Forgot your password?
typodupeerror

Comment: Re:Depends on the specs. (Score 2) 168

by Andy Dodd (#47973135) Attached to: Do Specs Matter Anymore For the Average Smartphone User?

Yeah. Windows, as far as I can tell, shows iowait CPU as normal CPU usage.

Linux, at least, shows iowait usage in a separate bin, letting you know when you're I/O bound.

Nearly every time I've found my system unresponsive/slow, I've noticed my CPU utilization bar on my system monitor widget is almost entirely green. Green = iowait.

In a number of cases, the iowait was high because my system was swap thrashing. If your system bogs down under heavy multitasking, it's much more likely you need more RAM and not more CPU.

Comment: Re:Because... (Score 2) 168

by Andy Dodd (#47973103) Attached to: Do Specs Matter Anymore For the Average Smartphone User?

Yup. In many cases, the newer SoCs in phones have improved performance-per-watt.

Not always though... If you want amazing performance-per-watt, you don't want a flagship SoC, you want a midrange one. The quad Cortex-A7 Snapdragon 400 blows away any member of the 600 or 800 family in terms of battery performance. This is, among other reasons, why most of the Android Wear OEMs have chosen Snapdragon 400 units and disabled the unnedded cores. (Motorola was the only exception, and their usage of an OMAP3 has led to major criticisms of battery life since it's made on an ancient manufacturing process and the Cortex-A8 is significantly less power-efficient than the A7 even on the same manufacturing process.)

I have a device with a 2.5 GHz Snapdragon 801. Most of the time I've capped the CPU frequency at 1.5-1.7 GHz and don't notice a difference.

Comment: Re:Battery life seems to be a killer (Score 4, Informative) 87

by Andy Dodd (#47835657) Attached to: Moto 360 Reviews Arrive

The whole "partial discharges are bad" thing dates way back to the old days of timer-based NiCd chargers. Back then, chargers were dumb - they terminated the charge cycle based solely on a timer, and not based on detecting a full charge. The infamous "memory effect" wasn't actually the thing at play in most cases there - it's actually extremely difficult to reproduce outside of a lab. The problem was simple overcharging due to dumb timed chargers. As far as users were concerned, the symptoms were the same as "memory effect" so the myth stuck.

In general, partial discharges/recharges of a lithium-ion are far less stressful to it than a full discharge/recharge. However - li-ion/li-po batteries tend to lose capacity more rapidly if they are routinely kept at a high state of charge. Lithium batteries are "happiest" when they hover around 50% state of charge. (This is one of the reasons Teslas default to only charging up to around 80% unless you specifically "top it off")

Lead-acids like most non-EV car batteries are quite different beasts - they are happiest when fully charged, and will lose capacity RAPIDLY to sulfation if let to sit when partially discharged.

Comment: Re:Phoronix = fail (Score 1) 294

by Andy Dodd (#47806353) Attached to: Ask Slashdot: Linux-Friendly Desktop x86 Motherboard Manufacturers?

I've had issues with Gigabyte products in the past - but never with ASRock or Asus. I have no idea why people are saying ASRock/Asus are out. I have a relatively new Sandy Bridge system with UEFI and no issues, and also have a headless Haswell system with UEFI and both are rock solid. Neither system has had a working Windows partition in a year or two (well, in the case of the Haswell, it has NEVER seen Windows.)

Comment: Not entirely surprising (Score 4, Interesting) 116

by Andy Dodd (#47728877) Attached to: NSA Agents Leak Tor Bugs To Developers

The NSA has two directives that often conflict with each other:
1) Protect communications that are critical to our nation's security. This is mostly military/government comms, but they have a role in securing banking and other civilian networks. An example of what comes from this side of the NSA is SELinux - which is now heavily used by Android to provide additional security against malware.
2) Compromise and monitor the communications of our enemies. These guys overstepping their bounds are what has been routinely making the news lately.

While I can't see an obvious reason for the guys in category 1 to want to strengthen Tor, it's possible. (Potentially on behalf of another agency... Think in terms of Tor's use by Chinese dissidents.)

I'm fairly certain the people in categories 1 and 2 don't get along with each other. While in theory their goals should not conflict (one focuses on our enemies, one focuses on strengthening friendlies), the truth is that it's hard for the guys in category 1 to strengthen friends without also making those tools available to our enemies - and the guys in category 2 are routinely overstepping their bounds and attacking friendlies.

Comment: So? Old news. (Score 5, Interesting) 53

by Andy Dodd (#47721721) Attached to: Experimental Drug Stops Ebola-like Infection

Success in a test tube and/or monkeys doesn't mean much as far as hope for a drug viable for humans. After all, the trials for Tekmira's drug are on hold by the FDA due to safety concerns ( http://www.cbsnews.com/news/ho... ).

Also, Tekmira is NOT the company that manufactured the drug used to treat Dr. Brantly and his coworker - that was Mapp Pharmaceutical's ZMapp

Comment: Article tries to condemn nuclear, fails (Score 4, Insightful) 249

by Andy Dodd (#47697223) Attached to: The Cost of Caring For Elderly Nuclear Plants Expected To Rise

"Closing the older nuclear plants is not an option for many EU countries, which are facing an energy capacity crunch as other types of plant are being closed or mothballed because they can't cover their operating costs, or to meet stricter environmental regulation."

In short: While nuclear isn't perfect, it currently sucks less than any other alternative available.

(Renewables just aren't scalable enough yet.)

Comment: Re:Why dont we (Score 1) 194

by Andy Dodd (#47677067) Attached to: The Billion-Dollar Website

That's pretty much how government contracts work.

It fails because:
1) The customer will change their requirements mid-stream, screwing everything up
2) Even if they don't, in some cases it's discovered once everything is complete that the system which meets all of the customer's requirements is utterly fucking useless in the real world. I believe this was a major role in healthcare.gov's failures - many of its issues were discovered post-launch

Comment: Poor documentation (Score 1) 38

"It is highly recommended to use a router configuration we're not going to document or even provide you a link to".

The document implies that at least one modification is a flash and RAM upgrade - but they don't even provide links to details of this modification and/or whether any other techniques are needed (how do you populate the bootloader in the new flash? Or does the SoC itself have a built-in recovery mode?)

Comment: Re:not big in UK (Score 2) 120

by Andy Dodd (#47654741) Attached to: Gas Cooled Reactors Shut Down In UK

Part of the problem is that without government incentives/subsidies, companies will go for the highest-profit methods of power generation available.

Which means that the only plants built will be fossil fuel plants.

I don't believe that we currently have the technology to fully switch to renewable and won't for a few decades. Nuclear provides that bridge - Ideally after one more generation of nuclear reactors (modern designs are FAR safer than the existing ones) we'll have the storage technology to properly use renewables. In the worst case that renewables are STILL not ready, by then we will hopefully have some descendants of the IFR breeder reactor design coming online. Last time I saw a calculation, I believe the estimate was that IFR designs could have supplied the entire electrical needs of the USA for a century using only existing LWR waste stockpiles. (One of the big benefits of the IFR is that extracts a FAR higher percentage of the available energy from nuclear fuel, and the end waste products are only hot for 200 years or so.)

Comment: Re: It's a still a nice PC. (Score 1) 337

by Andy Dodd (#47647199) Attached to: Microsoft Surface Drowning?

For a while, tablets provided superior portability at a low cost to laptops. Laptops in the same cost range as tablets were either flimsy or bulky or just plain crap all around.

Google seems to be doing a good job of ensuring that their hardware partners do a good job with Chomebook build quality. As a result, they've created laptops that are cheap, highly portable, and reasonably durable. My 10" tablet has been relegated to "alarm clock".

Although the moon is smaller than the earth, it is farther away.

Working...