AMD 3870k crashes

[vagxtr]

You are completely and unequivocally wrong on this one. Llano benefits very very heavily from each and every extra mhz in ram.

OMG. Every CPU does benefit from every MHz. But you're just overhyping things. Its nice that you point out how CPU benefit 10% or more (while not in fact) just by raising its IMC to work with 15% higher memory setting.

Do you work for AMD marketing department?

Only real benefit from this kind of higher speed ddr3 bus came for GPU but Llano is already doing that reordering/feeding job for GPU pretty lame. Not even mentioning Llanos VLIW5 architecture (AMD Sumo) used in its GPUs ain't nowhere that great in decent utilization of vast shader power it has. But aside of that, its gives up huge gains in RAW GFLOPS MARKETING NUMBERS.

 

[cdawall]

OMG. Every CPU does benefit from every MHz. But you're just overhyping things. Its nice that you point out how CPU benefit 10% or more (while not in fact) just by raising its IMC to work with 15% higher memory setting.

Do you work for AMD marketing department?

Only real benefit from this kind of higher speed ddr3 bus came for GPU but Llano is already doing that reordering/feeding job for GPU pretty lame. Not even mentioning Llanos VLIW5 architecture (AMD Sumo) used in its GPUs ain't nowhere that great in decent utilization of vast shader power it has. But aside of that, its gives up huge gains in RAW GFLOPS MARKETING NUMBERS.

Nice to see you are just throwing random numbers out with no idea how it works to begin with.

 

[vagxtr]

Nice to see you are just throwing random numbers out with no idea how it works to begin with.

So youre not just flaming while give up much of facts regarding an issue

 

[cdawall]

So youre not just flaming while give up much of facts regarding an issue

Sure can there buddy same 3.6ghz for both overclocked tests using 1866 and 21xx memory

source

here is another

source

Is that enough "proof" for you or should I pull a couple more of the dozens upon dozens of reviews that support memory overclocking is immensely beneficial to Llano?

 

[Aquinus]

OMG. Every CPU does benefit from every MHz. But you're just overhyping things. Its nice that you point out how CPU benefit 10% or more (while not in fact) just by raising its IMC to work with 15% higher memory setting.

Do you work for AMD marketing department?

Only real benefit from this kind of higher speed ddr3 bus came for GPU but Llano is already doing that reordering/feeding job for GPU pretty lame. Not even mentioning Llanos VLIW5 architecture (AMD Sumo) used in its GPUs ain't nowhere that great in decent utilization of vast shader power it has. But aside of that, its gives up huge gains in RAW GFLOPS MARKETING NUMBERS.

I'm not sure if you're trolling or if you actually think that what you're saying is true.

The APUs likes the extra memory bandwidth because of the iGPU, it is as simple as that. The CPU cores will hardly benefit from it. Reviews have shown that the iGPU really does like faster memory and that the gains are significant enough to consider running it. Your post is kind of confusing, before posting I tend to re-read and edit my posts before I reply. I suggest that you do as well to make sure you're conveying the point that you're trying to make, because you post is a little confusing because of the language you used.

You don't need to work for AMD's marketing department to do a little bit of research to find out how it really impacts Llano. Do you have an APU? What experience do you have with an APU that qualifies you to make this claim? No hardware? How about some sources?

How about this: http://www.anandtech.com/show/4476/amd-a83850-review/4

 

[cdawall]

I'm not sure if you're trolling or if you actually think that what you're saying is true.

The APUs likes the extra memory bandwidth because of the iGPU, it is as simple as that. The CPU cores will hardly benefit from it. Reviews have shown that the iGPU really does like faster memory and that the gains are significant enough to consider running it. Your post is kind of confusing, before posting I tend to re-read and edit my posts before I reply. I suggest that you do as well to make sure you're conveying the point that you're trying to make, because you post is a little confusing because of the language you used.

You don't need to work for AMD's marketing department to do a little bit of research to find out how it really impacts Llano. Do you have an APU? What experience do you have with an APU that qualifies you to make this claim? No hardware? How about some sources?

How about this: http://www.anandtech.com/show/4476/amd-a83850-review/4

Just a little quote from your source

Across our seven titles we found that on average DDR3-1600 resulted in a 12.6% increase in performance over DDR3-1333 at 1024 x 768. Moving up to higher resolutions only increased the advantage by under 2%. Using DDR3-1866 showed around a 20% increase in performance over DDR3-1333.

 

[acerace]

@vagxtr What now?

 

[vagxtr]

The APUs likes the extra memory bandwidth because of the iGPU, it is as simple as that. The CPU cores will hardly benefit from it.

That was my conclusion that CPU performance wont benefit from measly 15% in ram speed increase (1866 vs 1600) but GPU in Llano is in the first place botched VLIW5 that was and i wouldn't comment on that Corsair memory scaling graphs. Obviously thy highlight that on as they could benefit from 120%-150% increased memory price from 2133MHz modules of same size over 1600MHz ones.

I only said it ain't worth it to insist on 1866MHz running ram at 1.65V (Corsair Venganece 1866CL9) instead just sticking with its 1600CL9 speeds or it could go d-tour (original poster that has issues) and try to increase CL10 or even CL11 for working with Llano at 1866Mhz to get its stability instead of rising voltage to 1.65V

Just a little quote from your source

Hm those graphs show how much GAMING (GPU) improvement with using 1866 over 1600MHz ram? In best case scenario its 10% (HAWX: 32 vs. 29fps). Of course this gap will be widened at 1920x1080, but this fully featured A8 Llano GPU doesnt have that much power nor Trinity 6CUs for that matter does. It's a budget APU and for gaming on six year ago newish 1680x1050 resolution is more than good option

I wont comment on those Corsair investigation benchmarks. After all they're in memory selling business While their graphs are using generic say nothing percentage wise (from what standpoint??) presentation

Or to rephrase myself with editorial notes

Every CPU does benefit from every memory MHz. But you're just overhyping things. Its nice that you point out how CPU benefit 10% or more (while not in fact) just by raising its IMC to work with 15% higher memory setting.

Its not that CPU itself will benefit for huge amount of its performance while 1600MHz already provides decent enough memory bandwidth. Everything else is beneficial just for AMDs marketing, which is poor enough.

I know man here invested in ""superior 1866MHz memory"" that should work well at those declared speeds. But at what voltage? From my POV there's no reasonable point to go beyond 1600Mhz for Llano based APUs while Trinity might benefit from it but hopefully there should be some ddr3-1866 modules nowadays that are capable of running at 1.35 or 1.5V. Cmon its not 2010 that we should look with exaggeration 4GB ddr3-1600 modules that are capable to tun at 1.35V.

In my not so humble opinion is just poor that TWO 4GB 1866MHz modules should run at anything higher than 1.50V-1.55V. Then again i'm not using hios setup and he might just overvolting its memory "just to be safe" it aint failing "because of low voltage"

 

[cdawall]

That was my conclusion that CPU performance wont benefit from measly 15% in ram speed increase (1866 vs 1600) but GPU in Llano is in the first place botched VLIW5 that was and i wouldn't comment on that Corsair memory scaling graphs. Obviously thy highlight that on as they could benefit from 120%-150% increased memory price from 2133MHz modules of same size over 1600MHz ones.

I only said it ain't worth it to insist on 1866MHz running ram at 1.65V (Corsair Venganece 1866CL9) instead just sticking with its 1600CL9 speeds or it could go d-tour (original poster that has issues) and try to increase CL10 or even CL11 for working with Llano at 1866Mhz to get its stability instead of rising voltage to 1.65V

What exactly is 1.65v hurting? He purchased ram not designed for the board not on the QVL so it requires a smidge more voltage to get running.

Hm those graphs show how much GAMING (GPU) improvement with using 1866 over 1600MHz ram? In best case scenario its 10% (HAWX: 32 vs. 29fps). Of course this gap will be widened at 1920x1080, but this fully featured A8 Llano GPU doesnt have that much power nor Trinity 6CUs for that matter does. It's a budget APU and for gaming on six year ago newish 1680x1050 resolution is more than good option

Those don't look like gaming benchmarks click the link instead of posting away.

Sure can there buddy same 3.6ghz for both overclocked tests using 1866 and 21xx memory

http://media.bestofmicro.com/0/A/347338/original/sandra arith.png

http://media.bestofmicro.com/0/C/347340/original/sandra multimedia.png

http://media.bestofmicro.com/0/B/347339/original/sandra mem.png

http://img.techpowerup.org/121104/Capture010.jpg

source

Or to rephrase myself with editorial notes


Its not that CPU itself will benefit for huge amount of its performance while 1600MHz already provides decent enough memory bandwidth. Everything else is beneficial just for AMDs marketing, which is poor enough.

I know man here invested in ""superior 1866MHz memory"" that should work well at those declared speeds. But at what voltage? From my POV there's no reasonable point to go beyond 1600Mhz for Llano based APUs while Trinity might benefit from it but hopefully there should be some ddr3-1866 modules nowadays that are capable of running at 1.35 or 1.5V. Cmon its not 2010 that we should look with exaggeration 4GB ddr3-1600 modules that are capable to tun at 1.35V.

In my not so humble opinion is just poor that TWO 4GB 1866MHz modules should run at anything higher than 1.50V-1.55V. Then again i'm not using hios setup and he might just overvolting its memory "just to be safe" it aint failing "because of low voltage"

Who cares if it takes more voltage you aren't hurting anything? 1.65v is 10% over 1.5v thats within JEDEC spec.

 

[agent00skid]

The only reason I went into the memory thing, is that I HAVE a Llano setup with the same pair of memory sticks, and they do run at XMP profile at 1.5v. I even have it on a Gigabyte board.

And 1866 prices aren't much above 1600 these days.

 

[Aquinus]

Who cares if it takes more voltage you aren't hurting anything? 1.65v is 10% over 1.5v thats within JEDEC spec.

+1: Couldn't have said that better myself.
Believe it or not, my memory is 1.65v stock using the XMP profile for 2133 (9-11-10-28) and I've felt the memory chips (overclocked to 2333 (10-11-10-30)) and they're lukewarm. Additionally G.Skill said that they chips are supposed to take 1.65v. The only time you have to worry about the DRAM voltage is more on the IMC on the processor and AMD' can handle even higher voltages (albeit not too much more) where Intel's strongly recommends you not exceed 1.65v (since you could fry an INTELIMC... awww.)

But we're talking Llano here, not an Intel CPU so I would say run that puppy as fast as it will go and your Llano CPU will be grateful for the opportunity to run a little faster, but cdawall is 100% correct. 1.65v is within JEDEC specification (granted the top-end,) but within nonetheless.