Most effective cpu ?

[tom vilsack]

Probably. I spent a few hundred on a couple bags for it so yeah. Might consider getting the 1.3 once I see benchmarks then Id sell this one

Probably Apple's to Oranges (no pun intended) but looking at other laptops running the 5Y71-M...There Geekbench scores are not all that great. I'm wondering just how much of a speed increase we will see?

GeekBench 5Y71-M

 

[dexterbell]

Probably Apple's to Oranges (no pun intended) but looking at other laptops running the 5Y71-M...There Geekbench scores are not all that great. I'm wondering just how much of a speed increase we will see?

GeekBench 5Y71-M

Those are well under the 5Y51 and most are even lower than the 5Y31. Just shows how well Apple optimized OS X to run with these chips. It also shows how wildly inconsistent these benchmarks can be. One Dell Venue Pro 11 got a multi score of 3853 while another got 4303. Thats a pretty big difference.

 

[tom vilsack]

dexter:

If your looking for a good read about the intel-m...this link reviewing the new Asus Zenbook UX305 is a good read...(performance section is helpful in explaining thing's)

Asus Zenbook UX305

 

[leman]

What? They produce them all the same, yet there's a variance that just seems to line up with a 1.2, 1.3, and 1.4 gHz clock speed, and they separate them by QC testing? That I don't buy.

That is exactly how it works. I know it sounds contra-intruitive, but the crucial realisation is that every CPU — even if its made by the same machine using the same process — is slightly different. There are always 'individual' impurities and micro-variations,which affect energy efficiency. Some chips can clock up higher without using that much additional energy, and some will start getting unstable very soon. This is something that any overclocker knows very well.

So yeah, after the chips have been manufactured, Intel does extensive testing and sells some of the 'purer' ones as higher-rate models. More specifically AFAIK, the Core M is a basic Broadwell i7 dual-core with half of the GPU disabled.

----------

As developers scale code to higher core systems, they lose efficiency per core. TSX was designed to fix this through fine grain locking.

I doubt that TSX will benefit normal everyday computing that much, its primarily relevant for high-concurrency database servers. But its a definitively nice feature to have!

I would be also very quires to know where the difference between 31/51 and 71 is coming from... maybe 31/51 just has the relevant bits disabled?

 

[mutsaers-vr.nl]

But looking at these cpu values the spread is so big that you even do not need 6 sigma to see that INTEL does not have any controli over the processing !!!! Maybe they use the ASML crap machines to produce them.

That is exactly how it works. I know it sounds contra-intruitive, but the crucial realisation is that every CPU — even if its made by the same machine using the same process — is slightly different. There are always 'individual' impurities and micro-variations,which affect energy efficiency. Some chips can clock up higher without using that much additional energy, and some will start getting unstable very soon. This is something that any overclocker knows very well.

So yeah, after the chips have been manufactured, Intel does extensive testing and sells some of the 'purer' ones as higher-rate models. More specifically AFAIK, the Core M is a basic Broadwell i7 dual-core with half of the GPU disabled.

----------



I doubt that TSX will benefit normal everyday computing that much, its primarily relevant for high-concurrency database servers. But its a definitively nice feature to have!

I would be also very quires to know where the difference between 31/51 and 71 is coming from... maybe 31/51 just has the relevant bits disabled?

 

[shenan1982]

That is exactly how it works. I know it sounds contra-intruitive, but the crucial realisation is that every CPU — even if its made by the same machine using the same process — is slightly different. There are always 'individual' impurities and micro-variations,which affect energy efficiency. Some chips can clock up higher without using that much additional energy, and some will start getting unstable very soon. This is something that any overclocker knows very well.

Now I feel kinda stupid for paying Apple $200 for any processor spec bumps I've had in their laptops over the years, hehe

 

[iRun26.2]

Now I feel kinda stupid for paying Apple $200 for any processor spec bumps I've had in their laptops over the years, hehe

You shouldn't. You've been getting the better part. Essentially you are paying Intel for the extra work they had to do to sort (test) the chips. I'm sure those testing procedures are very complicated.

 

[Gav2k]

Which cpu saves more energy under normal conditions (e.g. text editing or web browsing) and produces less heat?

Core M-5Y31 @ 900 MHz - overclocked to 1.1 GHz
Core M-5Y51 @ 1.1 Ghz - overclocked to 1.2 GHz
Core M-5Y71 @ 1.2 GHz - overclocked to 1.3 GHz

Which one is prone to throttle under heavy load?

None of them are 'overclocked' so we shall clear that up to start with.

The 1.1 runs at the maximum base frequency specified by Intel @6w

The 1.2 and 1.3 both run below their maximum base frequency and thus require less power 5-5.5w.

What you need to factor in is that although the chips have 1.1/1.2 and 1.3 base frequencies they can all underclock themselves to conserve power when not needed.

The 1.2 and 1.3 make the logical choice due to the reducion in wattage required to perform at its base frequencies.

 

[low performer]

None of them are 'overclocked' so we shall clear that up to start with.

The 1.1 runs at the maximum base frequency specified by Intel @6w

The 1.2 and 1.3 both run below their maximum base frequency and thus require less power 5-5.5w.

What you need to factor in is that although the chips have 1.1/1.2 and 1.3 base frequencies they can all underclock themselves to conserve power when not needed.

The 1.2 and 1.3 make the logical choice due to the reducion in wattage required to perform at its base frequencies.

5Y31
Processor Base Frequency 900 MHz
TDP 4.5 W
Scenario Design Power (SDP) 3.5 W

Configurable TDP-up Frequency 1.1 GHz
Configurable TDP-up 6 W

Configurable TDP-down Frequency 600 MHz
Configurable TDP-down 3.5 W


5Y51
Processor Base Frequency 1.1 GHz
TDP 4.5 W
Scenario Design Power (SDP) 3.5 W

Configurable TDP-up Frequency 1.3 GHz
Configurable TDP-up 6 W

Configurable TDP-down Frequency 600 MHz
Configurable TDP-down 3.5 W


5Y71
Processor Base Frequency 1.2 GHz
TDP 4.5 W
Scenario Design Power (SDP) 3.5 W

Configurable TDP-up Frequency 1.4 GHz
Configurable TDP-up 6 W

Configurable TDP-down Frequency 600 MHz
Configurable TDP-down 3.5 W

(Source: intel)

One more question: A higher base frequency does not mean that the power consumption is higher while the processor is idle?

 

[Gav2k]

5Y31
Processor Base Frequency 900 MHz
TDP 4.5 W
Scenario Design Power (SDP) 3.5 W

Configurable TDP-up Frequency 1.1 GHz
Configurable TDP-up 6 W

Configurable TDP-down Frequency 600 MHz
Configurable TDP-down 3.5 W


5Y51
Processor Base Frequency 1.1 GHz
TDP 4.5 W
Scenario Design Power (SDP) 3.5 W

Configurable TDP-up Frequency 1.3 GHz
Configurable TDP-up 6 W

Configurable TDP-down Frequency 600 MHz
Configurable TDP-down 3.5 W


5Y71
Processor Base Frequency 1.2 GHz
TDP 4.5 W
Scenario Design Power (SDP) 3.5 W

Configurable TDP-up Frequency 1.4 GHz
Configurable TDP-up 6 W

Configurable TDP-down Frequency 600 MHz
Configurable TDP-down 3.5 W

(Source: intel)

One more question: A higher base frequency does not mean that the power consumption is higher while the processor is idle?

It depends on the chip used. All will clock down during idle periods

 

[leman]

But looking at these cpu values the spread is so big that you even do not need 6 sigma to see that INTEL does not have any controli over the processing !!!! Maybe they use the ASML crap machines to produce them.

Actually, most of it is marketing. Its like what GPU makers do. They know that their chip can stably run at 1Ghz. So they sell it clocked at 800Mz. Next year, they take the same chip, and sell it as a 'new' GPU clocked at 1Ghz.

The point being here, Intel could have probably clocked these CPUs higher (although its questionable if they can do it without losing power-efficiency). But it does not make any sense to do them from market standpoint. Either way, they are the best performer in class.

 

[dexterbell]

So the million dollar question....is the 1.2 worth the extra $300? Will there be any noticeable differences? Seems the only real world difference would be less heat generated by the 1.2

 

[tecnho]

Most effective cpu ?

So the million dollar question....is the 1.2 worth the extra $300? Will there be any noticeable differences? Seems the only real world difference would be less heat generated by the 1.2

There is more to this question, because those of us who are not interested in larger storage need to decide if going from 1.1 to 1.3 will be worth 250$. Alas, no 1.3 are available for testing yet.

 

[enduser99]

Actually Intel sells all the chips in the macbook for the same price (at least that's how Intel lists them, I don't know what price Apple is getting). It's Apple that socks it to us by charging a very high markup.

Thats complete BS. As someone who has an NDA with Intel and knows ODM pricing for these 3 specific chips, I can tell you for certain that the published price is not what Apple pays, and the 3 chips do indeed have different prices.

 

[squirrrl]

Thats complete BS. As someone who has an NDA with Intel and knows ODM pricing for these 3 specific chips, I can tell you for certain that the published price is not what Apple pays, and the 3 chips do indeed have different prices.

I'm sure that's not what they pay, but if they are offering the same price for each chip, retail, then I'm pretty sure that Intel isn't charging steep differences for each Core M with higher clock speeds. Which was my point.... that Apple is marking up the chips significantly, when I doubt they pay much difference between all of them. I understand if they are doing some extra engineering with the 51 and 71 versions, maybe that's what they are charging extra for? or maybe they just want the extra dough.

 

[enduser99]

....but if they are offering the same price for each chip, retail, then I'm pretty sure that Intel isn't charging steep differences for each Core M with higher clock speeds. Which was my point.... that Apple is marking up the chips significantly, when I doubt they pay much difference between all of them. I understand if they are doing some extra engineering with the 51 and 71 versions, maybe that's what they are charging extra for? or maybe they just want the extra dough.

Chips are priced according to what the market will bare, there is no relation to "extra work" for the higher clock CPUs, they (the 3 Y-series Broadwell Core-M's) are all made and tested the same. You are right, Apple definitely adds profit margin on top of the cost difference from Intel. Why shouldn't they?

 

[iRun26.2]

I'm sure that's not what they pay, but if they are offering the same price for each chip, retail, then I'm pretty sure that Intel isn't charging steep differences for each Core M with higher clock speeds. Which was my point.... that Apple is marking up the chips significantly, when I doubt they pay much difference between all of them. I understand if they are doing some extra engineering with the 51 and 71 versions, maybe that's what they are charging extra for? or maybe they just want the extra dough.

Higher margins are always a good company goal.

 

[squirrrl]

Chips are priced according to what the market will bare, there is no relation to "extra work" for the higher clock CPUs, they (the 3 Y-series Broadwell Core-M's) are all made and tested the same. You are right, Apple definitely adds profit margin on top of the cost difference from Intel. Why shouldn't they?

I'm just saying that clearly it is likely not supply/demand on Intel's side. The real equation seems to be cost = to what Apple thinks customers' perceived benefit for the upgrade is.

If it were simply a supply issue, Intel would list the price of the 71 higher than any other core M chips and that is simply not the case, at least according to their price list.

http://www.intc.com/priceList.cfm

We are all assuming that Apple is getting a price break, but even if they aren't, at $281 for the retail cost of ANY Core M, Apple is essentially passing off the entire cost of the chip in a 1.1 to 1.3 upgrade.

 

[enduser99]

... Intel would list the price of the 71 higher than any other core M chips and that is simply not the case, at least according to their price list.

http://www.intc.com/priceList.cfm

We are all assuming that Apple is getting a price break, but even if they aren't, at $281 for the retail cost of ANY Core M, Apple is essentially passing off the entire cost of the chip in a 1.1 to 1.3 upgrade.

No, that website is complete bunk related to the pricing. Intel doesn't sell these Core-M CPUs to consumers, the pricing there is just a placeholder, and the relative prices are irrelevant. 5Y71 costs more.

 

[squirrrl]

No, that website is complete bunk related to the pricing. Intel doesn't sell these Core-M CPUs to consumers, the pricing there is just a placeholder, and the relative prices are irrelevant. 5Y71 costs more.

That website is from Intel's own website dude... It's intel's ACTUAL website.

Their site. That they run. for their products. They have both intel.com and intl.com The latter being their stock name so it's their corporate (not consumer) site.

Pricing for these things is always a per contract basis but my point, which you don't address, is that the markup from the 1.1 to the 1.3 just doesn't makes sense according to difference in the price point between the two chips themselves. I'm not complaining, I know that Apple marks BTO things much higher than they otherwise would be, but I'm saying it's not a supply/demand issue from Intel's side, it Apple pricing up to what they think people will pay.

I'd like to see a comparative teardown of both the 1.1 and the 1.3 to see if there is any differences in the way they engineered their graphene heat sink as well to see if there are any physical changes other than the chip itself.

 

[enduser99]

That website is from Intel's own website dude... It's intel's ACTUAL website.

Their site. That they run. for their products. They have both intel.com and intl.com The latter being their stock name so it's their corporate (not consumer) site.

I know, and yet it is true. Like I said, I am in the computer business. I know ODM pricing for each of these chips. Companies publish BS, placeholders, errors all the time. Intel has no interest in advertising real pricing to the public for these parts.