AM3+ Crystal Ball

[Haywire]

When do you think the AM3+ mobos will start to be available? Spring... Summer... You guys seem to have a finger on that kind of thing.

Ian

[BaronMatrix]

When do you think the AM3+ mobos will start to be available? Spring... Summer... You guys seem to have a finger on that kind of thing.

Ian

Based on the availability of the AM2+ mobos, they should show up at the same time as the chips are launched.

[dm7000s]

When do you think the AM3+ mobos will start to be available? Spring... Summer... You guys seem to have a finger on that kind of thing.

Ian

Well, I am curious as well, so I did some thorough digging around. And it seems that not only the Bulldozer desktop and Llano roadmaps leaked, but also the Chipset roadmaps. From some very vague site (so I uploaded the pics again, btw the 9-series chipset series are the AM3+ launch chipset series):








The "detailed" (sorry about the size) specs of the chipsets:







So UNLESS the "NEW" stamp in the first image = Q1 launch, the new AM3+ chipset series all seem to become available at the start of Q2 (april).

[Haywire]

Thanks, I guess I can stop worrying about it till April.

Ian

[AussieFX]

You will probably find early boards before they're officially released. I ( and about half a dozen others) had an AM2+ board around 3 months before the "official" release but be warned there is nothing to compare to and it's easy to get burnt. My board was a cracker on AM2 (6400+) but was a bit of a dud when it came to K10 cpu's.

The roadmaps dm7000 posted look about right to me.

[Viral]

Strange how the NB specs for 950FX and 850FX are exactly the same yet 950 uses 1.6w more power.

The only difference I can spot on anything is the SB950 has twice as many GPP PCIe links.

Not sure how it was possible to change less than the transition from 750 to 850 but it looks like they'll manage to do it

[GummiRaccoon]

Strange how the NB specs for 950FX and 850FX are exactly the same yet 950 uses 1.6w more power.

The only difference I can spot on anything is the SB950 has twice as many GPP PCIe links.

Not sure how it was possible to change less than the transition from 750 to 850 but it looks like they'll manage to do it

990FX and 890FX paired with SB950 and SB850 respectively.

And if you see, the SB950 and SB850 are both 6W so that can't be the difference of 1.6W.

[Smartidiot89]

Strange how the NB specs for 950FX and 850FX are exactly the same yet 950 uses 1.6w more power.

The only difference I can spot on anything is the SB950 has twice as many GPP PCIe links.

Not sure how it was possible to change less than the transition from 750 to 850 but it looks like they'll manage to do it

990FX and 890FX paired with SB950 and SB850 respectively.

And if you see, the SB950 and SB850 are both 6W so that can't be the difference of 1.6W.

I am sure he's speaking of 890FX and 990FX. The later uses 19.6W(1.6W more) but still doesn't show any apparent advantages.

[Viral]

Yes, my comments on the SB were seperate from my questioning of where the extra power use comes from in the NB. SB commentary was merely because that was the only place I noticed any difference at all between the two platforms besides the socket.

I really did expect PCIe 3.0, but maybe it's just a bit too early for that... bad timing none the less.

[Death9000]

there would be no harm in them releasing AM3+ early imo as they are compatible with AM3 cpus.

[MKruer]

I am sort of hoping that AMD will mandate that all AM3+ system need to use UEFI over the BIOS. but i think that is a pipedream. At least that way the boot seems faster.

[Smartidiot89]

I am sort of hoping that AMD will mandate that all AM3+ system need to use UEFI over the BIOS. but i think that is a pipedream. At least that way the boot seems faster.

We can hope, all Intel Sandy-Bridge boards will have UEFI instead of good old BIOS.

[dm7000s]

Some more slides have appeared:

[Smartidiot89]

Some more slides have appeared:

You're welcome

[Pietro sk]

http://en.inpai.com.cn/doc/enshowcont.a ... ageid=7695

Based on these data, it is easy to infer the Core i7 which based on 45nm process, when core frequency up to 3.0GHz or more, there must be substantial increase in power consumption, while 32nm Sandy Bridge technology the increase in power consumption is little. So we can say that Sandy Bridge version Core i7 power control have a lot of improvement. This also is why intel did to up frequency too much after Core i7 reach 3.0GHz; Because if you want higher core frequency, you will need the new process to control the power at reasonable range.
This also shows that if other Sandy Bridge processors, which has lower frequency, the effect of power consumption control is not particularly obvious. As for the dual-core Sandy Bridge Core i3,the power differences will naturally smaller when compare with 32nm Core i3.

Hmm, old AMD 45nm vs brand new 32nm -- new chips must clock very high... (BE madness, people get ready )
("no problemo - 4GHz thuban" -VS- 32nm BE "octopus" ) ===> 5GHz on water (?) I think yes
2,5GHz 16c should "fit" HE suffix , IMHO

Maybe i´m too much positive, but i have good feeling about this

[FlanK3r]

Ihave gold Thuban, most people dont believe how its possible...but 4300 MHz full stable aircooled, better than my 2 chips x6 1100T and next others AMD chips. Most benchmarks can at 4400 MHz or more up to 4670 MHz for superpi 1M . This is not bad, right?:)
And of course, architecture of Bulldozer is friendly for higher freqency think. Its not secret, chips Zambezi will be 3.5 GHz+ on stock and info about Turbocore 2.0 +500 MHz on all loading cores! So, I think (speculating), Zambezi OC will be ussually about 4.5 GHz+ on air

[dm7000s]

First (decent) spot of an AM3+ motherboard (see also Pietro sk's post), the MSI Big Bang Conqueror (MS-7640 ver: 2.0):



The socket has printed AM3b on it (clickable):

[Pietro sk]

First (decent) spot of an AM3+ motherboard (see also Pietro sk's post),
the MSI Big Bang Conqueror (MS-7640 ver: 2.0):

big bang conqueror ?
thanks for the hires photos, it corrects previous (wrong) statement <2pin difference>
i see there AM3/AM3+ 3pin difference

[dm7000s]

big bang conqueror ?
thanks for the hires photos, it corrects previous (wrong) statement <2pin difference>
i see there AM3/AM3+ 3pin difference

According to the specifications socket AM3b seems to have 1-pin (hole) more (which of course does not say anything about how the pins are positioned):





See this pdf.

[FlanK3r]

nice!

[Aaronage]

Video of the AM3+ MSI Big Bang from TweakTown:
http://www.youtube.com/watch?v=FHwVSW64sn8

Nice!

[Viral]

Good looking board, seems MSI have followed ASUS wrt HS design. The colour scheme is nicer on this board though! Hopefully it is a stronger overclocker than their 890FX.

[Boundless]

A casual search (including reading some AMD spec
docs) didn't quickly answer these two purely
curiosity questions:

1. What frame buffer does a Fusion GPU have access to?
   Pretty clearly (from the BIOS spec), it has access to
   main RAM [UMA]. But does Fusion have any provision
   sideport or on-die RAM (other than cached UMA)?
2. How does the graphics signal get off the chip?
   I'd be surprised if there are pins for LVDS or TMDS
   (much less analog VGA). I'd tend to guess some
   external display driver chips are still needed.
   Are they just handed a base address for the current
   raster in RAM?

_______
So where do those analog VGA ports get their frames?

[tryathlon]

Sorry if I missed it on this thread, and because my AMD loyalty is just now coming in to focus, but what is/was the exact release date?

[brutis]

Sorry if I missed it on this thread, and because my AMD loyalty is just now coming in to focus, but what is/was the exact release date?

They are already out there http://www.newegg.com/Product/ProductLi ... MD%20990FX