Word of the Day: psychurgy. Mental energy or function
|
|
|
Solid State Drives suck
TOM’S BOYS ARE CASTING some disbelief magic on Intel’s claim to SSD’s battery life goodness. This one’ll be hard to digest at Satan Clara, but considering it’s true, it does throw a spanner in the works. The review claims that most “mainstream” SSDs actually reduce battery life instead of increasing it. They ran the numbers, they’ve got their stats up, and a lot o’ people will be plenty confused. If this sticks we smell a class-action in the works. Grit your teeth here.
Here’s an interesting one from Trusted Reviews: the Data Robotics Drobo & Droboshare DAS/NAS kit. The Drobo is the direct attached storage but couple it with the Droboshare and you get a network attached storage. It’s a clever bit of kit that rearranges your storage as you go. Pop in a hard drive (up to four) and it starts adding up storage and arranging redundancy automatically (according to your criteria, we presume). It’s performance is sub-par, though, but it’s an easy, no-fuss, solution. Read Dave’s review.
X-bit’s review of the HD4850 passed us by last week, but they’ve still got plenty to say about the mainstream card. Their reasoning is quite interesting, as the HD4850 launched at $199, while its competitor (the 9800GTX) launched at $349, initially, meaning the HD4850 still has some room for price maneuvering. Oh, and they also thought it kicked bottom. Read the review here.
You know what Microsoft does really well? We mean, really really well? You already know the answer isn’t “operating systems” – it’s “mice”. Yeah. For as long as we remember, Microsoft mice have been top-notch rodents that kept the peace between users and GUIs. Hardware Logic is reviewing Microsoft’s Wireless Laser Mouse 7000. The 7000 has a rechargeable battery and a flat recharging base, which is mostly what it’s about. You can buy this one for about $70. Read Matthew’s review here.
The bods at Techgage are engaged in a CPU cooler royal rumble. Using their thermal-handicapped QX6850 quad-core, they’ve sent off 22 coolers to the slaughter. William is trying to separate the boys from the men, when it comes to coolers, subjecting them to extreme conditions – but like he says: even most of the coolers that failed will still work well for the casual overclocker. As you can imagine, there is some cruel and inhumane treatment going on, right here.
TweakTown is fondling the Mvix MX-780HD kit, a media center device for your home entertainment system. It’s a replacement (sorta) for an HTPC, includes HDMI and 1080p decoding support, as well as an integrated HD. Mike enjoyed a stutter-free experience with this one, and all you need is to hook it up to your net (non-Gigabite) or 802.11g and send off the files. You can also add storage devices as it has USB host functionality. It seems like a great appliance, worth having if you can spare the money, it’s $349-ish (AUD?). Read on.
Lenovo’s tiny Thinkpad U110 is on display at Think Computers. The teeny 11.1-incher is a bit on the expensive side, but Bob thinks it’s got a lot going for it, like the LED-backlit screen and the 7-cell battery that almost gets you a full day’s work before requiring the almighty power socket. Go here if like to ogle expensive stuff.
Ryan at PC Perspective has rounded-up a modicum of 9600GT cards and put them through the paces in a single- and dual-GPU test. It’s SLI on budget, something the 9600GT has been doing handsomely for a while now. Crysis and COD4 are showing some pretty darn good results, with Crysis being almost playable at 1920x1200/High/8xAF… and that’s saying a lot already! Well, you better look at the numbers yourself. Catch them here. µ
Tags: Intel
Share this:
Share
If the MemoRight can do twice almost twice as much work and last almost as long as the Hitachi drive then it is close to being twice as efficient.

If the MemoRight and the Drive were subjected to the same workload then presumably the MemoRight would have spent almost an extra 50% of the time idling.

If you are going to compare power consumption then you have to compare equal workloads otherwise you'll just end up with meaningless numbers.
You're partially right.
Yes, it's not a normal situation for a laptop to be using 100% cpu al the time.

ButTom's test of choice runs an cpu/disk intensive task all the time.
If Flash drive is faster it will also put more data to CPU for calculations and this way you have to notice that CPU takes more energy when working (flash drive) then when it waits a lot for data (~10ms hdd access time)

So basically Tom's article is comparing something like:
50% cpu busy (say it takes 20W) + hdd
vs
90% cpu busy (say it takes 25W) + flash drive
and the higher power surge of "flash drive based laptop" comes from the higher power consumption by a higher load on cpu.

That's why the only real way to make a fair benchmark would be to have a predefined task (let us say open and close document 'X' 100 times) and the rest of the time laptop would be in idle state.
If You Read Specs You'll Notice that Testbed is 2- DDR2 400 Mhz/s ECC/REG(3,3,10) memry. Others use 667 ddr2. yet all compared equaly.So Highly Flawed TEST, If TEST At ALL.

Basicly Take SSD thats Runing 100 Mhz/S Data Output & try To Cram it into memory Thats 1/20 Speed for Read/write attributes & Something Gonna heat Up. SANDisk took Punishment Least well, using 3X More Power than other SSD Devices' to Control Rampage. Hey-Internal Flame War.Bummer.

Since Testbed was Home Built Roos, Rue is on You. theres always possibility entire system was corrupted in Home building Process, Wiping Ones Hands On Memory Stick could Be Another Humuiliation Laugh. Why. So You Don't Get IT?
SSD Prices Took Another tumble to FourFifie$ per 128 GB. here:
The firm is releasing the drives in 32GB, 64GB and 128GB capacities, and a claimed 120-143Mbs/80-93Mbps read write speeds. Seek times are less than 0.35ms, while the drives cost $169, $259 and $479 respectively IS THAT GOOD?
Respectively, Must Mean Ultie_Tom.
drashek

Also I posted Nvidia 9800 GTX at its last week price of $300, While I already posted it as its' true MSRP $199 earlier in week. Charts:CATCH UP!. Dumbo Hobos'
Ahso:Conroe Was Delayed Due to War IN Israel in 2006, contray to Intels Claim, giving X@ Run for Its money All Fall'6. BoToxETA is AKA BoToxEXO Developer, therfore NAME.Cheers. PPS I tried to post this elsewhere in AM, Yet Was Rejected, as Hardware Review Actually Taking comments today. edition2.
MirekCz, you bash the article for making an assumption and the turn around and make a big ass assumption of your own.

You're utterly wrong to say that performance = work and then go one to describe why the MemoRight device is so wonderful because although it shortens the battery life, it let's you get 2.25 times the amount of work done.

The simple fallacy of your argument is that no matter how fast the PC, the amount of work done with it depends not only on the component performance, but in far larger part on the performance of the user. Unless the laptop in question is simply sitting there churning through some huge processing task, the extra performance of the SSD is not worth the shorter battery life because a normal user will not see a productivity increase if the HDD in their system is 2.25 times faster, they will save a few seconds here and there, but since most people multi-task while using a laptop, it's fair to say that whatever benefit that extra performance brings it will get lost in the wash.

Let's think about this for a moment. If loading Word takes 10 seconds for the HDD, the MemoRight night just conceivably take 4 seconds. Whoa, 6 seconds. It'll be faster getting the document as well, but unless I'm opening War and Peace, I won't notice as my hand will still be recovering from pressing the OK button when the file loads. And so on, and so forth. This is exactly why that hugely high performance doesn't matter as much as perhaps it should, and battery performance matters far more.
I haven't looked at the details (that would be cheating), but I suspect that the benchmarks would have heavily stressed the system, continuously. They usually do. 

Such tests may be relevant to a server disk drive, but are NOT typical for real world use of a laptop.
Toms hardwares Test is actually not believable at all.

Actually it depends on the Power State if a Flash Drive can save you some energy. and in fact it can save a lot. If your computer is tuned towards Harddisks its no wonder that flash drives loose. Tune the power states to a more SSd friendly level and it will shine. For example it takes virtually no time to wake up an SSD from standby whereas a HDD need several seconds. Thats why you can always go into HDD power down with an SSD after a few seconds.
If you do not use fast Power down states for SSD and use a Benchmark which writes often enough on the SSD that it never reaches its standby state, no wonder the conventional Harddisk will win. Its basically apples to pears as you compare the one hardware in its highest powerstate to the other in its average to low powerstate.

Also you didn't mention the biggest advantage of SSDs being rugged. So no need for Lenovos Gyrosensors anymore to protect your precious Data. Chanche of a Headcrash is zero. If talking about SSD and HD, this should definitely be mentioned.
Also SSDs boot a lot faster and recover a lot faster from hibernation modus, thanks to their preferable random reads performance. That saves you a lot of time an energy, because nootebooks powersumption is very high while booting. Consider the time and how often it is booted during the lifespan of one Battery Charge.

The Author has obviously no idea what affects most when working with a Notebook and it seems that his understanding of Storage technologies is as little as a Walmart Salesman, cause the indept understanding of the technology and it benefits are missing.
The more I look at flash, the more I think it should be relegated right back to the fancy thumb drives we all like to keep in our pockets. 

These drives clearly arn't designed for large amounts of random I/O and personally I've had 2 drives that failed (they would format, but files would corrupt almost instantly) on me after 3-4 months of using readyboost on them. A page file is much the same idea and I can see problems devolping in a year or two from now when all these $1000 flash drives start to succom.

Don't get me wrong though, I just bought a flash drive for my desktop, but the only thing that its getting is C:\windows (or as close as I can get to just that dir) , with everything else on a recording medium that I can actually trust.
heavy loads of work for hd comes when the system have less thaan 2 gb of ram, or when it is performing an antivirus/defragmemtation/scandisk scan... personally i wouldn't perform a scan on battery power :) so tom's seems fair to me., calculations?, well i didn't check those, i just care for battery life (besides I don't have enough free time).
Erm ... If you try to reproduce the numbers in the graph on page13 by multiplying the performance figures by the runtime from the graphs on page12 you end up with very different numbers ...

MemoRight 117.4*((6*60)+38) = 46725.2 = 2.26
MTron 68.5*((6*60)+6) = 25071 = 1.21
Hitachi 48.9*((7*60)+3) = 20684.7 = 1
Crucial 40.3*((6*60)+3) = 14628.9 = 0.71
SanDisk 23.6*((7*60)+2) = 9959.2 = 0.48

The last number is the result when compared to the non-ssd disk since that is the purpose of the article!. It would appear from that that the memoright is 2.3 times as efficient as the hitachi because it almost matches the hitachi in runtime but completely thrashes it in performance. Since two SSD's are below the hitachi and 2 are above I'd say it's pretty inconclusive myself!
Yeah.. look at results on page 13 (vs my calculations).
They don't know how to multiply two numbers well.
They took one (random?) number and were multiplying drive time by it.

What they should have done was to take each drive performace and multiply it by corresponding driver runtime...
Just look at page 12 of Tom's flash test - mobilemark results.

Score for hdd: 48.9
Score for memoright: 117.4

So in the same period of time computer based with flash drive can do 117.4/48.9=2.4 times the amount of work! (I expect this work is heavily dependend on disk speed)

And what's with time?
HDD based laptop: 7:03 (423 min)
MEMORIGHT based laptop: 6:38 (398min)

So yes, theoretically flash drive takes more energy..right?
NO! Wrong!

In 6:38h flash based laptop can do 2.4*398/423=2.25 times the work of an hdd based laptop!

So yes, your flash based laptop will go out of energy sooner... but it will do over 2x the work of an hdd based laptop in the meantime.
The actual shorter life of flash based laptop comes to higher energy usage of the rest of the system that isn't stalled all the time by hdd access time.

Maybe Tom will try a benchmark where there's an X amount of work to do and after that laptop can go to idle mode for the rest of the time? This will be an apples to apples comparision.

Bad Tom.. no candy for you.