Feeds

back to article Intel promises '20X' power reduction with 'Haswell' chips

Intel claims that platforms built around its Haswell microarchitecture – the successor to the today's Sandy Bridge, scheduled for 2013 – will use one-twentieth the power of today's stingiest low-power platforms. "Haswell was designed to enable a 30 per cent reduction in connected standby power over the currently shipping …

COMMENTS

This topic is closed for new posts.
Linux

Ultrabooks

My X61s can do over one week on single charge when I Suspend To Ram it

doesn't make it useful during that time though

show me an ultraportable than can do over 24h on regular 9 cell battery and I'll become a believer until then, give me an ARM based laptop.

2
0

Sleep

Performance wise ARM still sucks compared compared to Intels offerings. Why would u want 24 hours on a single charge ? We all have to sleep, let your laptop charge while u sleep. Anyway it does sound like 12 hours plus will be possible with Haswell chips. Time will tell.

0
1

"Architecting" is still not a flipping word.

1
0

It will be, if it isn't already, by whatever measure you choose.

1
1

In the OED apparently

Quotes Keats using it in 1818

"“This was architected thus By the great Oceanus.”"

1
0
rav
Thumb Down

Yeah? But...

The graphics will still suck.

1
0

Standby time only an interesting metric in mobile phones

I assume this is nothing to do with ultrabooks, this is the result of taking the fight to mobiles.

There really isn't any reason to be "connected standby" rather then "real standby" unless you are expecting incoming calls.

My Android with a 1GHz Tegra2 is just fast enough to do everything I need (ITV player is slightly choppy, but the others are OK), so I don't see why having more power on tap will help.

This is like comparing compilers based on the size of "hello world", meaningless.

What would be interesting is to see how they compare with ARM on real workloads, not how they compare on full performance or no performance situations

0
0
Gold badge

Re: Standby time

"Standby time only an interesting metric in mobile phones"

Or tablets, which more resemble overgrown mobiles than slimmed down desktops.

Or any laptop that follows the rather faddish trend to "snooze" rather than turn off.

But yes, stand-by time *should* only be an interesting metric on phones.

1
0

>Or tablets, which more resemble overgrown mobiles than slimmed down desktops.

If you don't need to respond to data being pushed in then the extra 5 seconds it takes to attach to a router/base station is greatly outweighed by the extra hours you'd spend tethered to a charger - even if the chips are 20x as good.

Resemble a overgrown mobile they might, but always on just isn't sensible for pull data.

0
0
Thumb Down

How the hell can something be reduced by a factor of 20?

When something is increased by a factor of 20, it's larger by 20 times, or 2000 percent.

If something could be reduced by a factor of 20, it would be smaller by 20 times, or 2000 percent.

Yet once something is reduced by 100 percent, it no longer exists.

He meant to say that the new subsystem would draw 1/20th the power. In other words, its energy demand would be reduced by 95 percent.

0
0
Anonymous Coward

Re: How the hell can something be reduced by a factor of 20?

I'd divide by 20. Too simple?

4
0
Headmaster

20 times smaller is 5% of the original, or reduced by 95%.

If you can't do arithmetic, how do you know about the Register? Oh wait, "Maryland" is that your name or your location?

0
0
FAIL

20 times smaller???

The innumeracy is going from bad to worse.

To quote from 'factor change' at zonelandeducation.com:

'If a speed changed from 40 m/s to 20 m/s, we would say that the speed changed by a factor of 1/2.'

Need further evidence? From Common Errors in English Usage by Paul Brians:

'Mathematically literate folks object to expressions like “my paycheck is three times smaller than it used to be” because when used with whole numbers “times” indicates multiplication and should logically apply only to increases in size. Say "one third as large” instead.'

If Brians is not sufficiently authoritative, there's always the New York Times Style and Usage ('times less, times more'):

'"And do not write "times less" or "times smaller" (or things like "times as thin" or "times as short"). A quantity can decrease only one time before disappearing, and then there is nothing less to decrease further. Make it "one-third as much" (or as tall, or as fast).'

0
0
Anonymous Coward

You may not like the expression, but it is widely used and very clear: 20 times means a factor of 20 (note the use of multiplicative terms, not additive), more means multiply, less means divide. If you want to object to something, go look at all the journalists who, when a proportion changes from 20% to 21%, say that it increased by 1%. That at least is ambiguous.

0
0
Thumb Down

It's not about whether I like an expression

Very clear? I think not.

Under your definition, if I write 'reduced it by a factor of two-thirds,' readers will think I mean 'multiplied it by 3/2, making it 50 percent larger.' Do you seriously believe anyone will think that? Is that your intent?

Under your definition--but not under mine--to 'reduced unemployment by a factor of zero' would be impossible, since you can't divide by zero.

Under your definition, if you leave a value unchanged, you're reducing it by a factor of 1 (100 percent divided by 1). by your definition, reducing by a factor of 1 = increasing by a factor of 1 (100 percent x 1). That's jarringly counterintuitive.

Under your definition, if you reduce unemployment by 1 percent, you're

reducing it by a factor of 1.01 (since 1 divided by 1.01 = 0.99). Not at all obvious.

I completely agree with you that journalists who write 'increased by 1%' when they mean 'increased by 1 percentage point' are guilty of ambiguity and worse: stupidity or contempt for accuracy. Ditto for journalists who write '3 times greater' when they mean '2 times greater' or '3 times as great.'

0
0
Bronze badge

20 times what?

20 times is available everywhere if you just make it up as you go along.

Its all utter tosh without any TDP info.

0
0
Facepalm

Jan 0 Maths

"If you can't do arithmetic, how do you know about the Register?"

- perhaps in hindsight this is a retorical question Jan?

And does the 0 bit relate to your maths grade or just a general braincell count?

0
0
This topic is closed for new posts.