AMD LLano idles at 10.2 watts
#16
ubuntuf4n Wrote:Wow Nice Chart, where did you grab that one ?

Datapoint 1 versus datapoint 2

Besides that, as I posted earlier, the image quality of Llano seems to be head and shoulders above other IGP solutions:

Image

Now I only need to find a detailed Llano power consumption investigation.
Reply
#17
-EaS- Wrote:Comparable? More like 50-95% higher!

I don't know where your chart came from, who performed the tests, what they were testing, or what that chart is portraying. I gave the link for the lengthy review that led to to xBitLab's conclusion that I quoted. It was a lengthy head-to-head comparison of the E-350 and i3-2100T. For example, they found:

Image

There are other power consumption charts at:

http://www.xbitlabs.com/articles/cpu/dis...html#sect0

And remember, that's for a chip that is several times more powerful than the e-350.
Reply
#18
Zon2020 Wrote:I don't know where your chart came from, who performed the tests, what they were testing, or what that chart is portraying.

That chart is just a summary from the two datapoints I gave in post 16. It is a very detailed break-down of the different load scenarios that may happen during playback of media (which is the main purpose of a HTPC).

Not a simple "1080p playback" label which, as the chart shows, could mean as much as a 50% difference in power consumption even on the same system! So in itself "1080p playback" says nearly nothing. Smile

Quote:I gave the link for the lengthy review that led to to xBitLab's conclusion that I quoted. It was a lengthy head-to-head comparison of the E-350 and i3-2100T. For example, they found:

Image

True. But xbit-labs also tested E350 on a MSI board but 'conveniently' seem to have forgotten to use those numbers. Here is what the numbers would have looked like if they would've included them:

Image

Do you think it would have made a difference in the "1080p playback" test? I do!

I'm a little allergic to reviews that seem to conveniently forget things, and thus seem to write an article around pre-cooked conclusions.

Quote:There are other power consumption charts at:

http://www.xbitlabs.com/articles/cpu/dis...html#sect0

And remember, that's for a chip that is several times more powerful than the e-350.

True. The i3-2100T is a lot more powerful, and actually has a perfect consumption/performance ratio. But for plain media playback I think the zacate is unbeatable, especially when -as xbit-labs says in its review- "Every Watt Counts".

BTW: In general, if performance is that important on a HTPC, than why not recommend a renderfarm like the ILM studios have?
Because else what exactly is the difference between "3 times around the block and 3 cups of coffee" (slower CPU) versus "3 times around the block and 2 cups of coffee" (faster CPU)? - waiting is still the name of the game!

Llano now brings something new to the table: A lot more powerful IGP, which seems to have quite some effect on (post-processing) image quality as first reviews show. Now its only waiting for some more detailed power-consumption numbers.
Reply
#19
ubuntuf4n Wrote:Wow Nice Chart, where did you grab that one ?

Wondering how long a zacate could survive with the above mentioned temps (max.specif.temp ?)

E-350 Max Temp is of 90º, at least 3 years. And anyway, most of the poor temps is due to crapy thermal paste and application.

RogerS Wrote:Hopefully a vendor will make an Mini-ITX board suited for the A8-3500M processor....

As stated by wsume99 it's a mobile procesor, but as I showed here, Asrock already has some sugar for you...


Image

Some specs from the unfinishhed web: http://www.asrock.com/mb/overview.asp?Mo...ifications
Reply
#20
Zon2020 Wrote:Have you seen a review regarding power consumption for an i3-2310M? I'd be curious how they actually stack up. You can't tell anything from the llinked review.

I thought it was curious that this review would compare a dual core Llano to a quad core i7 Sandy Bridge and then pronounce "Intel's quad-core Core i7 Sandy Bridge processor eats over twice as much power under full load." Yeah, why would that surprise anyone? It's also twice as powerful. Indeed, a couple pages earlier in the review, in discussing performance, they called the Llano "no match for the quad-core Core i7 2820QM."

This is a useless apples-to-oranges comparison. Those two chips don't compete against each other. Why didn't they do a meaningful comparison to the i3-2310M? And just yesterday Intel released several more ultra-low-power mobile Sandy Bridge processors.
Well the A8-3500M is a Quad core CPU. I'd say that a fair comparison is a mobile i5. If you look at the link from the OP you'll see that the Llano beats both the Arrandale and Sandy Bridge i5 mobile chips it was compared to. It wasn't as big a gap when compared to the i7 but the Llano still performed better. From what I've read so far it really comes down to what you need/want. If you want compute power then go Sandy Bridge but if you want the best IGP out there with decent compute power then go Llano. So if you're building a HTPC then most likely the Llano would be the best choice. Now if AMD could just get their linux drivers out of the crapper then they'd dominate the HTPC arena with zacate and Llano.
HTPC: Win 7 Home 64-bit | MB | CPU | GPU | RAM | Case | PSU | Tuner | HDDs: OS, Media | DVD Burner | Remote
Media server: unraid 4.7 | CPU | MB | RAM | Case | PSU | HDDs: Parity-2TB, Data-2x2TB
Reply
#21
SpectreX Wrote:So a ION with a 80mm case fan is cooler than an Zacate with a 40mm fan?
The case fan in the Atom+ION build is a low 900 rpm case fan that's nowhere near the cpu, while the Zacate has a 40mm right on top the heatsink (that's what the mobo came with) AND a 6000rpm 40mm case fan blowing towards it (and it's in a M350 case which well known for being well vented due to the extensive meshing).
If I switch off the 40mm case fan (but still using the 40mm cpu fan), the Zacate idles at 55C (and goes over 80C during high load)!

There is no question that Atom+ION runs much cooler than Zacate, Zacate has other benefits but low temps aren't one of it's strong points, that's for sure.
Reply
#22
-EaS- Wrote:That chart is just a summary from the two datapoints I gave in post .

So let me get this straight. Do I understand that the chart you posted is not from some testing lab but is something YOU put together from separate articles involving separate tests of separate equipment by different people and you presented it as if it was some head-to-head comparison?

On the other hand the xbitlabs actually is a lengthy 12 page report of a head to head analysis running the same tests on both systems. The chart I included is but one of very many from their test results.

Sorry you don't like their conclusions.
Reply
#23
wtf? I do not believe Smile
Reply
#24
Zon2020 Wrote:So let me get this straight. Do I understand that the chart you posted is not from some testing lab but is something YOU put together from separate articles involving separate tests of separate equipment by different people and you presented it as if it was some head-to-head comparison?

On the other hand the xbitlabs actually is a lengthy 12 page report of a head to head analysis running the same tests on both systems. The chart I included is but one of very many from their test results.

Sorry you don't like their conclusions.

Image
Reply
#25
Zon2020 Wrote:So let me get this straight. Do I understand that the chart you posted is not from some testing lab but is something YOU put together from separate articles involving separate tests of separate equipment by different people and you presented it as if it was some head-to-head comparison?

Please, don't worry. It's about as close to a real-world scenario as it *can* get. Different systems and different equipment *is* the real-world. BTW: Everyone is throwing numbers around here, comparing their "numbers" to others "numbers", I even see you doing it.

Besides: Both datapoints come from reviews stripped of bias (since they are single system reviews), use *system* based power consumption numbers (which is where *everyone* deals with, and *is* real-world), use purpose build parts (efficient low power powersupplies etc. - so "as good" as it can get), give a very detailed break-down on the *same* various media load scenarios with regards to power consumption and just use numbers drawn from the socket. *Very* real-world.

On the other hand: The xbit-labs review is exactly the *opposite* of a head-to-head comparison. They're purposely leaving out data that would've made them write a 180 degree turned conclusion. A pre-cooked conclusion, with an article written around it. Knowingly! Very *non* head-to-head. Very questionable.

Next to that: Strange setups (880W power supplies?), completely untransparent on what "1080p playback" means. As hollow as it can get.

Quite a usefull review from xbit-labs don't you agree?

Quote:On the other hand the xbitlabs actually is a lengthy 12 page report of a head to head analysis running the same tests on both systems. The chart I included is but one of very many from their test results.

So if I write a 13 page review it will become the next bible in HTPC land? I'm sure you'll agree that the length of a review doesn't say anything.

Quote:Sorry you don't like their conclusions.

Correction: Sorry you don't like their pre-cooked conclusions.

Agreed!
Reply
#26
-EaS- Wrote:Please, don't worry. It's about as close to a real-world scenario as it *can* get. Different systems and different equipment *is* the real-world. BTW: Everyone is throwing numbers around here, comparing their "numbers" to others "numbers", I even see you doing it.

Besides: Both datapoints come from reviews stripped of bias (since they are single system reviews), use *system* based power consumption numbers (which is where *everyone* deals with, and *is* real-world), use purpose build parts (efficient low power powersupplies etc. - so "as good" as it can get), give a very detailed break-down on the *same* various media load scenarios with regards to power consumption and just use numbers drawn from the socket. *Very* real-world.

On the other hand: The xbit-labs review is exactly the *opposite* of a head-to-head comparison. They're purposely leaving out data that would've made them write a 180 degree turned conclusion. A pre-cooked conclusion, with an article written around it. Knowingly! Very *non* head-to-head. Very questionable.

Next to that: Strange setups (880W power supplies?), completely untransparent on what "1080p playback" means. As hollow as it can get.

Quite a usefull review from xbit-labs don't you agree?



So if I write a 13 page review it will become the next bible in HTPC land? I'm sure you'll agree that the length of a review doesn't say anything.



Correction: Sorry you don't like their pre-cooked conclusions.

Agreed!

Yeah, a reputable review doesn't fit your desired outcome, so it's "biased." Yeah, that must be it.

So I suppose SilentPCReview is "biased" too, since their results also contradict your contrived chart:

http://www.silentpcreview.com/article1167-page5.html
Reply
#27
Zon2020 Wrote:Yeah, a reputable review doesn't fit your desired outcome, so it's "biased." Yeah, that must be it.

So I suppose SilentPCReview is "biased" too, since their results also contradict your contrived chart:

http://www.silentpcreview.com/article1167-page5.html

Ah yes.... don't address the points made, but just spit out random links....

Where does it contradict the points I made?
Reply

Logout Mark Read Team Forum Stats Members Help
AMD LLano idles at 10.2 watts1