AMD Carrizo Pre-release thread

[Abwx]

Yes i know what that means exactly, it works like that since SB in both desktop and mobile.

Also Intel cant fake it, TDP value is what OEM use to calculale the apropiate cooling system, there whould be massive problems if Intel where faking TDP in any way.

Surely, and you re just fueling the point that perf/Watt comparisons cant be made by taking numbers at face value, a significant set of correlating datas is neessary before doing conclusions.

So far AMD released enough datas about Carrizo, we know at wich frequency the CPU part reach 15W.

This certainly is more worthy of confidence than Intel who release a list of CPUs that start at 1.6GHz and end at 2.6GHZ and all are 15W, even if the laws of physics say that the latter will have a TDP 2.64x higher than the former...

Btw, if 15W was possible at 2.6 then the 4GHz DT SKL would be specced at 60W, curious that people think that what is not possible for a DT would be a cakewalk once stuck in a laptop, must be another of thoses die hard myths..

 

[Abwx]

It seems prefectly reasonable that that same chip can run the much less demanding CB at 3.1 ghz at 15W.

You realize that this mean that the DT BDW 4C/8T would be Cinebenching at barely 28W at 3.1GHz..?..

And at 56W at 4.35GHz..?..

So why what is not possible in DTs would be magicaly doable once one stick the device in a laptop..?..

I think that you know what is a basic square law, that is f(x) = x^2, so why the need to forever come with assumptions that are forever in a collision course with the most elementary laws of physics..?.

 

[Shivansps]

Surely, and you re just fueling the point that perf/Watt comparisons cant be made by taking numbers at face value, a significant set of correlating datas is neessary before doing conclusions.

So far AMD released enough datas about Carrizo, we know at wich frequency the CPU part reach 15W.

This certainly is more worthy of confidence than Intel who release a list of CPUs that start at 1.6GHz and end at 2.6GHZ and all are 15W, even if the laws of physics say that the latter will have a TDP 2.64x higher than the former...

Btw, if 15W was possible at 2.6 then the 4GHz DT SKL would be specced at 60W, curious that people think that what is not possible for a DT would be a cakewalk once stuck in a laptop, must be another of thoses die hard myths..

Its not like that, TDP is NOT calculated for the CPU base clocks, neither is for the CPU at max turbo, its calculated for CPU+IGP at base clocks(for the ULVs).

So if a cpu is 15W for CPU+IGP is clear that it can turbo CPU alone using the IGP tdp room.

AND you need to consider TDP2 value, its hidden and you cant see it, but if the cooling system can handle it you can use it. And thats a good thing.
But one thing is clear, if CPU can mantain turbos and CPU Package power is inside the CPU TDP, them there is nothing strange/hidden/facking about it. Period.

 

[Abwx]

Its not like that, TDP is NOT calculated for the CPU base clocks, neither is for the CPU at max turbo, its calculated for CPU+IGP at base clocks(for the ULVs).

So if a cpu is 15W for CPU+IGP is clear that it can turbo CPU alone using the IGP tdp room.

I know that the CPU can use about all the TDP if the GPU is unused or that there s hidden values but that s really not the point.

The point is that there s a frequency at wich a SoC will dissipate 15W when loaded by CB 11.5.

Now for intel it s the complete smog, they are deliberatly hiding the number, all the way while releasing a whole line of so called 15W CPUs that actualy encompass a ratio of 3 TDP wise.

Of course there are means to extract this number from existing reviews, it just happen that it doesnt correlate with what some people are mandating others to believe, among other that Intel command the laws of physics.

For Carrizo it is at 2.25-2.3GHz and the data has been made available by AMD long before NBC tested a laptop, now if someone pretend that an Intel CPU does better he just has to show us the equivalent numbers published by Intel for their products, where are they..?..

Here perhaps, at the alinea 10 ?.. :

Or eventualy here :

Didnt someone say that Carrizo was competitive with i3..?..

 

[Enigmoid]

I know that the CPU can use about all the TDP if the GPU is unused or that there s hidden values but that s really not the point.

The point is that there s a frequency at wich a SoC will dissipate 15W when loaded by CB 11.5.

Unless you are looking at a i7-U chip at turbo the frequency at which intel's U lineup consume more than 15W on CB 11.5 is greater than the maximum frequency they are allowed to run at.

Didnt someone say that Carrizo was competitive with i3..?..

In terms of CPU performance, looking at reviews with more reasonable 1600 mhz RAM (no OEM is going to use 2133 mhz RAM on lower end notebooks) the CPU of the 8700p is slightly behind an i3U. An FX 8800p would be slightly above an i3-U but still in the same class.

 

[Dresdenboy]

Edit: My work laptop is a full power mobile, (i.e. not U model) sandy bridge i5, and it is plenty fast enough, although if there is a lot of IT crap running in the background it can drag at times.

I also had a SB laptop (i7 - I think a ~35W one) with dGPU and heavy cooler (2.7 kg). Due to the HDD it also suffered from having to regularly run the background scans (software inventory). Only due to that encrypted HDD and the background tasks it often felt like a slug. At 15W now, 1.4 kg and with SSD the new one feels like an improvement in nearly all areas.

 

[ET]

That is not true, the Intel U tier models are still a whole class faster than 4core Bay Trail Atoms.

Looking at Notebookcheck reviews, the A10-8700P at 15W with 1333 memory is 5-60% faster than a A8-7410 on the CPU side (with more benchmarks showing the high difference), and about twice as fast in GPU tests and games. That's a pretty big difference in my book.

The average customer isn't that smart, but they understand its a bad idea to pay more than 500 dollar for a CPU which doesn't even make it to 2.5 GHz.

Sounds like you have no idea what an average consumer is like. Go ask a hundred people in the street what the CPU speed is in their PC.

 

[maarten12100]

What weighs a lot in this case, is that OEMs decide how long it will run, because they will pick the accu size. Because that factor can't be controlled, I'd like to focus on things that CAN be compared, such as performance gains.

A smaller battery would reduce the weight obviously. By the way you can absolutely control the battery type and size if it's not enough just mount some extra cells in parallel and you're done.

Carrizo is comparable to an i3U in terms of CPU performance. It is significantly weaker than a 940m.

http://www.notebookcheck.net/HP-Pavilion-17-g054ng-Notebook-Review.147794.0.html

Looks like a 940m is around 50-100% stronger.

With a dgpu you also have fewer concerns about tdp limitations. Just make sure that notebook doesn't overheat.

The 840/940m is going to be consuming about as much as the entire Carrizo SoC configured at 25W so it should perform better. You're basically looking at a little less than double the power consumption under load.

Well, dual graphics MAY fail to deliver the promised performance in some cases, due to bad driver optimisation, but there are enough benchmarks which prove it's a promising concept, which actually delivers good power when playing certain games. For me, I'd just be glad to have the option to use DX. After all, having a IGP and dGPU which are almost identical and NOT be able to use them together, would be a waste of potential.

A 40% gain for a 100% increase in power consumption is not good in my book. In DX12 however we will almost definitly see really nice scaling which could make it worth it.

 

[Abwx]

the CPU of the 8700p is slightly behind an i3U. An FX 8800p would be slightly above an i3-U but still in the same class.

You should better read the review, this will spare us obvious lies like this one..

The i3s are behind, and this despite using more power than the 8700P.

http://www.notebookcheck.com/Test-HP-Pavilion-17-g054ng-Notebook.147334.0.html

The i3s used as comparisons, 15"6 of course to help a little, and the two are more power hungry than Carrizo.

http://www.notebookcheck.com/Test-Asus-F555LJ-Core-i3-5010U-GeForce-920M-Notebook.143512.0.html

http://www.notebookcheck.com/Test-Lenovo-ThinkPad-Edge-E550-Notebook.137307.0.html

To get better scores with the i5 they inflate the power numbers and call it a day.

Now if you want to further discuss those numbers bring us first Intel s official numbers, i m not interested in contradicting what is known to be non sense by the very writers of thoses posts, their real point being to give Intel credits that are undue.

Read again my answer to you in a previous post :

You realize that this mean that the DT BDW 4C/8T would be Cinebenching at barely 28W at 3.1GHz..?..

And at 56W at 4.35GHz..?..

So why what is not possible in DTs would be magicaly doable once one stick the device in a laptop..?..

All you re doing is to recycle those beliefs in every single one of your posts, using arguments that end being based on physical impossibilities.

 

[sm625]

http://www.notebookcheck.com/Test-HP-Pavilion-17-g054ng-Notebook.147334.0.html

This is a disaster of epic proportions. The HP Pavilion 17-g013ng with an i7-5500U gets 32% more battery life in the H.264 test, and 30% more idle battery life. Same exact battery. AMD should just call it quits right now because there is no point. The HP Pavilion 17-g013ng will end up being cheaper simply due to economy of scale. And it even has a frickin nvidia gpu!

 

[Abwx]

The HP Pavilion 17-g013ng with an i7-5500U gets 32% more battery life in the H.264 test, and 30% more idle battery life

Lol, a 41Wh battery, 5.3W at the main and still 10 hours battery life..?.

Yet the Carrizo use 4.1W at the main, has the same battery and they measured it at about 8H..?.

Not that they used a device charged at only 93% that is cut at 9% remaining charge while the Intel cut at 5%...

Anyway thank you for your expertise and prove that Intel numbers are taken at face value, i have another exemple if you want where the laptop use 24W at the main in the stress test even if HVinfo show the CPU at 21W...


Why this one doesnt last as long despite having a 20% bigger battery :

http://www.notebookcheck.com/Test-Update-Acer-Aspire-E5-771G-71PT-Notebook.135585.0.html

20% bigger battery and 20% lower battery time...

 

[dahorns]

This is a disaster of epic proportions. The HP Pavilion 17-g013ng with an i7-5500U gets 32% more battery life in the H.264 test, and 30% more idle battery life. Same exact battery. AMD should just call it quits right now because there is no point. The HP Pavilion 17-g013ng will end up being cheaper simply due to economy of scale. And it even has a frickin nvidia gpu!

Yeah, I can't believe any Abwx could have ignored that bit.

HP Pavilion 17-g013ng
CPU: Intel Core i7-5500U 2.4 GHz (Intel Core i7)
dGPU: NVIDIA GeForce 940M - 2048 mb
Memory: 12,288 MB , DDR3-1600, Dual-Channel
Screen:17.3 Inch, 16:9, 1920x1080 Pixel, Samsung, IPS
Hard-Drive: Seagate ST1000LM014 Solid State Hybrid Drive, 1000 GB
Battery: 41 Wh
Price: 899 Euro

HP Pavilion 17-g054ng
CPU: AMD A10-8700P 1.8 GHz (AMD A-Series)
dGPU: AMD Radeon R8 M365DX, Dual-Graphics (Radeon R6 Carrizo)
Memory: 8192 MB, Dual-Channel Memory
Screen: 17.3 Inch 16:9, 1920x1080 Pixel, Samsung, TN-Panel, LED-Backlight
Hard-Drive:Seagate Momentus SpinPoint 1000 GB, 5400 U/Min
Battery: 41 Wh
Price: 599 Euro

The Broadwell based system is almost twice as fast in CPU tasks, faster in gaming (in some instances almost twice as fast) and has 20-40% longer battery life. . .

 

[Enigmoid]

You should better read the review, this will spare us obvious lies like this one..

The i3s are behind, and this despite using more power than the 8700P.

http://www.notebookcheck.com/Test-HP-Pavilion-17-g054ng-Notebook.147334.0.html

The i3s used as comparisons, 15"6 of course to help a little, and the two are more power hungry than Carrizo.

http://www.notebookcheck.com/Test-Asus-F555LJ-Core-i3-5010U-GeForce-920M-Notebook.143512.0.html

http://www.notebookcheck.com/Test-Lenovo-ThinkPad-Edge-E550-Notebook.137307.0.html

To get better scores with the i5 they inflate the power numbers and call it a day.

There is simply not enough data to make any conclusions. Looking at the maximum amount of power the platform can consume does not allow one to infer the efficiency of the platform.

Those power numbers are also the peak power numbers, not average.

 

[Abwx]

Yeah, I can't believe any Abwx could have ignored that bit.


Battery: 41 Wh

10H for the i7 that consume 5.3W and 8H for Carrizo and his 4.1W using the same 41Wh battery..

You are right, thoses numbers should be ignored, if they were true the i7 5500U i linked, and wich use a 48WH battery, would have no trouble lasting more than the 8H it display...

http://www.notebookcheck.com/Test-Update-Acer-Aspire-E5-771G-71PT-Notebook.135585.0.html

Now can the intel sales forces keep this thread alone, although they are welcomed if they can provide us with Intel s equivalent of this :

If there s no such infos please stop polluting the thread with fairy tales and assumptions that have no basis since it s Intel s very marketing to hide the numbers ,and then send an army of marketers to spin the facts.





Now back on topic an interesting IPC comparison at Planet3D between the CMT designs, including EXV, in the following benches :

• Cinebench 11.5 x64
• Cinebench 15 x64
• 7-Zip 9.38 x64
• Truecrypt 7.1a
• Fritz (Schach-Benchmark) 4.2
• LAME 3.99.5 x64
• AIDA64 Engineer 5.30

http://www.planet3dnow.de/cms/18564...cavator-leistungsvergleich-der-architekturen/

 

[AtenRa]

HP Pavilion 17-g013ng
CPU: Intel Core i7-5500U 2.4 GHz (Intel Core i7)

Price: 899 Euro

HP Pavilion 17-g054ng
CPU: AMD A10-8700P 1.8 GHz (AMD A-Series)

Price: 599 Euro

That is 300 Euro more

 

[monstercameron]

10H for the i7 that consume 5.3W and 8H for Carrizo and his 4.1W using the same 41Wh battery..

You are right, thoses numbers should be ignored, if they were true the i7 5500U i linked, and wich use a 48WH battery, would have no trouble lasting more than the 8H it display...

http://www.notebookcheck.com/Test-Update-Acer-Aspire-E5-771G-71PT-Notebook.135585.0.html

Now can the intel sales forces keep this thread alone, although they are welcomed if they can provide us with Intel s equivalent of this :

If there s no such infos please stop polluting the thread with fairy tales and assumptions that have no basis since it s Intel s very marketing to hide the numbers ,and then send an army of marketers to spin the facts.





Now back on topic an interesting IPC comparison at Planet3D between the CMT designs, including EXV, in the following benches :


http://www.planet3dnow.de/cms/18564...cavator-leistungsvergleich-der-architekturen/

wow carrizo is a nice improvement is some areas, amd gimme carrizo desktop naow!

 

[mrmt]

wow carrizo is a nice improvement is some areas, amd gimme carrizo desktop naow!

Which one, the one from AMD slides or the one reaching actual products?

 

[Abwx]

Lol, Maxon caught cheating their own bench, i often pointed that the R15 revision was a scam to help their compiler provider.

Cinebench 11.5 MT :

Cinebench R15 MT :

Cinebench 11.5 ST :

Cinebench R15 ST :

http://www.planet3dnow.de/cms/18564...er-architekturen/subpage-rendering-cinebench/

 

[ShintaiDK]

Please elaborate, where is the cheat?

All you show is AMDs performance went up with new uarchs.

Let me guess, you forgot to cross compare with Intel.

CB 11.5 ST:
4790K 9.55
6700K 10.22

CB 15 ST:
4790K 181
6700K 183

 

[monstercameron]

Which one, the one from AMD slides or the one reaching actual products?

from the planet3d uarch benchmarks.

 

[Enigmoid]

10H for the i7 that consume 5.3W and 8H for Carrizo and his 4.1W using the same 41Wh battery..

You are right, thoses numbers should be ignored, if they were true the i7 5500U i linked, and wich use a 48WH battery, would have no trouble lasting more than the 8H it display...

http://www.notebookcheck.com/Test-Update-Acer-Aspire-E5-771G-71PT-Notebook.135585.0.html

Now can the intel sales forces keep this thread alone, although they are welcomed if they can provide us with Intel s equivalent of this :

If there s no such infos please stop polluting the thread with fairy tales and assumptions that have no basis since it s Intel s very marketing to hide the numbers ,and then send an army of marketers to spin the facts.


Now back on topic an interesting IPC comparison at Planet3D between the CMT designs, including EXV, in the following benches :

http://www.planet3dnow.de/cms/18564...cavator-leistungsvergleich-der-architekturen/

Screen's are different.

The battery quicks are nothing new.

http://www.notebookcheck.net/Apple-MacBook-Air-13-2015-Notebook-Review.144375.0.html

Look at the MBA. 54 whr battery, 2.7W idle, 26Hr battery life idle.

Most likely some other power saving measures are kicking in.

 

[Abwx]

There is simply not enough data to make any conclusions. Looking at the maximum amount of power the platform can consume does not allow one to infer the efficiency of the platform.

Those power numbers are also the peak power numbers, not average.

http://forums.anandtech.com/showpost.php?p=37637227&postcount=2301

Ultimately the thread is about Carrizo, not about Intel s shenanigans to hide their power numbers and get the things explained by sales forces.

As said ad nauseam provide official numbers, in the waiting mines are the relevant ones.

Please elaborate, where is the cheat?

All you show is AMDs performance went up with new uarchs.

Let me guess, you forgot to cross compare with Intel.

26% improvement in CB 11.5 and 13.6% in CB R15, the comparisons you re using are not relevant a all, if you want relevant ones check here :

http://www.notebookcheck.com/Test-HP-Pavilion-17-g054ng-Notebook.147334.0.html

Look at the scores of the i3s in respect of Carrizo in both version of the bench, they are behind in the former and ahead in the latter, thing is that you wont see such scam with Povray successive versions..

wow carrizo is a nice improvement is some areas, amd gimme carrizo desktop naow!

If the slide below is genuine there will be one but only in 12 months.

 

[Evleos]

Seems like the work AMD put into the new FPU pay off, from 50 -> 99% IPC improvement.

http://www.planet3dnow.de/cms/18564...vergleich-der-architekturen/subpage-aida-fpu/

 

[Enigmoid]

Seems like the work AMD put into the new FPU pay off, from 50 -> 99% IPC improvement.

http://www.planet3dnow.de/cms/18564...vergleich-der-architekturen/subpage-aida-fpu/

Its AVX 2

 

[LTC8K6]

https://en.wikipedia.org/wiki/Advanced_Vector_Extensions#CPUs_with_AVX2