AMD Piledriver rumours ... and expert conjecture

[Reynod]

We have had several requests for a sticky on AMD's yet to be released Piledriver architecture ... so here it is.

I want to make a few things clear though.

Post a question relevant to the topic, or information about the topic, or it will be deleted.

Post any negative personal comments about another user ... and they will be deleted.

Post flame baiting comments about the blue, red and green team and they will be deleted.

Enjoy ...

 

[noob2222]

^^ DUN DUN DUN DUN!!!....
cue 'ride of the [strike]haswell[/strike] valkyries' theme...
that aside, how are intel going to destroy motherboards? in desktop platform, low end motherboard brands will suffer a bit (looking at ya, jetway, ecs and biostar et al). even they are trying to improve themselves. mainstream motherboard brands are very likely to remain 'unscathed' from the 'destruction'.
worst case scenario: motherboards become segmented into two main branches - one that just acts as a platform for expansion slots and ports, another that caters to overclockers and enthusiasts. wait, that is already happening...
this could be just as bad if not worse -
http://semiaccurate.com/2012/09/28/asus-makes-an-offer-to-buy-asrock/
...since intel isn't doing the merger, i guess it's less evil and more innovation-friendly.... 😉

might want to actually read the article.

It is critical for motherboard makers that they are able to differentiate their products from rival products and an SoC strategy that collects more features into the CPU will take away choices from the motherboard industry

so motherboards makers can only make a pcb with copper traces from the cpu over to the pci slots. and they are going to make how much off this?

These SoC like designs will first of all come to the mobile computer market, while we expect that the south bridge will still be a separate part on the mother board even for stationary systems running Broadwell. We should be able to expect the full integration of all the segments including desktops no earlier than 2015, in connection with the launching of their Skylake architecture.

If the small vendors are leaving over haswell's power controller, what are the rest going to do when they don't even have a SB to install and charge for doing so?

Oh, ya, the overclockers industry ... http://wccftech.com/intel-develops-revolutionary-overclocking-features-haswell-processors/

looks like Intel wants to control that too, but what is the "OEM Overclocking lock bit". Is that another attempt to lock out bios-level overclocking?

Could be interesting (or really bad) to see what happens in 3 years.

 

[noob2222]

Correction: Steamy on 990FX means no DDR4. You can have a new chipset without changing socket. Intel only just did it a few months ago.

not sure why everyone is worried about ddr4, its not like Intel is going to adopt it next year.

http://www.xbitlabs.com/news/memory/display/20120404214706_Intel_to_Start_DDR4_Usage_with_Server_Platforms_in_2014.html

2014

Heck the final specifications have just now been published, litterally this week. http://www.engadget.com/2012/09/26/final-ddr4-specification/

 

[mayankleoboy1]

http://www.obr-hardware.com/2012/09/exclusive-some-geforce-gtx-780-details.html

GTX780 does not have GK110 chip. OBR has been spot-on wiht nvidia news for some time now.
Maybe it is the GK110 with the Double FP parts removed.....

 

[de5_Roy]

25GB's? I remember my Nvidia 9500GT had 32GB of bandwidth and that card is weaker then a 4670 and a A10 is a little more powerful then a 5670.

ddr3 version of radeon hd 6670 gets 28gb/s. amd a8 3870k igpu can reach 6670-level with high overclock.
http://www.tomshardware.com/reviews/a8-3870k-apu-overclocking-guide,3260.html
radeon hd 6670 with gddr5 can perform even better as it gets more vram bw (gddr5 = around 2x ddr3 at stock). what i'm saying is that if the 6670(ddr3) can perform like this with only 28 gb/s and better with gddr5, llano and trinity's igpus can scale with higher memory bw too, if they can reach at least 25 gb/s.

might want to actually read the article.

It is critical for motherboard makers that they are able to differentiate their products from rival products and an SoC strategy that collects more features into the CPU will take away choices from the motherboard industry

so motherboards makers can only make a pcb with copper traces from the cpu over to the pci slots. and they are going to make how much off this?

i read it alright. don't see an issue with motherboard vendors. most of entry level motherboards are same as others. very few have standout features like the ones from msi with pcie 3.00 slot or asrocks xfast lan/usb.
smaller vendors sell poorly designed bios, cheapo components, worse reliability, cut corners in designing even now. they already sell "pcb with copper traces from the cpu over to the pci slots". most of them don't even offer good power delivery system. why not improve those before freaking out over vrm integration. since both intel and amd will be pushing soc chips by 2015, motherboard companies will have to comply and find a way to work with them. i guess we'll see in 3 years.

These SoC like designs will first of all come to the mobile computer market, while we expect that the south bridge will still be a separate part on the mother board even for stationary systems running Broadwell. We should be able to expect the full integration of all the segments including desktops no earlier than 2015, in connection with the launching of their Skylake architecture.

If the small vendors are leaving over haswell's power controller, what are the rest going to do when they don't even have a SB to install and charge for doing so?

mergers like asus+asrock will happen and desktop motherboard market will shrink further. DIY segment will become 'high end' (in reality, overpriced with less choices) segment. although it is too early to tell.
another effect is that motherboard vendors' eagerness to stand apart from rivals can make adoption of newer tech like thunderbolt or lightning bolt faster.

Oh, ya, the overclockers industry ... http://wccftech.com/intel-develops-revolutionary-overclocking-features-haswell-processors/

looks like Intel wants to control that too, but what is the "OEM Overclocking lock bit". Is that another attempt to lock out bios-level overclocking?

Could be interesting (or really bad) to see what happens in 3 years.

if oem overclocking lock bit refers to locking overclockable cpus like the i5 on oem machines... it's not like overclocking oem machines is easy/possible even now.
intel integrating vrm before amd could benefit amd's desktop platform. when intel jacked up their quadcore prices over $180, amd approached sub $160 level with cheap quadcores. their higher priced cpus got price cuts over time and settled in sub $190 level. amd has always been friendlier towards overclockers and enthusiasts than intel. motherboard companies don't want single cpu vendor to monopolize, so they will gravitate towards amd more provided that amd doesn't take the same path as intel.

 

[sarinaide]

Some time next year AMD are due to announce architectures for 2015 and 2016, which basically puts an end to AMD are not competing fiasco.

 

[the3holepuncher]

So, not to steer the conversation, but back to the topic of Piledriver's release: Trinity is supposed to be out on tuesday, are they releasing piledriver at the same time, or is it the 23rd as previously thought?

It's just weird to me, mind you this is my first chip release that I've followed anywhere near release date, but I figured it would be similar to the release of a new car, or a summer blockbuster, where it's everywhere, in everything, etc, for months before release.

Is this tactic of keeping quiet normal for CPU releases, or AMD? Is it an attempt to keep Intel from guessing where to make the price cut in their product line to pull sales away? Or is October not truly confirmed, and it has the possiblity of becoming a "Q1 2013" release? I kinda need to know, because I have everything for my first rig, except for the chip, because I found out it was supposed to be getting released within a week or two(this was maybe a week ago)

 

[cgner]

Bulldozer was everywhere, that didnt help it be a good chip. This time they dont waste time on marketing.

 

[the3holepuncher]

then what they need is a )(*&(*ing countdown timer on their main website, to release. simple, effective, no nonsense. atleast then we know when to be ready to put in our orders for those that are already committed to buying one.

 

[blackkstar]

Didn't AMD release a CPU that worked in AM2 and AM3 sockets before? And if PCIE 3.0 is backwards compatible with PCIe 2.0, couldn't a PCIe 3.0 controller on steamroller work fine in an AM3 board? Or do they really and honestly need more pins for that?

 

[obsama1]

To get PCIe 3, all they need is a new chipset. No new socket is necessary.

 

[fazers_on_stun]

PC sales are only slowing in the west, emerging markets like china and india have huge demand for PCs. Which just so happen to be where AMD hopes to sell a lot of the APUs.

IIRC AMD blamed slowing sales in China for their Q2 earnings performing 10% under previous estimates..

 

[esrever]

AMD didn't really put marketing up until bulldozer launched. There were some people saying it would be great and some saying it would be bad but AMD didn't officially have much marketing out until it all was out in public and everyone notice it underperformed. There were only a few slides about bulldozer before it actually launched.

 

[cgner]

and overhyping didnt help

 

[viridiancrystal]

And now there is almost no hype around Piledriver. If it is a good improvement then it may surprise quite a few people.

 

[king smp]

a solid %15 improvement in IPC would be awesome

 

[mayankleoboy1]

a solid %15 improvement in IPC would be awesome

would probably happen with SteamRoller.

 

[king smp]

I have a 1100T at 4ghz and dont see PD as being a worthy upgrade so far
hopefully Steamroller will be

 

[jdwii]

I have a 1100T at 4ghz and dont see PD as being a worthy upgrade so far
hopefully Steamroller will be

i'd have to agree.

 

[gamerk316]

Everything I'm seeing with PD points to it being more of a power lessening revision. SR looks like the big performance jump.

Of course, by then, Broadwell will be out...

 

[cgner]

I have a 1100T at 4ghz and dont see PD as being a worthy upgrade so far
hopefully Steamroller will be

sayem

 

[de5_Roy]

"AMD Trinity A10-5800K vs Intel Ivy Bridge i5-3470 - Discrete GPU Gaming Performance"
http://vr-zone.com/articles/amd-trinity-a10-5800k-vs-intel-ivy-bridge-i5-3470--discrete-gpu-gaming-performance/17272.html

 

[jdwii]

"AMD Trinity A10-5800K vs Intel Ivy Bridge i5-3470 - Discrete GPU Gaming Performance"
http://vr-zone.com/articles/amd-trinity-a10-5800k-vs-intel-ivy-bridge-i5-3470--discrete-gpu-gaming-performance/17272.html

Nice at times Amd even had the intel chip beat but like the comments i want to see it benchmarked fully at the I3. If Trinity is priced at the I3 its going to be quite a deal.

 

[jdwii]

http://www.amazon.com/gp/product/B0095VP8D4?ie=UTF8&tag=cw01-20&linkCode=as2&camp=1789&creative=9325&creativeASIN=B0095VP8D4

Not bad for 135$!

 

[the3holepuncher]

I'm upgrading from a hand-me-down hp with a phenom 64 9550. I'm holding out for the piledriver, since it can't be any worse than the bulldozer, and might be much better. I've already got all my components ordered, and it's the last piece of my rig, which is why I'm soo impatient. And even if it's just a bulldozer with lower power consumption, I've used an 8150 HP phoenix, and it's actually quite nice, I just want the newer chip.

 

[sarinaide]

The problem with the 1100T and 1090 was that they were good chips, just to late and overlooked due to expectations on BD. That is not going to change anytime soon and the Thuban will still be a good CPU regardless. But like all, time will necessitate its change, but it will remain a novalty, it will remain the one that wasn't so bad after all.