How do you think Bristol Ridge will do as a mobile processor compared to Carrizo?

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

How do you think Bristol Ridge will do as a mobile processor compared to Carrizo?

  • Much better

  • Somewhat better

  • a little bit better

  • no change


Results are only viewable after voting.

cbn

Lifer
Mar 27, 2009
12,968
221
106
With Bristol Ridge DDR-4 2133MHz will be supported on the 15W TDP models as well.

So, im expecting a nice performance boost on the 15W TDP models, lets say 20% compared to Carrizo.

The clocks on the 15W are so low though when CPU and iGPU are both being used (see post #12).

I am having a hard time imagining how DDR4 would help.
 

AtenRa

Lifer
Feb 2, 2009
14,003
3,362
136
The clocks on the 15W are so low though when CPU and iGPU are both being used (see post #12).

I am having a hard time imagining how DDR4 would help.

Yea i had to say that i also expect higher clocks/Watt from Bristol Ridge.
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
DDR-3 2133MHz is only supported on the 35W TDP Carrizo models.

There are no Carrizos with TDP higher than 15W. All of the current models support cTDP and DDR-2133 becomes unlocked when cTDP is set to 21W or higher. Only A6-series parts are restricted to DDR-1600, regardless of the cTDP.
 

AtenRa

Lifer
Feb 2, 2009
14,003
3,362
136
There are no Carrizos with TDP higher than 15W. All of the current models support cTDP and DDR-2133 becomes unlocked when cTDP is set to 21W or higher. Only A6-series parts are restricted to DDR-1600, regardless of the cTDP.

Yes i know bad wording, i was thinking of when the OEM allow 35W TDPs.

Now since DDR-4 entry-level frequency is 2133MHz (Jedec), then every Bristol Ridge Laptop even at 15W will come with DDR-4 2133MHz. And as i have said im expecting higher clocks per Watt, then we could see a nice performance boost at 15W. I will dare to say that 15W Bristol Ridge will reach or even exceed Carrizos 35W performance.
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
I will dare to say that 15W Bristol Ridge will reach or even exceed Carrizos 35W performance.

Just because of DDR4? It just isn´t possible since it is the same piece of silicon as Carrizo (a refresh without any physical differences).
 

AtenRa

Lifer
Feb 2, 2009
14,003
3,362
136
Just because of DDR4? It just isn´t possible since it is the same piece of silicon as Carrizo (a refresh without any physical differences).

Not from DDR-4 only but from higher clocks per watt as well.

I will have to remind everyone that Kabini/Temash is the same as Beema/Mulins on the same 28nm process and yet very different in clocks/watt and performance.
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
Not from DDR-4 only but from higher clocks per watt as well.

I will have to remind everyone that Kabini/Temash is the same as Beema/Mulins on the same 28nm process and yet very different in clocks/watt and performance.

Kabini and Mullins are completely different things. Kabini is 16h 00-0Fh while Mullins is 30-3Fh. Kabini has Jaguar cores while Mullins is Puma based. They are different generation parts with various tweaks in Mullins. Meanwhile Bristol Ridge is the same exact generation as Carrizo (15h 60-6Fh). Trinity and it´s refresh Richland were the same generation (15h 10-1Fh) and Kaveri and it´s refresh Godavari were these same generation (15h 30-3Fh). No physical differences, just software configuration.

There will be no change in "performance per watt", obviously. High-end Bristol Ridge will perform better since it is less TDP restricted than Carrizo (15W vs. 35W), support higher memory frequency (2133 vs. 2400MHz) and has higher maximum frequency due the higher TDP.

Stoney Ridge meanwhile is a different design, belonging to different family (70-7Fh). If there is going to be any further improvements in performance-per-watt, then they´ll be in Stoney Ridge and not in Bristol Ridge. I´m not holding my breath for futher improvement in Stoney Ridge either thou.
 
Last edited:

dark zero

Platinum Member
Jun 2, 2015
2,655
140
106
There really arent many 35 watt laptops anymore. Pretty much the only ones are quad intel, and those are usually paired with a relatively powerful dgpu.
That's why I say that the OEMs are killing themselves the notebook industry by their sheer stupidity.

Bussiness needs 35 watts laptops and they gives them the U trash in exchange. Someone needs to ask MS to make a notebook with 35 W Intel and AMD processor.

As an example, I am forced to use that, is nice to not to heat, but.... to have pretty low performance? That exchange marks a lot of time lost.
 

AtenRa

Lifer
Feb 2, 2009
14,003
3,362
136
Kabini and Mullins are completely different things. Kabini is 16h 00-0Fh while Mullins is 30-3Fh. Kabini has Jaguar cores while Mullins is Puma based. They are different generation parts with various tweaks in Mullins. Meanwhile Bristol Ridge is the same exact generation as Carrizo (15h 60-6Fh). Trinity and it´s refresh Richland were the same generation (15h 10-1Fh) and Kaveri and it´s refresh Godavari were these same generation (15h 30-3Fh). No physical differences, just software configuration.

There will be no change in "performance per watt", obviously. High-end Bristol Ridge will perform better since it is less TDP restricted than Carrizo (15W vs. 35W), support higher memory frequency (2133 vs. 2400MHz) and has higher maximum frequency due the higher TDP.

Stoney Ridge meanwhile is a different design, belonging to different family (70-7Fh). If there is going to be any further improvements in performance-per-watt, then they´ll be in Stoney Ridge and not in Bristol Ridge. I´m not holding my breath for futher improvement in Stoney Ridge either thou.

OK i had a second look in the Bristol Ridge specs and it seems 15W TDP will only support up to 1866MHz DDR-3 or DDR-4.
And if perf/watt is not that higher than Carrizo then the performance gains will be very small.
 

monstercameron

Diamond Member
Feb 12, 2013
3,818
1
0
I wish we could make a kickstarter just to build the optimal amd laptop and demonstrate to oems how it should be done. This is just getting too ridiculous...
 

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106
I wish we could make a kickstarter just to build the optimal amd laptop and demonstrate to oems how it should be done. This is just getting too ridiculous...

Sure you wouldn't just create a product with even less sales and higher cost? The chips are still the same, and they are the problem.
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
Note that the APUs that actually have all CUs enabled (eg A10) are hard to come by where I live and usually are just massively overpriced eg. about same price as better performing i5 + dGPU.

Are you referring to desktops or mobile?

Mobile actually.

That is what I thought you meant.

I am just asking because only the FX mobile processors have all 12 CUs enabled. The A10 mobile processors actually only have 10 CUs enabled (unlike the desktop A10s).
 

coercitiv

Diamond Member
Jan 24, 2014
7,225
16,982
136
Sure you wouldn't just create a product with even less sales and higher cost? The chips are still the same, and they are the problem.
Pretty sure he wouldn't do that. The 1st "fix" consists in removing any dGPU OEMs insist to add to the mix and using the extra TDP room for a proper cTDP implementation (15W while on battery, 20-35W TDP while plugged in). 2nd (optional) fix is to remove the optical disk drive and use the savings to ensure a decent battery and a somewhat slimmer unit.

Platform cost goes down, and from a marketing standpoint all they need to do is show the damn thing plays League or other MOBAs.

Does anyone know if Bristol Ridge comes with updates on video decode side?
 

monstercameron

Diamond Member
Feb 12, 2013
3,818
1
0
Pretty sure he wouldn't do that. The 1st "fix" consists in removing any dGPU OEMs insist to add to the mix and using the extra TDP room for a proper cTDP implementation (15W while on battery, 20-35W TDP while plugged in). 2nd (optional) fix is to remove the optical disk drive and use the savings to ensure a decent battery and a somewhat slimmer unit.

Platform cost goes down, and from a marketing standpoint all they need to do is show the damn thing plays League or other MOBAs.

Does anyone know if Bristol Ridge comes with updates on video decode side?
Pretty much this, maybe add ssd as an option or if possible a vrr capable display.
 

AtenRa

Lifer
Feb 2, 2009
14,003
3,362
136
A good and Cheap AMD Laptop like the one coercitiv was talking,would cannibalize OEMs Intel Laptop sales.

so the solution,

Crap AMD expensive Laptops, problem solved.
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
I wish we could make a kickstarter just to build the optimal amd laptop and demonstrate to oems how it should be done. This is just getting too ridiculous...

Sure you wouldn't just create a product with even less sales and higher cost? The chips are still the same, and they are the problem.

I don't think the chips are the problem. If we can get 35W Bristol Ridge 4C + 512sp to average around 2.5 Ghz (CPU) and 750 Mhz (iGPU) in games then that is a usable level of performance.

If I were to guess the bigger issue is volume and therefore my hope is AMD would cancel Bristol Ridge A10 for the AM4 desktop and instead re-bin those chips for FP4 (mobile). Then we could have two levels of 4C + 512sp mobile chips: An FX SKU and a A10 SKU. Then the 4C + 384sp die can be called A8.

FX Bristol Ridge: 4C + 512sp
A10 Bristol Ridge: 4C + 512sp (slightly slower bin)
A8 Bristol Ridge: 4C + 384sp

No A6 (Since Stoney ridge 2C/192sp exists to cover dual cores on 35W FP4)
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
One thing I am hoping for on Windows 10 Bristol Ridge laptops (and I am sure no OEM will care about) is Full Linux support (with secure boot) out of the box.

More devices available = more folks wanting to improve AMD on Linux.

P.S. Another thing I think AMD can do encourage Linux development is to re-enter competition with Nvidia at the low end desktop dGPU. This again, increases Radeon numbers....which naturally attracts developers. And I wouldn't be surprised if cancelling A10 Bristol Ridge for AM4 desktop helps this process along.
 
Last edited:

erunion

Senior member
Jan 20, 2013
765
0
0
Pretty sure he wouldn't do that. The 1st "fix" consists in removing any dGPU OEMs insist to add to the mix and using the extra TDP room for a proper cTDP implementation (15W while on battery, 20-35W TDP while plugged in). 2nd (optional) fix is to remove the optical disk drive and use the savings to ensure a decent battery and a somewhat slimmer unit.

a thin 15W chassis can't support a 35W turbo. AMD spent years stuck competing against modern and thin Intel ULV designs with 35W chips. With Carrizo AMD finally has a real ULV chip, but with the accompanying performance drop. Going back to thick 35W designs isn't a magic solution. It will put them right back where they were, competing against Intel+discrete gpus.

The problem is the chips. There is no single area that AMDs big APU have an undisputed lead over Intel, so AMD/oems are stuck trying to find a middle ground sweetspot where they offer a superior mix of price/feature/performance. They've tried different formulas but haven't found it yet.
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
Unless you have actually tried a laptop which produces > 42W of heat, it is actually pretty hard to imagine how much noise and heat that actually is. On desktop standards that amount is basically nothing, but on a laptop it is something which I cannot personally accept.

I have the Carrizo reference system, which can be configured for 35W / 42W cTDP. While the performance at 35W cTDP is impressive, everything else is just horrible. The system will run extremely hot, loud and it will drain the battery in no time. Mind you that neither the cooling, size of the housing or the battery are anywhere close to the worst, loudest or smallest you can get with this kind of a system. The system has a dedicated cooling element for the APU, the housing is rather large and the battery has capacity of > 4700mAh (IIRC).

Regardless I use the higher cTDP options just for testing, since anything higher is just too hot, loud and consumes too much power.

Carrizo / Bristol Ridge on 16nm FF+ with better memory controller (>DDR4-2933) on it and a UVD with VP9 support would be pretty perfect :eek:
 

monstercameron

Diamond Member
Feb 12, 2013
3,818
1
0
a thin 15W chassis can't support a 35W turbo. AMD spent years stuck competing against modern and thin Intel ULV designs with 35W chips. With Carrizo AMD finally has a real ULV chip, but with the accompanying performance drop. Going back to thick 35W designs isn't a magic solution. It will put them right back where they were, competing against Intel+discrete gpus.

The problem is the chips. There is no single area that AMDs big APU have an undisputed lead over Intel, so AMD/oems are stuck trying to find a middle ground sweetspot where they offer a superior mix of price/feature/performance. They've tried different formulas but haven't found it yet.
There is some truth to what you are saying although you seem to be implying that oems only choose the best components. That they won't choose amd because amd is mostly inferior to Intel. Unless I'm mistaken, that seems very far from reality. Not that it is always sou d business but oems do like to experiment amd even outside of that having a second source of components is a smart thing to do.


All that said amd apus, as has been shown by benchmarks, aren't that far in terms of performance and probably near identical user experience. The only places Intel has much more to offer is best and servers.

Maybe I'm just too biased to see what the proponents of Intel see.
 

monstercameron

Diamond Member
Feb 12, 2013
3,818
1
0
Unless you have actually tried a laptop which produces > 42W of heat, it is actually pretty hard to imagine how much noise and heat that actually is. On desktop standards that amount is basically nothing, but on a laptop it is something which I cannot personally accept.

I have the Carrizo reference system, which can be configured for 35W / 42W cTDP. While the performance at 35W cTDP is impressive, everything else is just horrible. The system will run extremely hot, loud and it will drain the battery in no time. Mind you that neither the cooling, size of the housing or the battery are anywhere close to the worst, loudest or smallest you can get with this kind of a system. The system has a dedicated cooling element for the APU, the housing is rather large and the battery has capacity of > 4700mAh (IIRC).

Regardless I use the higher cTDP options just for testing, since anything higher is just too hot, loud and consumes too much power.

Carrizo / Bristol Ridge on 16nm FF+ with better memory controller (>DDR4-2933) on it and a UVD with VP9 support would be pretty perfect :eek:
The vp9 unicorn...I don't think amd will ever support vp9 on fixed function hardware at this point. They didn't even announce support for it on their Polaris PR dump.

Quad channel ddr4 would also be nice...
 

erunion

Senior member
Jan 20, 2013
765
0
0
There is some truth to what you are saying although you seem to be implying that oems only choose the best components. That they won't choose amd because amd is mostly inferior to Intel. Unless I'm mistaken, that seems very far from reality. Not that it is always sou d business but oems do like to experiment amd even outside of that having a second source of components is a smart thing to do.
.

OEMS absolutely do want to carry AMD laptops, but they are faced with the challenge of making AMD laptops that fit somewhere in their product lineup.