16ms will be fast enough for a 60Hz refresh rate).

In reality right now you need to find independant benchmarks, and that's when you find that a Viewsonic Pro series screen advertised as '4ms' actually can do all transitions in under 9ms, while a generic '2ms' screen might do some of its transitions in the 2ms range, but also have a large number of transitions that take 15+ms.

In which case the viewsonic is a much more responsive screen in practice.

We're still stuck with the less-than-honest measurement, which was my original point.

Absolutely. I would go as far as to say, its misleading and pointless too.

Re:Engineering is a trade-off

[sexconker]

I've always thought grey-to-grey was going from [128,128,128] to [0,0,0] (or [255,255,255]) and back.

Either way, grey-to-grey is faster than a standard measurement (be it white-to-black or white-black and back, or whatever "standard" panel manufacturers use), almost always by a factor of 2.

When you find an LCD with specs that list 2 different measurements, the GTG or G2G one is slower. See http://www.newegg.com/Product/Product.aspx?Item=N82E16824014173 for example.

Either way, you're absolutely right - the specs are often more like speculation that specifications.
 

Re:Engineering is a trade-off

[vux984]

I've always thought grey-to-grey was going from [128,128,128] to [0,0,0] (or [255,255,255]) and back.

from:
http://www.presentationtek.com/2008/03/13/lcd-display-response-time/

"The two most common measurement methods are "rise-and-fall" and "gray-to-gray". A technical specification for rise-and-fall is available from VESA (the Video Electronics Standards Association). Rise-and-fall is measured as the time taken for a pixel to change from black to white (rise) and back to black again (fall). Gray-to-gray is the time to change from one shade of gray to another. Transitions between the fine graduations of gray-to-gray can be 3 to 4 times slower than those of rise-and-fall because of the lower driving signal for the transition."

and from

http://www.xbitlabs.com/articles/other/display/response-compensation.html

"It may seem that the process of switching from black to gray should go faster than switching from black to white since the crystals are turned round by a smaller angle, but the fact is the applied electric field not only determines the orientation of the crystals, but also the speed of their turning round. The force that affects the crystals is proportional to the square of the applied electric field. So it takes a four times smaller force to turn the crystals round by a two times smaller angle."

This again confirms that g2g (gtg) is slower than black-white.

And the reason it takes longer for g2g transitions is that you use smaller electric fields to turn the crystals, which in turn means the crystals orient slower. Thus the smaller the transition, the slower the transition.

Overdrive works because it hits them with a strong field to get them to quickly orient to a new position (in the wrong position), and then hits them with another strong field to get them to the target intensity, back near where they started. And it works out that these two quick transitions with strong fields is faster than one slow transition with a weak field.

Either way, grey-to-grey is faster than a standard measurement (be it white-to-black or white-black and back, or whatever "standard" panel manufacturers use), almost always by a factor of 2.

My experience, and the links above agree, is the opposite. I have no idea what the newegg ad is trying to disclose... maybe thats the gtg transition speed with and without overdrive? Or maybe its the rise-fall and the gtg with overdrive?

Re:Engineering is a trade-off

[Anonymous Coward]

You don't seem to know what overdrive is. You set it for one frame to a voltage that results in a steady-state value too far, but you pick that so it's dead on at the next frame. Then, at the next frame, as it arrives at the correct color, you give it the voltage for that color. No overshoot.

The only way you'd get overshoot is if you used response data for a panel with different characteristics than the panel being driven. And as sure as I am that no cheap Chinese panels would ever wind up being matched to cheap Chinese controllers with other response data, I'd recommend going with quality panels, just in case.

Re:Engineering is a trade-off

[sexconker]

I've always seen it listed as grey-to-grey being faster. When the first big jump in speeds came about, manufacturers sometimes qualified that as being based on a grey-to-grey measurement, and reviews would always mention what the true response time was (usually double of what was advertised.

If grey-to-grey means ANY change between greys, then the reasons for faster grey-to-grey times could include:

Specific manufacturing techniques to boost the speed of certain transitions.
Like I mentioned earlier, getting a 50% swing (0 -> 128 or 255 -> 128) could be very beneficial, since you could do a half transition quickly, then cover the rest of the distance.
0% - 52% = 0% -> 50% + 50%-52%.
Likewise, you could overdrive to 50% and then back off a bit to reach 48%. You're still playing against the power of the square, though, and any benefit would have to outweigh the the penalty of doing a smaller transition (2% in this case).

Treating [0,0,0] -> [255,255,255] as grey-to-grey. They're technically greys.
More specifically, the "standard" may have been based on some average or worst case scenario. Consider going from [0,255,0] to [255,0,255]. You've got competing fields now (in terms of direction) and different fields (in terms of strength) since you're not going from a grey to a grey.

Marketing hogwash. I don't buy the panels in bulk, and I can't be bothered to dig around indian and chinese sites to look up their true technical specification, if they're even published anywhere.

no one turns on their computer and has their hard

[DeeVeeAnt]

Except Vista users apparently ;>

Author = Clueless and Stupid

[slaker]

I have a 32GB SSD in my T61. My real life usage shows that I get between 30 and 45 more minutes of battery life out of my SSD-equipped notebook than on my other T61, which has a 160GB 7200rpm drive in it, when both of them are on the "medium" power saving setting in Windows.

Re:Author = Clueless and Stupid

[WillAdams]

My experience matches yours. I had a huge increase in battery life when I swapped out the HD in my Fujitsu Stylistic 2300 for a CF-IDE adapter w/ a 2GB (booting Windows 2000) and 4GB (data, swap space) CF cards in it.

Ran much cooler and was absolutely silent (no fan).

I was quite bummed when the system died.

William

Re:Author = Clueless and Stupid

[ascendant]

Ran much cooler and was absolutely silent (no fan).

I was quite bummed when the system died.

LOL. If these two things are related, I won't stop laughing for days.

Am I the only one...

[initdeep]

Who thinks that maybe the overall bettery life would be more affected by the HEAT the laptop is putting out?

After all, most laptops I've ever dealt with or owned are compact little furnaces that have their hard drives, cpu, gpu, and everything else shoved in as little space as possible in order to make room for things like batteries and keyboards.....

thus if they remove a high heat generating device (even more so with a 7200rpm drive) wouldn't logic also assume that you reduce the amount of heat needing to be removed from the system?

thus the fans would run less often, and therefore drain less battery power.
not too mention the proposed savings by simply accomplishing tasks "faster" when it comes to disk access.

Re:Am I the only one...

[MozeeToby]

Well, since power usage is directly related to heat output, we could benchmark SSD's versus standar drives much more scientifically; we could even use the author's broken benchmarking tool. Put the hard drive into a calorimeter, run the benchmark for 10 minutes and measure the total heat output. Repeat for the drive you are benchmarking against and presto, actual, scientific numbers on power usage for just that device, rather than simply plugging away until the battery runs dry.

Re:Am I the only one...

[compro01]

For the most part, power use = heat dissipation. Practically all the power that is used is lost as heat (with a little going towards light).

Re:Am I the only one...

[Anonymous Coward]

>thus the fans would run less often, and therefore drain less battery power.

While you are technically correct, a laptop fan uses about 1 watt of power at full tilt. Laptop batteries are typically in the 70 wH range

In a system running 3 hours, with a constantly running ran at maximum speed, that means about 4% of the system power is being sucked down by the fan. Or 8 minutes less runtime.

Re:Am I the only one...

[atraintocry]

For two devices of such different design, I wouldn't expect the efficiency to be the same. One might put out less heat while drawing more power from the battery, for instance. So I'm not sure what measuring the heat would tell you about the rest of the power consumption.

Re:Am I the only one...

[Kuvter]

Offtopic: I think an overall "bettery life" would be getting out of the house more often, and not infront of a computer screne. But yet I still read /. to help maximize the time well spent while not outside.

Re:Am I the only one...

[MozeeToby]

It's called conservation of energy; in a closed system (such as a colorimeter) all the energy will be converted to heat.

Re:Am I the only one...

[Anonymous Coward]

Or measure the current used with an ammeter on the battery (or the AC/DC converter with the batery removed).

But still, other points of contention remain, such as the SSD computer getting more work done since it's faster at I/O.

Or the fact that normally computers will read a little from the drive and then stop and then read a little more. In a continuous scenerio, the HDD gets a better score than normal use. In normal use the platters need to spin up and stop and then spin up again, wasting power.

Re:Am I the only one...

[Kaenneth]

I thought about that previously, but what about energy not lost as heat? Such as through the electromagnetic spectrum, or though physical force/vibrations. I know hard drives can act a bit like a gyroscope, in that a spinning one is harder to manipulate than a stopped one.

I would expect SSD's to be utterly silent, but I don't know how their RF noise compares to hard drives (electro-magnetic motor?)

Oh my

[Endo13]

And here I've been, thinking all this time that the real benefit to SSDs was that your laptop could fall off a table while running without killing the hard drive. Silly me.

Re:Oh my

[Ilgaz]

If you have Macbook Air and I don't recommend trying it, you would be amused to figure it still cares about sudden motion sensor blocking the access to SSD drive when it falls.

I have read about it on a Developer's blog who has been really surprised. It could be Apple's philosophy to ship the exact same OS X on every machine, there is a plist tweak to disable the sudden motion sensor.

Operating systems and their disk requirements

[tepples]

if your computer has enough RAM, it will access the drive quite infrequently, especially in many cases where power drain would be of concern, such as in those UMPCs.

...which tend to have reduced memory.

...and which ship with operating systems designed to run in reduced memory. Puppy Linux, for instance, hits the disk less than Windows XP on the same PC.

I really haven't found this to be true any more. My computer hits the disk pretty much all the time (for logging if nothing else)

Then increase the buffering on the logs. You can get away with it on a laptop because a laptop has a built-in 2-hour-plus UPS.

Re:Operating systems and their disk requirements

[drinkypoo]

Then increase the buffering on the logs. You can get away with it on a laptop because a laptop has a built-in 2-hour-plus UPS.

I would, but I'm running hardy-proposed, and I want to know why my computer crashed.

Good idea, though.

Re:Operating systems and their disk requirements

[tepples]

Then increase the buffering on the logs. You can get away with it on a laptop because a laptop has a built-in 2-hour-plus UPS.

I would, but I'm running hardy-proposed, and I want to know why my computer crashed.

If an app crashed, commit the log buffers. If the kernel crashed, is there a reason you can't wait to troubleshoot this until you're at mains power?

Power Efficient? They can barely aim!

[SDF-7]

Seems unlikely to be all that power efficient to me... hauling around that huge bulk, and it sure didn't seem like Executor really made all that difference at the Battle of Endor from a capital ship perspective. (Probably because most turbolaser batteries seem to have really lousy guidance,
after all).

Oh... you meant those SSDs.... my bad.

Re:Power Efficient? They can barely aim!

[dkf]

To the person who modded the parent Informative, I salute you on behalf of Slashdot!

Re:Power Efficient? They can barely aim!

[Guysmiley777]

(Probably because most turbolaser batteries seem to have really lousy guidance, after all).

Hey! YOU try shooting down a mosquito with a .50 caliber machine gun some time and let me know how it goes. Who could have predicted the pesky rebels would resort to using FIGHTERS?!

SSD can keep CPU Busy

[Anonymous Coward]

Maybe while normal hard drives can indeed halt CPU consumption while waiting for data, U will wait longer to get things done, and that time is wasted.

Old and inaccurate

[jlp2097]

The article states nothing new - there are two very interesting blogs from Lenovo which already stated the same in August 07 (!). To quote:

Solid state HDDs promise to save power compared to traditional hard disk technology. And they will. However today's generation of SSDs have no power savings benefit compared to traditional HDDs. The big reason is that current SSDs with a Serial ATA interface are actually Parallel ATA hard disk drives with a serial bridge chip. They don't offer support for low power interface states and the architecture has a potential for data-losing error conditions when recovering from a low power state like suspend or hibernate. In the future, there will be native SATA solutions which will solve many of these problems and will at the same time offer a real power savings benefit which should increase battery life.[1]

An updated quote from a newer blog:

Power Consumption - All SSDs are going to save you battery life on your notebook, but some will save you more than others. Again, the native SATA drives will give you better battery life.[2]

To summarise: old news and mostly outdated with very recent SSD drives.

[1]: SSD part 1 (Aug 07)

[2]: SSD part 2 (March 08)

Check Page 14

[pjrc]

In TFA, there is a graph on page 14 with power consumption measurements for the 5 drives tested.

The SanDisk SSD shows 1.0 watt active, 0.5 watt idle.

The Hitachi drive (magnetic) shows 3.2 watts active, 1.1 watts idle.

So even if the SanDisk drive spent 100% of its time in active mode and the Hitachi drive was always idle, the SanDisk drive should still provide longer runtime.

However, their runtime test (page 12) shows 7:03 runtime with Hitachi, 7:02 with the SanDisk.

All they have to say about this is:

Most of the power consumption measurements are in line with our results in Mobilemark 07. However, it has become clear that idle and maximum power do not provide the full picture when we talk about flash SSDs.

Well, something clearly is wrong here.

Re:Check Page 14

[Rhys]

Idle cycles while waiting on the physical HD to do something = power savings due to speedstep/descendants of it.

No big mystery, just an author measuring something that's useless and irrelevant to make a splash.

Re:Check Page 14

[Anonymous Coward]

Yes, this is wrong.

If you have a 25W CPU and 1W to 3W disks then
the quicker disk will have less battery because the CPU will be less time waiting for IO and more time working (consuming battery).

This is a completly wrong test!

Of course they're better

[jollyreaper]

SSD's are like 8 frickin' times longer than ISD's, of course they'd be more powerful.

iPod counterexample

[sunderland56]

The battery life of a flash-based iPod nano is basically identical to that of a HD-based iPod Classic. However - the battery in a Classic is much larger.

Obviously the use model for both devices is the same.

Upgrade

[Anonymous Coward]

No, Tom, no one turns on their computer and simply waits for the battery to die. However, no one turns on their computer and has their hard drive constantly thrashing either.

I guess you're not on Vista yet?

Re:running a synthetic benchmark 100% of the time.

[Anonymous Coward]

What? If the bulk of the power for a conventional HDD is in the spinning of platters, shouldn't constant spinning of the drive bias the results in the other direction? I think what this test shows is that the power consumed by such a drive should average out over time such that the motor isn't using much more power then flash is to write.

Fan

[travelmug]

What about the energy savings from reduced ammount of fans?

Re:Fan

[deroby]

I wouldn't be all that surprised if it turns out that the INCREASED need for cooling caused the wicked results of this test.
=> SSD is faster at delivering info from 'disk'
=> CPU spents less time in 'kernel wait mode' due to I/O bottleneck
=> CPU heats up more because it does more work / time
=> fan(s) draws more power to exhaust the extra generated heat from the cpu

Re:Fan

[emj]

~0.3 Watt according to the Thinkpad X60 battery optimization guide

Thinnovation

[Anonymous Coward]

That was a bold move buying the first notebook from apple containing a NAND flash hard drive. You know how it usually goes for Apple early adopters...

"now take a deep breath and bend over...this will only be uncomfortable until version 2 comes out."

Already debunked....

[Junta]

The testing methodology was flawed to draw any conclusions. The problem is the CPU may have been more active due to less IOWait states. AS a resulte, the drive consumption may be lower, and the benchmark was not throttled to the platter disk performance. The benchmark might have run many more times during the test.

Take with grain of salt

[John Jamieson]

The title "Some SSD devices are inefficient" just does not drive page clicks. So Tom sensationalizes some facts, omits others, runs questionable synthetic benches and Voila... Slashdot delivers the page hits.

Tom sometimes takes known "problems" and dramatizes them. After all, an objective view should not be allowed to stand in the way of revenue generation.

As others have pointed out.
1. The reason for inefficient drives was usually power hungry interface logic chips.
2. The newer drives are all better according to a Lenovo blog (thanks jlp2097)

And to top it off the numbers don't even add up (see PJRC's nice post above).

Re:Take with grain of salt

[Eil]

This is pretty much the very reason I stopped reading Tom's Hardware about 6 years ago. They went from the useful site with good info to sensationalistic 20-page stories with no (or worse, misleading) content.

Valid Data -- Yet Invalid Conclusion

[Anonymous Coward]

Intel rolled out a 100+ person pilot program of it's currently sampling SSD. Users noticed a giant increase in productivity for HD-limited activities and significant battery life increases (25%+).

While the data from the experiment is valid, it's not reflective of a client workload, where a user produces x spreadsheets and y emails in a typical day. Under the experiment performed, you're performing a poor-man's version of a enterprise workload, where the results are well known -- SSDs won't save power per server. However, since enterprise workloads are normally severly HD limited, this may lead to less servers required to do the same amount of work.

I actually believe they are wrong!

[DRAGONWEEZEL]

Here is why...

What is the power consumption / bit processed?
They didn't state that.

If the drives are that much faster than spinning disk, they are procing (whops slipped into wowspeak...) I mean they are doing more work per mA making them more efficient. It's like in a car... j/k. (about making another car analogy)

Re:I actually believe they are wrong!

[Anonymous Coward]

I don't know if you made it as far as page 14, performance x battery runtime index:

http://www.tomshardware.com/reviews/ssd-hdd-battery,1955-13.html

The real issue appears to be that SSDs don't use their "idle" rating as often as spindle drives... yet.

Re:running a synthetic benchmark 100% of the time.

[Rakshasa Taisab]

No... Let's say you use 10 minutes of constant read/write activity for your benchmark.

Now, consider how much power this uses. Probably going to put the flash and platter HDD's quite close in power consumption. (As this test shows)

Then consider a RL example; writing the same amount of data, intermittently over a period of an hour. The flash drive uses pretty much the same amount of power, while the platter drive needs to keep itself spinning for 50 minutes extra.

Now, a normal laptop would probably be doing it's hibernating, etc, directly from ram, thus you'd probably get even more power efficiency out of the flash drive than the example above.

Re:running a synthetic benchmark 100% of the time.

[Perl-Pusher]

Your correct also as a previous post pointed out here The CPU on a disk machine should be idle more. Count how many times the program ran in the time tested. If a program runs straight through to completion say twice as many times as the hard disk. Then it really isn't less efficient.

SSD vs ISD

[8ball629]

I'd say that the SSD is most definitely NOT more power efficient than the ISD given the enormous size difference.

how about this theory...

[Anonymous Coward]

When the disk is seeking, the CPU is idle.
This was a test of how long the battery lasts when running a benchmark, not how far through / how many iterations of the benchmark it performs.

It's conceivable that due to the lower access time of the SSDs the CPU is spending less of it's time idling, and therefore consumes more power per given unit of time.
As the CPU consumes a substantial amount of power this could explain the discrepancy.

Re:running a synthetic benchmark 100% of the time.

[NorseWolf]

Hell, even read/write 10% of the time is too much for normal usage.

You obviously aren't using Windows Vista.

I thought it was a good article

[BlueZombie]

Seems like many of the objections can be roughly translated as: "If Bob can toss 60 shovels of dirt a minute and Ray can only toss 50, and both walk out in the same field and each dig a ditch for an hour ... " Tom's site asked "which one consumed more calories?" Instead of "which one moved more dirt?" Either is a valid measurement to take. What I take away from the article is not "OMG they LIED" or even "OMG Tom LIED". Instead, it is that different SSD's have widely varying performance and power profiles that may or may not be better than more traditional solutions for any given task.

No.

[je ne sais quoi]

No, that's not right:

Power Consumption : 80GB 4200RPM | HDD 64GB SSD
Read/Write : 0.9W | 0.24W/0.36W
Idle : 0.30W | 0.035W
Standby : 0.07W | 0.035W
Sleep : N/A | 0.015W

The power savings are dramatic [for the macbook air]; idle power is reduced by nearly a factor of 10, standby power is cut in half and read/write power is at worst a third of the mechanical drive. The 10x idle power reduction is important since that's where your drive should be spending most of its time, and if the average power savings is somewhere in the 0.25 - 0.5W range you're looking at improving battery life on the order of tens of minutes.

Source.

You fail.

SSD + Cameras?

[Anonymous Coward]

From the use of my digital camera I have the feeling that SDHC consumes a lot of power. After a transfer of ~4G photos much of the battery is gone.

So, if the underlying technology of SDHC and SSDs is the same then SSDs should consume quite a lot of power.

Wish they'd hire me to do the benchmarks instead.

[Gldm]

I'd probably do a better job, considering I've caught them doing benchmarks wrong before.

But hey, why be right when you can generate a controversy that attracts millions of hits instead?

Try Concurrency Testing on SSD

[GeekMarine72]

After much tinkering, I was able to slide a Samsung SSD 32GB drive into my T43P laptop. (Yes, duct tape was involved). The SSD benchmarks faster. Actions like deletes are laughably fast. Battery runtime duration is noticeably longer (as my activity hasn't changed). The one thing I am certain of is that the SSD / OS (Vista) struggles with concurrent reads / writes. I am replacing it with a very fast and well rated Western DGG HDD as I've found the SSD not what I was expecting. GM

AHA!

[wonkavader]

Yes, that seems to make a lot of sense.

The SANDisk part which had a lower theoretical wattage even in activity than the idle of the HDD shows that there's a missing factor, and this seems like a pretty good guess at what it is.

And this brings us back to the need for a battery life benchmark which has wait states in it, so that the performance of the machine using HDD and SSD is forced to be equal.

Power savings? What?

[Aphoxema]

Huh, I've heard some things about SSDs but I don't remember hearing any mentions that they're supposed to save more power. The only advantage I thought was size (and shape, you could make an SSD a 12"x1"x.25" or 3"x3"x3", but a circle will always get more space inside a square) and no moving parts so shock wouldn't be as dangerous.

It's switching billions of little semiconductors, it's going to take power to do that. There's sure to be less inductive reactance than two motors and a head that goes ZAP ZAP ZAP, but not (a whole lot) less power.

Its not all about laptops

[unity100]

majority of computer usage still goes on on desktops. and at a desktop, if solid state can deliver twice the speed of a conventional hard drive, i dont give a damn about power saving. hard drive is the slowest component of a pc as of now. speeding them up has much importance.

Or even real world intensive use

[Sycraft-fu]

Even when I use something that hits my drive hard, it is far from continuous. In fact flash drives might have a chance to shine here as well as often one of the things that increases activity is the need for random access, which magnetic media isn't all that fast at. For example my harddrive works much harder than it has to when I'm doing audio mixdowns because it has to access multiple wave files at the same time. Most of the time is spent seeking from file to file, it takes comparatively little time to read the actual data needed. Flash, of course, has much faster random access. Thus for something like this it'd be loaded lighter, despite it being an "intense" use.

Re:Or even real world intensive use

[gfody]

I wonder if they used FAT32 for the SSD and NTFS for the HDD. I think a lot of the potential efficiency in SSD is being pissed away by file system drivers optimized for HDD.

You're ALL wrong!

[krazul]

I was about to jump on the 'tomshardware are fools' bandwagon but guess what: they are absolutely correct.

The chart on page 13 is multiplying the mobilemark07 result (some number)* battery time. many people are assuming that the SSDs finished the mobilemark07 test many more times than the HDD before finally using up the battery capacity. but what they are missing is that the mobilemark07 result is already taking into account the time required to finish the work.

in other words, there is a fixed amount of work being done, and mobilemark measures the time needed to complete the work. and returns a number based on that.

so it may be true that the SSD is finishing the benchmark faster, but not fast enough to offset the amount of energy it is using up to get that fixed amount of work done.

let's say my name is SSDguy and twin's name is HDDguy.

Tom gives us each an identical stack of 250 blank laser printer pages and says, ok now you each need to make identical paper airplanes out of every piece of paper.

HDDguy is old tech. He uses his hands and folds the airplanes. he does a pretty good job. 280 airplanes in 10 minutes. AirplaneMark08 score = 28.

I, SSDguy am new tech. I can fold them with my mind. no messy moving parts (arms). 300 airplanes in 10 minutes. AirplaneMark08 score = 30.

Next, Tom makes us run laps until we collapse and fall asleep. When we wake up, he gives us each the same meal and he places us in front of an infinite supply of page stacks and says "start folding". Assuming we have the same amount of energy stored in our bodies, we both fold until we collapse.

I, SSDguy stop first, after 6 hours (360 minutes). HDDguy stops after 6.5 hours (390 minutes).

So now you say, "Well, SSDguy works faster, so he easily did more work." Only problem is that if you do the math, HDDguy beats me by a small margin, similar to the tomshardware results. If you don't believe me, do the math yourself.

Is this a kinder, gentler Slashdot?

[Provocateur]

All the acronyms are well-explained

SSD check.
SLC check.

HDD er

Must be the new guy.

I was under the impression...

[tuvoky_wo]

...that SSDs were designed to be more durable than HDDs due to lack of moving parts. If they use the same amount of electricity as a HDD for the same operations then so be it.
Perhaps someone will design a SSD which only powers memory regions which are being accessed at the time as opposed to the all-on/off nature of current drives.

What's the power for a web server?

[reiisi]

Especially for a web server serving mostly static content?