Niiiiiiiice... would that make an x86 CPU Card doable...? ;)
On 11/19/2013 2:43 PM, luke.leighton wrote:
http://www.amd.com/us/press-releases/Pages/amd-2014-mobile-apu-2013nov13.asp...
at laaaast.
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On Tue, Nov 19, 2013 at 7:54 PM, Christopher Havel laserhawk64@gmail.com wrote:
Niiiiiiiice... would that make an x86 CPU Card doable...? ;)
uhh.. yeah :)
On 11/19/2013 04:43 PM, Christopher Havel wrote:
On 11/19/2013 4:40 PM, luke.leighton wrote:
uhh.. yeah :)
Great! I'll take two, when they're ready :D maybe three... we'll see...
Ah... the 'shut up and take my money' attitude at it's best!
Let's get the first A20 cards out there and sink some of that money into the few other card designs that luke has mapped out.
http://rhombus-tech.net/freescale/iMX6/news/ http://rhombus-tech.net/jz4760/news/
By that time the AMD APUs (6-12 months?) should actually be available. Hopefully also enough revenue in the pipe to speed along development of a card based on them too.
*Walks off to the kitchen* Computer! Tea, Earl Grey, Hot.
On 11/19/2013 4:53 PM, Scott Sullivan wrote:
Ah... the 'shut up and take my money' attitude at it's best!
http://i0.kym-cdn.com/photos/images/original/000/264/241/9e9.gif
in other words: exactly! ;)
*Walks off to the kitchen* Computer! Tea, Earl Grey, Hot.
RED ALERT!! RED ALERT!! You're not Patrick Stewart!
(Sorry if I'm being obnoxious -- I tried to resist. Really, I did. I think I must be a capacitor because it didn't work very well. No harm meant, just a bit of good humor. Without the ice cream.)
2013/11/19 Christopher Havel laserhawk64@gmail.com
On 11/19/2013 4:53 PM, Scott Sullivan wrote:
Ah... the 'shut up and take my money' attitude at it's best!
http://i0.kym-cdn.com/photos/images/original/000/264/241/9e9.gif
in other words: exactly! ;)
*Walks off to the kitchen* Computer! Tea, Earl Grey, Hot.
RED ALERT!! RED ALERT!! You're not Patrick Stewart!
(Sorry if I'm being obnoxious -- I tried to resist. Really, I did. I think I must be a capacitor because it didn't work very well. No harm meant, just a bit of good humor. Without the ice cream.)
http://www.iowascribe.com/uploads/6/9/0/1/6901504/4926612_orig.jpg
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
On Tue, 2013-11-19 at 19:43 +0000, luke.leighton wrote:
http://www.amd.com/us/press-releases/Pages/amd-2014-mobile-apu-2013nov13.asp...
at laaaast.
Read carefully? It doesn't say 2W - it sez 2x performance.
On Wed, Nov 20, 2013 at 8:54 AM, joem joem@martindale-electric.co.uk wrote:
On Tue, 2013-11-19 at 19:43 +0000, luke.leighton wrote:
http://www.amd.com/us/press-releases/Pages/amd-2014-mobile-apu-2013nov13.asp...
at laaaast.
Read carefully? It doesn't say 2W - it sez 2x performance.
roadmap document, page 4.
On Wed, Nov 20, 2013 at 10:08 AM, luke.leighton luke.leighton@gmail.comwrote:
On Wed, Nov 20, 2013 at 8:54 AM, joem joem@martindale-electric.co.uk wrote:
On Tue, 2013-11-19 at 19:43 +0000, luke.leighton wrote:
http://www.amd.com/us/press-releases/Pages/amd-2014-mobile-apu-2013nov13.asp...
at laaaast.
Read carefully? It doesn't say 2W - it sez 2x performance.
roadmap document, page 4.
“Mullins ” APU
2-4 “Puma” CPU Cores GCN Graphics Compute Units AMD Security Processor ~2W SDP
On Wednesday, November 20, 2013 04:12 PM, Hrvoje Lasic wrote:
On Wed, Nov 20, 2013 at 10:08 AM, luke.leighton <luke.leighton@gmail.com mailto:luke.leighton@gmail.com> wrote:
On Wed, Nov 20, 2013 at 8:54 AM, joem <joem@martindale-electric.co.uk <mailto:joem@martindale-electric.co.uk>> wrote: > On Tue, 2013-11-19 at 19:43 +0000, luke.leighton wrote: >> http://www.amd.com/us/press-releases/Pages/amd-2014-mobile-apu-2013nov13.aspx >> >> at laaaast. > > Read carefully? It doesn't say 2W - it sez 2x performance. roadmap document, page 4."Mullins " APU 2-4 "Puma" CPU Cores GCN Graphics Compute Units AMD Security Processor ~2W SDP
For the ultra low power, APU they use SDP (Scenario Design Power), whereas for the other processors, they don't mention anything, so I'd guess they use TDP (Thermal Design Power).
Anandtech says SDP is ~2W, but TDP is 4.5W for AMD Mullins APU. http://www.anandtech.com/show/7514/amd-2014-mobile-apu-update-beema-and-mull...
I'm not sure how this can be compared to ARM SoC power consumption, as they may use a different metric.
On Wed, Nov 20, 2013 at 10:44 AM, Jean-Luc Aufranc <cnxsoft@cnx-software.com
wrote:
On Wednesday, November 20, 2013 04:12 PM, Hrvoje Lasic wrote:
On Wed, Nov 20, 2013 at 10:08 AM, luke.leighton luke.leighton@gmail.comwrote:
On Wed, Nov 20, 2013 at 8:54 AM, joem joem@martindale-electric.co.uk wrote:
On Tue, 2013-11-19 at 19:43 +0000, luke.leighton wrote:
http://www.amd.com/us/press-releases/Pages/amd-2014-mobile-apu-2013nov13.asp...
at laaaast.
Read carefully? It doesn't say 2W - it sez 2x performance.
roadmap document, page 4.
“Mullins ” APU
2-4 “Puma” CPU Cores GCN Graphics Compute Units AMD Security Processor ~2W SDP
For the ultra low power, APU they use SDP (Scenario Design Power), whereas for the other processors, they don't mention anything, so I'd guess they use TDP (Thermal Design Power).
Anandtech says SDP is ~2W, but TDP is 4.5W for AMD Mullins APU.
http://www.anandtech.com/show/7514/amd-2014-mobile-apu-update-beema-and-mull...
I'm not sure how this can be compared to ARM SoC power consumption, as they may use a different metric.
_
Hm, you are right, i just googled it. Shady marketing...
On Wed, 2013-11-20 at 16:44 +0700, Jean-Luc Aufranc wrote:
On Wednesday, November 20, 2013 04:12 PM, Hrvoje Lasic wrote:
On Wed, Nov 20, 2013 at 10:08 AM, luke.leighton luke.leighton@gmail.com wrote: On Wed, Nov 20, 2013 at 8:54 AM, joem joem@martindale-electric.co.uk wrote: > On Tue, 2013-11-19 at 19:43 +0000, luke.leighton wrote: >> http://www.amd.com/us/press-releases/Pages/amd-2014-mobile-apu-2013nov13.asp... >> >> at laaaast. > > Read carefully? It doesn't say 2W - it sez 2x performance.
roadmap document, page 4.“Mullins ” APU 2-4 “Puma” CPU Cores GCN Graphics Compute Units AMD Security Processor ~2W SDP
For the ultra low power, APU they use SDP (Scenario Design Power), whereas for the other processors, they don't mention anything, so I'd guess they use TDP (Thermal Design Power).
Anandtech says SDP is ~2W, but TDP is 4.5W for AMD Mullins APU. http://www.anandtech.com/show/7514/amd-2014-mobile-apu-update-beema-and-mull...
I'm not sure how this can be compared to ARM SoC power consumption, as they may use a different metric.
The entire EOMA68-A20 dual core with RAM and ethernet working is about 0.3A @ 5V or about 1.5W. A third less if its idling.
On Wednesday, November 20, 2013 05:10 PM, joem wrote:
On Wed, 2013-11-20 at 16:44 +0700, Jean-Luc Aufranc wrote:
On Wednesday, November 20, 2013 04:12 PM, Hrvoje Lasic wrote:
On Wed, Nov 20, 2013 at 10:08 AM, luke.leighton luke.leighton@gmail.com wrote: On Wed, Nov 20, 2013 at 8:54 AM, joem joem@martindale-electric.co.uk wrote: > On Tue, 2013-11-19 at 19:43 +0000, luke.leighton wrote: >> http://www.amd.com/us/press-releases/Pages/amd-2014-mobile-apu-2013nov13.asp... >> >> at laaaast. > > Read carefully? It doesn't say 2W - it sez 2x performance.
roadmap document, page 4.“Mullins ” APU 2-4 “Puma” CPU Cores GCN Graphics Compute Units AMD Security Processor ~2W SDP
For the ultra low power, APU they use SDP (Scenario Design Power), whereas for the other processors, they don't mention anything, so I'd guess they use TDP (Thermal Design Power).
Anandtech says SDP is ~2W, but TDP is 4.5W for AMD Mullins APU. http://www.anandtech.com/show/7514/amd-2014-mobile-apu-update-beema-and-mull...
I'm not sure how this can be compared to ARM SoC power consumption, as they may use a different metric.
The entire EOMA68-A20 dual core with RAM and ethernet working is about 0.3A @ 5V or about 1.5W. A third less if its idling.
Thanks. But Is it possible to estimate the power consumption of EOMA68-Mullins based on SDP or TDP for the same tasks? I understand manufacturers use SDP / TDP to provide adequate cooling solutions for x86 systems. So I'd assume the CPU power consumption would be less than 4.5W or even 2W, but we would still need to add the power consumption of other components (e.g. RAM), to get an idea.
I'm not sure how this can be compared to ARM SoC power consumption, as they may use a different metric.
The entire EOMA68-A20 dual core with RAM and ethernet working is about 0.3A @ 5V or about 1.5W. A third less if its idling.
Thanks. But Is it possible to estimate the power consumption of EOMA68-Mullins based on SDP or TDP for the same tasks? I understand manufacturers use SDP / TDP to provide adequate cooling solutions for x86 systems. So I'd assume the CPU power consumption would be less than 4.5W or even 2W, but we would still need to add the power consumption of other components (e.g. RAM), to get an idea.
If I go by my MK802 experience, I would say put in the max ratings and supply adequate ventilation or the box will crash when doing the grunt work when someone loads decent software.
The current EOMA68-A20 is hot enough - so I guess not much more option to pump more heat in.
That would imply EOMA68-Mullins would need vent holes and ability to shift around 6W of heat, which in turn would imply a miniature fan or a heat pump. Not looking forward to that because you may as well start with a big Intel chip if power consumption is not an issue.
As it takes months and months to make these into boards and get software going, personal opinion is that it is easier just to stick with making selling the existing ones, and then to wait a few months for chip geometries to shrink and then for someone bring out more powerful ARM with lower power consumption and greater features.
On Wed, Nov 20, 2013 at 2:13 PM, joem joem@martindale-electric.co.uk wrote:
That would imply EOMA68-Mullins would need vent holes and ability to shift around 6W of heat, which in turn would imply a miniature fan or a heat pump.
mass-volume. no fans. returns too great on mechanical parts. failure. plus heat pump as part of the standard would mean that *all* products would need to have a heat pump. that is extra overhead that would destroy the standard because low-cost products would no longer be low-cost.
so.... no. no fans. no heat pumps.
l.
On 11/20/2013 03:35 PM, luke.leighton wrote:
On Wed, Nov 20, 2013 at 2:13 PM, joem joem@martindale-electric.co.uk wrote:
That would imply EOMA68-Mullins would need vent holes and ability to shift around 6W of heat, which in turn would imply a miniature fan or a heat pump.
mass-volume. no fans. returns too great on mechanical parts. failure. plus heat pump as part of the standard would mean that *all* products would need to have a heat pump. that is extra overhead that would destroy the standard because low-cost products would no longer be low-cost.
so.... no. no fans. no heat pumps.
water cooling?
On Thu, Nov 21, 2013 at 12:01 PM, Vladimir Pantelic vladoman@gmail.com wrote:
mass-volume. no fans. returns too great on mechanical parts. failure. plus heat pump as part of the standard would mean that *all* products would need to have a heat pump. that is extra overhead that would destroy the standard because low-cost products would no longer be low-cost.
so.... no. no fans. no heat pumps.
water cooling?
i'm not sure if you're joking, vladimir :) assuming you're not: it would be the sort of thing that would be ok on a CPU Card, if it wasn't for the fact that it's only 5mm thick. no on the I/O boards because again, that would mean that *every* single system would have to cope with the *possibility* of a 6W (or above) CPU Card.
this will all be fine for the Type III PCMCIA Cards, that's on the roadmap at some point. there it will be absolutely fine to put in thermal cooling by default onto I/O boards because the cost of such systems would be higher anyway because the cost of the CPU Cards would also be correspondingly higher (and also much faster clock speeds).
l.
arm-netbook@lists.phcomp.co.uk
-
Christopher Havel -
Hrvoje Lasic -
Jean-Luc Aufranc -
joem -
luke.leighton -
mike.valk＠gmail.com -
Scott Sullivan -
Vladimir Pantelic