Obsolete before old: What’s the lifespan of your laptop?

[Chinese-Dragon]

This is old news. Google 'light bulb conspiracy'.

It is not even a conspiracy.

It is all out in the open and anyone who thinks about it for 5 minutes will understand it.

Companies want to make money, Governments earn tax from that money, the fat cats get rich. It's the same story throughout human history and nothing has changed.

They don't even have the decency to try and hide it from us.

 

[third eye]

It is not even a conspiracy.

It is all out in the open and anyone who thinks about it for 5 minutes will understand it.

Companies want to make money, Governments earn tax from that money, the fat cats get rich. It's the same story throughout human history and nothing has changed.

They don't even have the decency to try and hide it from us.

... and applies to all products across the board.

Not just phones.

 

[Vicar]

Experts on ancient Greek culture said that people back then didn't see their thoughts as belonging to them. When ancient Greeks had a thought, it occurred to them as a god or goddess giving an order. Apollo was telling them to be brave. Athena was telling them to fall in love. Now people hear a commercial for sour cream potato chips and rush out to buy, but now they call this free will. At least the ancient Greeks were being honest.

 

[Developereo]

It is not even a conspiracy.

It is all out in the open and anyone who thinks about it for 5 minutes will understand it.

Companies want to make money, Governments earn tax from that money, the fat cats get rich. It's the same story throughout human history and nothing has changed.

They don't even have the decency to try and hide it from us.

Two different concepts.

The light bulb conspiracy is about planned obsolescence. Light bulbs, and fridges and such, typically do not change drastically from year to year. It has been possible for decades to build light bulbs that can last far longer than they actually do.

That concept of planned obsolescence does not apply to electronics, especially computer hardware. As per Moore's Law, computer processing power doubles every 18 months. This is not a conspiracy by computer companies, this is technological innovation driven by competition. Your phone becomes obsolete, not because the manufacturers have dark intentions, but because technology keeps advancing. And, far from being superfluous, computer software makers keep demanding more processing power from the hardware so they can offer more functionality to the end user.

As an aside, the reason newer chips keep adding cores instead of making the CPU faster (GHz) has to do with the abysmal speed of main memory (RAM), which simply hasn't kept pace with processor speed. Upon cache miss, modern CPUs spend a lot of their time twiddling their thumbs waiting for main memory to deliver data, so the hardware designers decided to put multiple processors on a chip to provide parallelism.

 

[R0H1T]

Two different concepts.

The light bulb conspiracy is about planned obsolescence. Light bulbs, and fridges and such, typically do not change drastically from year to year. It has been possible for decades to build light bulbs that can last far longer than they actually do.

That concept of planned obsolescence does not apply to electronics, especially computer hardware. As per Moore's Law, computer processing power doubles every 18 months. This is not a conspiracy by computer companies, this is technological innovation driven by competition. Your phone becomes obsolete, not because the manufacturers have dark intentions, but because technology keeps advancing. And, far from being superfluous, computer software makers keep demanding more processing power from the hardware so they can offer more functionality to the end user.

As an aside, the reason newer chips keep adding cores instead of making the CPU faster (GHz) has to do with the abysmal speed of main memory (RAM), which simply hasn't kept pace with processor speed. Upon cache miss, modern CPUs spend a lot of their time twiddling their thumbs waiting for main memory to deliver data, so the hardware designers decided to put multiple processors on a chip to provide parallelism.

No just no, its the number of transistors on a chip that doubles not the computation power !
Infact even with Intel, the best in biz, they go for a process node(half a node for most others) shrink every ~2yrs or so !

The race for adding cores has long ended, just like the GHz race at the beginning of the last decade, simply because of the fact that adding more cores will just make the chip hotter thus companies nowadays go for improving efficiency & the underlying microarchitecture rather than adding cores or GHz(speed/frequency) to their chips !
I hate it when people with half baked knowledge put a topspin on stuff they're not fully aware of

 

[1ndy]

Actually everybody wants something better than others, i won't blame companies alone they are just doing their business. This is natural and won't change.

 

[Developereo]

No just no, its the number of transistors on a chip that doubles not the computation power !
Infact even with Intel, the best in biz, they go for a process node(half a node for most others) shrink every ~2yrs or so !

You don't seem to understand the connection between the two. Smaller components, placed closer together, result in reduced latencies, allowing the clock rate to be increased. Increased clock rate translates into more instructions per second, which is measured as computing power.

The race for adding cores has long ended, just like the GHz race at the beginning of the last decade, simply because of the fact that adding more cores will just make the chip hotter thus

The heat signatures and looming quantum effects are a problem no matter how the components are arranged on a chip. They are a problem both for faster cores and more cores. That is a separate problem in itself and is not the issue here.

My comment was specifically why hardware manufacturers are focusing on adding cores rather than increasing clock rates. The modern processor is plenty fast enough in real life situations, since the bottleneck is RAM speed. If RAM speeds were to increase significantly, then the balance might shift again. Until then, the path to superior performance lies with exploiting parallelism rather than making the uniprocessor faster,

Uniprocessors were left in the dust decades ago in the supercomputer race. All modern supercomputers are massively parallel. Granted, the parallelism is off-chip, but many of the same arguments for maximizing performance apply onto the chip as well.

companies nowadays go for improving efficiency & the underlying microarchitecture rather than adding cores or GHz(speed/frequency) to their chips !

Duh!

It is precisely this refinement of "microarchitecture" which favors using on-chip real estate to add cores rather than speed up the processor. Even with the evolving and improving cache-coherence schemes, pipelining and branch-prediction strategies, a cache miss is a disaster for a modern processor. And cache misses WILL happen. That is a fact of life and that's where the bottleneck of RAM speed kicks in. Unless RAM speeds catch up to processor speeds any time soon, which is extremely unlikely, the only alternative is to have other cores standing by doing useful work.

 

[Chinese-Dragon]

Actually everybody wants something better than others, i won't blame companies alone they are just doing their business. This is natural and won't change.

Exactly. It's not their fault that consumers are so willing to devour up everything regardless of whether they need it or not.

 

[R0H1T]

You don't seem to understand the connection between the two. Smaller components, placed closer together, result in reduced latencies, allowing the clock rate to be increased. Increased clock rate translates into more instructions per second, which is measured as computing power.



The heat signatures and looming quantum effects are a problem no matter how the components are arranged on a chip. They are a problem both for faster cores and more cores. That is a separate problem in itself and is not the issue here.

My comment was specifically why hardware manufacturers are focusing on adding cores rather than increasing clock rates. The modern processor is plenty fast enough in real life situations, since the bottleneck is RAM speed. If RAM speeds were to increase significantly, then the balance might shift again. Until then, the path to superior performance lies with exploiting parallelism rather than making the uniprocessor faster,

Uniprocessors were left in the dust decades ago in the supercomputer race. All modern supercomputers are massively parallel. Granted, the parallelism is off-chip, but many of the same arguments for maximizing performance apply onto the chip as well.



Duh!

It is precisely this refinement of "microarchitecture" which favors using on-chip real estate to add cores rather than speed up the processor. Even with the evolving and improving cache-coherence schemes, pipelining and branch-prediction strategies, a cache miss is a disaster for a modern processor. And cache misses WILL happen. That is a fact of life and that's where the bottleneck of RAM speed kicks in. Unless RAM speeds catch up to processor speeds any time soon, which is extremely unlikely, the only alternative is to have other cores standing by doing useful work.

I hate to break it to you but here goes nothing ~

1) Increasing the number of transistors doesn't necessarily double or treble the performance of that chip nor does it increase the frequency of the said chip, CPU performance/frequency is at best loosely dependent on transistor count ! Latencies at interchip level are so miniscule that a move from 32nm to 22nm wouldn't do much to the performance bottleneck, infact its the other components like RAM, HDD, GPU, mobo & their respective interconnects that reduce the potential performance of a chip hence the move towards SoC !

2) I dunno how we got into the quantum mechanics part here but hey as you said its irrelevant in this case so I'll leave it at that !

3) By microarch I meant refining the way a chip works its akin to tuning an engine that drives a vehicle which is quite different to shrinking the process node, based on which the chip is fabricated, & its the latter which you were talking about incase of Moore's law ! The part of cache enhancement & improving the rest is where the microarch refinement kicks in, lastly we have DDR4 RAM in production now that will be faster than the previous gen DDR3 however the clock rates will never match that of a CPU besides we won't have DDR4 using processors anytime before 2015 at the earliest so any complaints in this regard are moot !

 

[Developereo]

1) Increasing the number of transistors doesn't necessarily double or treble the performance of that chip nor does it increase the frequency of the said chip, CPU performance/frequency is at best loosely dependent on transistor count ! Latencies at interchip level are so miniscule that a move from 32nm to 22nm wouldn't do much to the performance bottleneck, infact its the other components like RAM, HDD, GPU, mobo & their respective interconnects that reduce the potential performance of a chip hence the move towards SoC !

You STILL don't get it how component size has a direct effect on processor speed and I won't waste time explaining further. I never claimed that the resulting impact is linear; that's just a furphy you threw in just now to save face.

I am enjoying how you are bringing in RAM, HDD, GPU into this particular point of Moore's law and CPU speed. These factors have NO effect on a particular CPU's processing speed. Read again: NONE. The CPU runs at a particular speed -- these other off-chip factors put the CPU in a wait state, which has zilch to do with Moore's law.

2) I dunno how we got into the quantum mechanics part here but hey as you said its irrelevant in this case so I'll leave it at that !

No, I don't expect you to know about the concerns for quantum effects in modern chip design.

3) By microarch I meant refining the way a chip works its akin to tuning an engine that drives a vehicle which is quite different to shrinking the process node, based on which the chip is fabricated, & its the latter which you were talking about incase of Moore's law ! The part of cache enhancement & improving the rest is where the microarch refinement kicks in, lastly we have DDR4 RAM in production now that will be faster than the previous gen DDR3 however the clock rates will never match that of a CPU besides we won't have DDR4 using processors anytime before 2015 at the earliest so any complaints in this regard are moot !

You are just paraphrasing exactly what I wrote in a predictable backtrack since your bombastic post was full of hot air and no substance.

 

[R0H1T]

You STILL don't get it how component size has a direct effect on processor speed and I won't waste time explaining further. I never claimed that the resulting impact is linear; that's just a furphy you threw in just now to save face.



No, I don't expect you to know about the concerns for quantum effects in modern chip design.



You are just paraphrasing exactly what I wrote in a predictable backtrack since your bombastic post was full of hot air and no substance.

Still denying that you're mixing advanced physics with CPU's & what not aye, typical & for the last time~

1) Ever heard of overclocking ? You can easily get twice the performance from the same die & the same component(transistors) size by increasing the frequency at which they work so would you care to explain why component size matters ? The smaller process nodes actually helps in processor efficiency as in for instance ~

A proc running at 2Ghz fabricated on 22nm node will be roughly ~30% cooler or can be clocked that much higher as compared to the same proc on 32nm node, all things being equal viz microarch, transistor count et al !

I don't claim to know more about quantum mechanics or tunneling effect at super high frequencies but don't assume whatever you're posting(wrt computing performance) is the gospel of truth here !

 

[Developereo]

Still denying that you're mixing advanced physics with CPU's & what not aye, typical & for the last time~

1) Ever heard of overclocking ? You can easily get twice the performance from the same die & the same component(transistors) size by increasing the frequency at which they work so would you care to explain why component size matters ? The smaller process nodes actually helps in processor efficiency as in for instance ~

A proc running at 2Ghz fabricated on 22nm node will be roughly ~30% cooler or can be clocked that much higher as compared to the same proc on 32nm node, all things being equal viz microarch, transistor count et al !

This is getting FUN!

The advertized clock rate for a CPU is the recommended rate for optimal, long life performance. Overclocking will squeeze more performance, but only at the expense of processor life. The chip is designed with certain performance parameters in mind, which are dictated by the component technology. You cannot overclock a CPU beyond a certain point which is not supported by the underlying chip components.

Are you going to tell me that you will take a 1980s chip and "overclock" it to run at 2GHz?

I don't claim to know more about quantum mechanics or tunneling effect at super high frequencies but don't assume whatever you're posting(wrt computing performance) is the gospel of truth here !

You managed to google tunneling effect.

Good for you, although you need more googling to understand what it actually means.

 

[R0H1T]

This is getting FUN!

The advertized clock rate for a CPU is the recommended rate for optimal, long life performance. Overclocking will squeeze more performance, but only at the expense of processor life. The chip is designed with certain performance parameters in mind, which are dictated by the component technology. You cannot overclock a CPU beyond a certain point which is not supported by the underlying chip components.



You managed to google tunneling effect.

Good for you, although you need more googling to understand what it actually means.

Alright wise guy don't start acting like a t!ck now !
Firstly there is no Moore's law unlike E=mc^2, it was just an estimate put forth by the Intel pioneer at the beginning of that firm some ~40yrs ago, google it !
"Overclocking" ~ you really are clueless aren't ya, search any number of tech forums/sites & then comeback to me on that FYI I'm running my 2008 Wolfdale CPU at 50% overclock on stock cooler, you can easily get over twice the performance with water cooling & nearly four times on LN2, so you only need to supply more power & have better cooling with the latter being more important !

I first read about tunneling effect some 10yrs ago in 9th or 10th grade so please spare me the googling part

 

[Developereo]

Alright wise guy don't start acting like a t!ck now !
Firstly there is no Moore's law unlike E=mc^2, it was just an estimate put forth by the Intel pioneer at the beginning of that firm some ~40yrs ago, google it !
"Overclocking" ~ you really are clueless aren't ya, search any number of tech forums/sites & then comeback to me on that FYI I'm running my 2008 Wolfdale CPU at 50% overclock on stock cooler, you can easily get over twice the performance with water cooling & nearly four times on LN2, so you only need to supply more power & have better cooling with the latter being more important !

I notice you evaded my point that you can only overclock up to a certain point beyond the recommended value. Ultimately, the maximum speed is dictated by the chip's component technology.

 

[R0H1T]

I notice you evaded my point that you can only overclock up to a certain point beyond the recommended value. Ultimately, the maximum speed is dictated by the chip's component technology.

Nope infact the earlier celerons at 65nm could exceed ~8GHz on LN2 but nowadays IVB won't cross ~7Ghz on 22nm with the same cooling ! Actually the lesser the number of transistors the higher they can be clocked(more raw perf) know why cause more transistors means greater heat so less is the frequency achieved hence the need for multicore proc to be clocked lower vs single core models of yester years ! Infact placement of transistors on a chip is even more imp to prevent hotspot formation that reduce the overall perf of the chip !
You are partly right with the process node part but microach refining plays a greater role here for instance the latest IVB is at best ~20% faster than the previous gen SNB with the same microarch coming from 32nm fab to 22nm trigate