Intel Broadwell Thread

[monstercameron]

That's a pretty bad 3DMark Physics score...Moorefield in a smartphone does better.

do you have a link to that? the only icestorm score i see is for the z3560 and its around 17,000. Mind you that was in a tablet formfactor and a phone would score much lower.

 

[bullzz]

@monstercameron - ya somethings wrong abt that 3DMark score. I expect at least Denver level performance from GPU

also its a little unfair the compare the 2 laptops you posted. one is fanless and the other is 15 inch laptop with fan

 

[monstercameron]

@monstercameron - ya somethings wrong abt that 3DMark score. I expect at least Denver level performance from GPU

also its a little unfair the compare the 2 laptops you posted. one is fanless and the other is 15 inch laptop with fan

and if the other was 13'' with an ips display and flash based storage, im confident it would be more competitive in terns of power draw.

 

[Abwx]

also its a little unfair the compare the 2 laptops you posted. one is fanless and the other is 15 inch laptop with fan

If allowed to reach 43°C like the Envy it could be fanless, no doubt...

http://www.notebookcheck.com/Test-HP-Envy-x2-j001ng-Convertible.129799.0.html

http://www.notebookcheck.net/HP-15-g005ng-Notebook-Review.126027.0.html

and if the other was 13'' with an ips display and flash based storage, im confident it would be more competitive in terns of power draw.

Once you ponderate the set ups differences there will be about..no differences in matter of experience and power comsumption, probably that you could even squeeze Beema, like you point it, to lower power drain by reducing frequency by 10%, it would be still enough when looking at the CB 11.5 and other scores...

 

[bullzz]

@Abwx - those 2 laptops dont compete in the same segment/price/performance. which is why its even more unfair

@monstercameron - true but the beema laptops also have 20% more battery to compensate for additional power draw

 

[monstercameron]

@Abwx - those 2 laptops dont compete in the same segment/price/performance. which is why its even more unfair

@monstercameron - true but the beema laptops also have 20% more battery to compensate for additional power draw

check again, these oems keep sabotaging amd and saddle them with small 30whr batteries

 

[dahorns]

If allowed to reach 43°C like the Envy it could be fanless, no doubt...

That's a bit of a leap. First, the Beema setup is much larger, and thus dissipates more heat. Second, you'll notice the Beema setup has 12 zones above 30 Celsius. The Envy x2 has only 7.

Once you ponderate the set ups differences there will be about..no differences in matter of experience and power comsumption, probably that you could even squeeze Beema, like you point it, to lower power drain by reducing frequency by 10%, it would be still enough when looking at the CB 11.5 and other scores...

The Beema has a 30% larger battery, lower resolution screen, twice the thickness, 20% more weight, half the single threaded performance (.6 vs 1.06 CB 11.5), and 10% less battery life while surfing.

Oh, not to mention the Envy comes ahead in the only gaming benchmark (i.e., non-synthetic) listed for both:

Beema vs Core M
Tomb Raider (2013)
low - 34 vs 40.8 (+20%)
med - 17.8 vs 20 (+12%)
high - 11.9 vs 14.7 (+23%)

 

[dahorns]

benchmark-------| beema | broadwell
3dmark physics--| 19842 | 20946
3dmark graphics-| 35540 | 34908
power max-------| 24.8W | 18.2
l

Actually, going back through the review I don't see those numbers anywhere. Where did you get them? I don't see anything that even looks like a physics score. Here is what I see:

(Core M vs Beema)
3DMark 11--------: 731 vs 753
3DMark Icestorm-: 30,404 vs 30,225
3DMark Cloud-----: 3,580 vs 2,747
3dMark Fire strike: 387 vs 442

 

[Abwx]

Actually, going back through the review I don't see those numbers anywhere. Where did you get them? I don't see anything that even looks like a physics score. Here is what I see:

(Core M vs Beema)
3DMark 11--------: 731 vs 753
3DMark Icestorm-: 30,404 vs 30,225
3DMark Cloud-----: 3,580 vs 2,747
3dMark Fire strike: 387 vs 442

Personaly i saw this :

(Core M vs Beema)

3DMark 11--------: 731 vs 831
3DMark Icestorm-: 30,404 vs 32088
3DMark Cloud-----: 3,580 vs 2882
3dMark Fire strike: 387 vs 517
CB 11.5 ST : 1.06 vs 0.6
CB 11.5 MT : 1.89 vs 2.01
Price : 1000€ vs 400€

Also in games the 6410 does better in F1 2013 than the Envy in F1 2014, granted it s not the same versions but there wasnt only Tomb raider, also the Envy is better than most other such devices so it s a best case scenario, i spare everybody numbers taken from the Yoga 3..


http://www.notebookcheck.com/Test-HP-Envy-x2-j001ng-Convertible.129799.0.html

http://www.notebookcheck.net/HP-15-g005ng-Notebook-Review.126027.0.html

I m sure a 6410 is not that more expensive than a 6310 and should be, substancialy, cheaper than a Core M whose production cost mandate at least 100$ selling price if they want to retain their usual margins on these kind of product.

check again, these oems keep sabotaging amd and saddle them with small 30whr batteries

The 15"6 HP i linked has a 41Wh battery, that s correct but still short of the usual 48Wh we find in these formats.

 

[dahorns]

Personaly i saw this :

(Core M vs Beema)

3DMark 11--------: 731 vs 831
3DMark Icestorm-: 30,404 vs 32088
3DMark Cloud-----: 3,580 vs 2882
3dMark Fire strike: 387 vs 517
CB 11.5 ST : 1.06 vs 0.6
CB 11.5 MT : 1.89 vs 2.01
Price : 1000€ vs 400€

Just so it is clear, you're citing from a different Beema device. I was posting the scores from the device that monstercamersons linked, the Lenovo B50-45.

My point was really that the physics subscore was not present. I don't think notebookcheck tracks those subscores.

Also in games the 6410 does better in F1 2013 than the Envy in F1 2014, granted it s not the same versions but there wasnt only Tomb raider, also the Envy is better than most other such devices so it s a best case scenario, i spare everybody numbers taken from the Yoga 3..

F1 2013 and F1 2014 are different games. That is why I only posted Tomb Raider.

I'm also not sure how you can state that the Envy is "better than most other such devices." I mean, there is quite literally only one other Core M device that has been reviewed.

 

[Abwx]

Just so it is clear, you're citing from a different Beema device. I was posting the scores from the device that monstercamersons linked, the Lenovo B50-45.

My point was really that the physics subscore was not present. I don't think notebookcheck tracks those subscores.


F1 2013 and F1 2014 are different games. That is why I only posted Tomb Raider.

I'm also not sure how you can state that the Envy is "better than most other such devices." I mean, there is quite literally only one other Core M device that has been reviewed.

Fair point about the subscores, there s some submissions at Futuremark site.

http://www.3dmark.com/3dm/3266166

MC posted a link with a Beema 1.8-2.4 6310 while i posted one with a 2.0-2.4 6410 with also better battery.

As for the Envy it has better perfs than the Yoga, you could get even better ones but according to the review the chip used 7.444W at peak, max comsumption is 18.2W, to get better perfs you ll have to raise the Watt bar significantly higher and this would eat in the autonomy that is already nothing exceptional.

 

[dahorns]

Fair point about the subscores, there s some submissions at Futuremark site.

http://www.3dmark.com/3dm/3266166

MC posted a link with a Beema 1.8-2.4 6310 while i posted one with a 2.0-2.4 6410 with also better battery.

As for the Envy it has better perfs than the Yoga, you could get even better ones but according to the review the chip used 7.444W at peak, max comsumption is 18.2W, to get better perfs you ll have to raise the Watt bar significantly higher and this would eat in the autonomy that is already nothing exceptional.

The power charts are interesting, but I'm not sure how much they really tell us about the 5y70. First, they measure the whole device. Second, the results are . . . interesting. For instance, one of the comparative devices in the Envy's review is the HP Spectre 13 (4202y Haswell).

Battery Size
Spectre 13-: 50 wh
Envy x2----: 33 wh

Power Consumption (Average load, Stress Test load):
Spectre 13-: 16, 14.5
Envy x2----: 15.3, 18.2

Battery Life (Surfing with Lan, Battery Eater):
Spectre 13-: 7h 25m, 2h 45m
Envy x2----: 5h 23m, 2h 41m

 

[Abwx]

The power charts are interesting, but I'm not sure how much they really tell us about the 5y70. First, they measure the whole device. Second, the results are . . . interesting. For instance, one of the comparative devices in the Envy's review is the HP Spectre 13 (4202y Haswell).

Battery Size
Spectre 13-: 50 wh
Envy x2----: 33 wh

Power Consumption (Average load, Stress Test load):
Spectre 13-: 16, 14.5
Envy x2----: 15.3, 18.2

Battery Life (Surfing with Lan, Battery Eater):
Spectre 13-: 7h 25m, 2h 45m
Envy x2----: 5h 23m, 2h 41m

In the Lan surfing test the HW has about the same power comsumption, normalizing the battery capacity the improvement is 10%, and that do not include the set ups differences.

On the battery eater test there should be a score to ponderate the power comsumption, comsumption without the perfs is not very insightfull.

Globaly this chip has better GPU perfs than a BTrail but is not significantly better CPU wise, compared to a Mullins/Beema it does hold even less water if we are to compare the perf/watt/price ratios, the suggested prices of items is just plain rip off with features like SSD and HD screens used as some make up to create value upon a quite average product.

 

[dahorns]

Globaly this chip has better GPU perfs than a BTrail but is not significantly better CPU wise[.]

Do you have anything to back this up? In most tests, particularly single-threaded tests, 5y70 is much better than bay trail. Here, check it out. Most of those tests were done with the Yoga 3. It would be interesting to see how the Envy x2 performs.

http://www.notebookcheck.com/Intel-Core-M-5Y70-SoC.125624.0.html

http://www.notebookcheck.net/Intel-Atom-Z3770-Tablet-SoC.101424.0.html

FYI, "ponderate" is a very odd word choice. I think you're using it correctly. But I haven't seen it used in modern English. I doubt many English speakers would understand you if you said it.

*Not a criticism* just hopefully helpful language advice.

 

[Abwx]

Do you have anything to back this up? In most tests, particularly single-threaded tests, 5y70 is much better than bay trail. Here, check it out. Most of those tests were done with the Yoga 3. It would be interesting to see how the Envy x2 performs.

http://www.notebookcheck.com/Intel-Core-M-5Y70-SoC.125624.0.html

http://www.notebookcheck.net/Intel-Atom-Z3770-Tablet-SoC.101424.0.html

FYI, "ponderate" is a very odd word choice. I think you're using it correctly. But I haven't seen it used in modern English. I doubt many English speakers would understand you if you said it.

*Not a criticism* just hopefully helpful language advice.

Thank for the precisions, otherwise there s a N3530 in the Beema HP test i linked, the Baytrail CB 11.5 MT score is 1.7, to compare with the Envy 1.89, the ST scores do not matter, most apps use several cores and thoses wich are still ST are such because they do not need more grunt to be fast enough.

 

[monstercameron]

Do you have anything to back this up? In most tests, particularly single-threaded tests, 5y70 is much better than bay trail. Here, check it out. Most of those tests were done with the Yoga 3. It would be interesting to see how the Envy x2 performs.

http://www.notebookcheck.com/Intel-Core-M-5Y70-SoC.125624.0.html

http://www.notebookcheck.net/Intel-Atom-Z3770-Tablet-SoC.101424.0.html

FYI, "ponderate" is a very odd word choice. I think you're using it correctly. But I haven't seen it used in modern English. I doubt many English speakers would understand you if you said it.

*Not a criticism* just hopefully helpful language advice.

I dont think abwx is a native english speaker.

The single threaded perf is good, but it doesnt clock very high [short 2.6GHz turbo] maybe 14nm isn't that great or just badly optimized systems build around it.

The gpu is alright matches 128 GCN 1.0 @ 800MHz

what is so special about broadwell?

 

[Enigmoid]

Thank for the precisions, otherwise there s a N3530 in the Beema HP test i linked, the Baytrail CB 11.5 MT score is 1.7, to compare with the Envy 1.89, the ST scores do not matter, most apps use several cores and thoses wich are still ST are such because they do not need more grunt to be fast enough.

Very rarely will you be running something that scales well to 4 threads on these types of machine. Twice the ST performance is very good and in day to day usage (generally around 2 to maybe 3 threads) the BW machine will be far better. Lots of apps still hit a single core hard. I can easily tell the difference between IVB at 3.2 Ghz and at 1.2 Ghz on flash heavy sites despite having 4 physical cores.

 

[dahorns]

I dont think abwx is a native english speaker.

I assumed he wasn't. I appreciate it when others correct my usage in other languages, so I thought I'd extend the same courtesy.

 

[Abwx]

Very rarely will you be running something that scales well to 4 threads on these types of machine. Twice the ST performance is very good and in day to day usage (generally around 2 to maybe 3 threads) the BW machine will be far better.

You cant have your cake and eat it, at two threads it will throttle the frequency , the ST score is valid only with a single core active and the other idling, you wont have twice the ST perf when there s 2 threads, let alone 3.

what is so special about broadwell?

Nothing in this power range, it s typicaly a 12-15W chip that had been squeezed into power devices wich it s actualy not meant for, or rather there s some technical limitations that retain it from performing as well, in respect of the competitions, as its more power hungry siblings, or simply compared to itself but at say 15W.

First limitation is that the CPU part is over the point at wich IPC delta/power delta is still greater or equal to 1 in the current state of the uarches, it is below 1 since SB at least in Intel uarch, so lower IPC designs like BT or Mullins/Beema have no trouble being on par perf/watt wise despite being one and two nodes late respectively, at equal nodes it would be a slaughter.

Second limitation is the poor perf/watt of Intel s GPU, this is an area that is not compensated by the 14-16nm node.

Lastly, their current process does not seem to be up to their own expectations, if any indication the current chips are surely cherry picked and this is telling about the dispersions they must have in their process, with yields surely in the low double digit, something like 20-25%.

 

[monstercameron]

You cant have your cake and eat it, at two threads it will throttle the frequency , the ST score is valid only with a single core active and the other idling, you wont have twice the ST perf when there s 2 threads, let alone 3.

good point, it seems the TDP limits have severely caused the devices to throttle.

 

[Phynaz]

You cant have your cake and eat it, at two threads it will throttle the frequency , the ST score is valid only with a single core active and the other idling, you wont have twice the ST perf when there s 2 threads, let alone 3.



Nothing in this power range, it s typicaly a 12-15W chip that had been squeezed into power devices wich it s actualy not meant for, or rather there s some technical limitations that retain it from performing as well, in respect of the competitions, as its more power hungry siblings, or simply compared to itself but at say 15W.

First limitation is that the CPU part is over the point at wich IPC delta/power delta is still greater or equal to 1 in the current state of the uarches, it is below 1 since SB at least in Intel uarch, so lower IPC designs like BT or Mullins/Beema have no trouble being on par perf/watt wise despite being one and two nodes late respectively, at equal nodes it would be a slaughter.

Second limitation is the poor perf/watt of Intel s GPU, this is an area that is not compensated by the 14-16nm node.

Lastly, their current process does not seem to be up to their own expectations, if any indication the current chips are surely cherry picked and this is telling about the dispersions they must have in their process, with yields surely in the low double digit, something like 20-25%.

Citation needed on all the above.

 

[Abwx]

Citation needed on all the above.

You could as well have added for the below, anyway you could have done better for your post 6666..

 

[mikk]

Here is a huge Core M review: http://www.notebookcheck.com/Im-Test-Intel-Core-M-5Y70-Broadwell.129544.0.html


Lots of gaming benchmarks. Performance isn't groundbreaking, Haswell-U is much faster.


Interesting frequency log from Dota 2: http://www.notebookcheck.com/fileadmin/Notebooks/Sonstiges/Prozessoren/Broadwell/dota2.png


GPU runs roughly at 400 Mhz only most of the time and CPU only 800 Mhz. No wonder performance isn't great. For a consistent performance over several minutes and longer Broadwell requires 10+ watts it seems.

 

[witeken]

Note how it destroys Apple A8X in Sunspider. Unfortunately AT hasn't measured power consumption.

 

[mikk]

Sunspider is a short benchmark, favouring Core-M and its Turbo. The drop comes in longer benchmarks.