Printable View
This will be useful as info for Intel enthusiast.
This X79 chipset could be capable to support quad-channel DDR3 memory frequency up to 2666 MGHz and above.:eek:
This chipset is capable to use SATA 6 Gb/s and SAS at the same rate.:eek:
Codename Patsburg is for confimation purposes.
Thanks.
:up::up::up:
Yes, Patsburg is the codename for the chipset(s) based on X79, Waimea Bay platform. SATA6 and SAS are indeed confirmed on as many as 10 ports, Intel went all the way this time around.
Still, lack of USB3 is a major step back in my opinion. I can live without it for now, but this should last for a few years still, like X58 did. Also, PCIe lanes instead of PCIe 3.0, even if they were crippled like in our chipsets, just half the speed. But at least they should be PCIe 3.0 to be future proof.
Anyway, let's wait and see.
Regarding mem, I very much doubt those 2666MHz, it's spec'ed at 1600MHz if I'mnot mistaken, maybe some mobs will go as high as 2400MHz but 2666? Seriously! And if you consider that 4 sticks will load the controller heavilly, forget it.
Dear Manel:
This wonderful MB from Gygabite has features that i think their unconclused cause this is a UD3 and the generation for this MB goes to UD9.
Attachment 116617
I hope as the enthusiats for this mobo are in the correct way cause this is a Extreme edition with support for USB 3.0 too.
About mem at quad core settings and the rate 2666 MGHz and above is neccesary for a change at the extreme Dimension.
This Gygabite UD3 shows that and Intel x79 last assembly at the final release will have all the new tech needed to be a Extreme Edition as Intel knows.
Good for us.:clap::clap::clap:
This mobo was shown on Computex, one of the 12. Not that I don't think that Gigabyte knows how to make boards but I wouldn't buy them for myself.
They have finally understood that Extreme boards need to be black! Still too flashy for me though. Nothing designed on that side of the world will take my money.
What I'm really eager to see is Intel's own take on the X79, the Siler and Thorsby.
Anyway, as I said, if the specs are correct and final, I guess I'll pass this one too. Skip to the next iteration. Anyway, time for tweaking is not much right now.
Thanks Manel for your quick n' kindly reply.
You have for reference this:
http://lab501.ro/wp-content/uploads/2011/05/x79a.jpg
Intel Corp. not dissapoits us i hope.
Siler n' thorsby will be in the market soon, will see.:up::up::up:
Well, in the meanwhile i remain.:)
Still, I'll believe 2666 when I see it. Anyway, quad channel even at 2133 (which is not uncommon today) will give you enough bandwidth for any application today, no proc will starve for data.
Next year DDR4 will raise the bar in terms of speed, new platform needed again :-)
As for USB3, anyone can slap a NEC/Renesas controller in a mobo. I was hoping Intel would support it in the chipset, I don't think that would be hard since IB will come out right after and the companion chipsets (7 series, Z77/5, H77, Q77/5, B75) already support it. And those PCIe 2.0 lanes should have been PCIe 3.0 already. And also PCI should be dead already, why go back if SNB doesn't already. Those are the only gripes I have with X79.
On the other hand, SAS is cool although hardly used on the desktop machines.
Have you guys also seen the news on Larrabee? Or should I say the Knight's Bridge project. It's only oriented fo HPC now but it seems to go well. Intel has already released some pics of a PCIe boad it will give CUDA a run for it's money it seems. No GPU based on it expected though. Intel engineers don't exactly excel on software and driver development.
On another note, for the first time BITS was first released on Intel than on their own page. Well, not exactly. Just noticed that although the main page doesn't show it, v438 is already available there on the download page.
Yes, those were the boards but I want Siler!!!
:-)
For the Efficient Performance Server market, the Platform is called Romley-EP, Patsburg+Sandy Bridge-EP, for 1 and 2 socket. For 4 sockets there is the Romley-EP 4S.
Wow, PCI SIG has plans to release (specs in mid 2013) a cabled version of PCIe (4 lanes PCIe 3.0, 32Gbps), to replace Thunderbolt (over 3x faster today, TB will be 50Gbps in 2015), USB 3.0 and eSATA. The funny thing is PCIe is the underlying structure (bus) for these controllers. The idea is to eliminate the protocol overhead of these protocols and make PCIe itself an interconnect. It can even supply 20W of power for any devices. The connector will be compact and flat to fit even in notebooks and tablets.
Why didn't they think of it before?
Check this out:
http://www.tomshardware.com/reviews/...oard,2963.html
Intel's own Hawthorn Farm, where our boards are born and take shape.
Unfortunatelly, they're ultimatelly made we know where, don't we?
Some info:
The next phase, he says, involves an X79 enthusiast board for the LGA 2011-based interface, which will support Sandy Bridge-E-based processors with four memory channels. From the ground-up, he wants this board designed to be LN2-friendly. It’s to sport the highest-rated POSCAPs (a solid electrolytic capacitor with a sintered tantalum anode and semiconductor cathode) with high thermal tolerances and a more open processor interface area to accommodate large pots.
The GPGPU programming technologies are here cause Intel is pushing ahead this tech.
Attachment 116650
Attachment 116651
Attachment 116652
This is awesome, It's Intel Manel.:clap::clap::clap:
I think as reported that Intel is hiding for strategy purposes all the material that Intel siler could have finally.
d'ont worry about USB3.0's, those will be included in the final version.
:up::up::up:
Awesome!
Would this be a Tesla competitor?
It does look somewhat Geforce-ish :p
I cannot wait for X79!
John
Paty, those are cool pics, hadn't seen those yet, only of the die and the board like your first pic.
Yes, this is a Tesla competitor and Intel has already sent out I believe 130 of these babies to developers. The major argument for Intel is that it's still x86, so any app already existing will be easy to port. While for Tesla you need to deep into CUDA. I believe CUDA is very C'ish (haven't digged into it however) which is piece of cake for any programmer, but still you need to develop from ground.
I have a great respect for Intel, and I'll always buy Intel instead of anything else if I can, but I'd get rid of x86 in a flash. We're using decades old tech (with some updates of course) for the sake of compatibility and I think it's holding everything back.
I'll take your word for it regarding Siler, but all info out points in the other direction. Unless you're referring to external controllers, which for me doesn't make sense anyway. It's not that Intel doesn't have the means to do it, I know they are always the last to adopt new stuff but that's the way for big outfits (I work for one), but it kinda disappoints.
Imagine the power inside that box in the last pic, 8 Knight's Bridge boards. Each contains 50 simple x86 cores based off P54. This is MIC (Many Integrated Core) tech from Intel and it might be the future.
I couldn't find info regarding the interface, if it's already PCIe 3.0 or 2.0, I believe still the latter for now.
There are quite a few articles out there abut this (read some a while back, don't know if anything new has come up) and the performance attainable in the near future.
This was quite a chalenge in design, I remember the cores are connected by a low latency ring bus.
I wonder if they used RISC cores (maybe Itanium based) instead. Awesome. But that would make compatibility with x86 (and the whole point of using these cores for easy porting) difficult or impossible. I remember the problems of running existing soft on Itanium, low performance due to emulation made it be dumped by some manufacturers.
Intel RST 10.6.0.1002 and audio driver updates.
These are pic's that reveals the socket for 2011 Proc. really as i think it must be.
It's not phases or capacitor's tech more important than the tech embebbed with our Intel Proc. 2011, that i see with Graphic availability if you want into this "Piece of Art".
See what this will be the socket but the Proc.2011 will be more powerful than the Bul**** from the other brand.
Attachment 116755
Attachment 116756
:up::up::up:
Paty
Those are indeed Socket 2011 pics, from a Gigabyte board.
I guess we'll have to get used to the new retention mechanism.
Intel engineers go a different route than the taiwanese makers, regarding phase count. What they do is in fact engineering, not marketing bull**it using trillions of power phases to hide the design deficiencies.
They get the same or better power delivery behaviour then other designs, better thermals.
Also, they'll start using different (better) components.
That will eventually come with a $$$ premium.
Come out of hiding Siler. Intel is keeping it under wraps quite well, I believe they really want to dazzle everyone with it.
nVidia taped out Kepler, the 28nm successor to 40nm Fermi. It seems we might see GF6xx by the end of the year already. AMD also seems to be on track with their 28nm offering.
Mass avail maybe next year only.
And the war begins... again!
Let's see what's new. MS is quiet regarding DX, 11.1 or 12 should be in the works but no info that I know of recently.
Too sad after reading the articles below about a delay for the Sandy bridge - E: :(
http://www.tcmagazine.com/tcm/news/h...k-january-2012
http://vr-zone.com/articles/sandy-br...012/12816.html
I still think this delay is for a good review from the Intel engineering Group to deliver better products.
:up::up::up:
Yes, disappointing indeed. Still,mayb it's for a good reason, and they're doing some last minute tweaking! But I guess it's just a business move though. They have nothing to fear from the other camp so why rush it?!
Maybe they're updating the chipset with all the goodies that are missing at the moment, but I don't believe they'll go for significant silicon changes right now.
Let's hope they deliver a great platform for the next couple of years.
Intel, please ditch PCI already, save the silicon for USB 3.0 and PCIe 3.0 instead. By then, I'll go for it!!!
Here we can see some diferences with IvY bridge and Sandy Bridge.
Attachment 117147Attachment 117148
Attachment 117149Attachment 117150
Attachment 117151Attachment 117152
Well, it appears that Sandy bridge with 2 real cores makes the difference.
P.S.: The voltage could be a new advantage for these Pro.'s.
:clap::clap::clap:
friend , you will see this pic with 3 Proc. 2011.
22nm ivy bridge,32nm sandy bridge i5-2300,32nm sandy bridge pentium g620.
Attachment 117153
Well, for now we have not too much to share as info or previous info.
:up::up::up:
Hi
I assume those UEFI shots are from an Ivy Bridge CPU, looking at the cache sizes.
CPU-Z seems to not detect yet the IB CPU, but I guess this one is not in the same league as the also tested i5-2400, mostly lower scores, which should not happen with equivalent procs. It's only a dual core CPU vs the quad SB.
Also, GFX are different, IB seems to use integrated graphs and SB is using GTX580, which could explain the MAXON scores.
Still, very promising indeed.
If you get a shot of an 2011 SB-E, that would be cool!
My eys might be tricking me, but I see a difference of 2 pins between IB and SB and not only one, as socket 1155 and socket 1156 suggest. Is it just me?
Forget it, those in the middle might not even be pins, and then I see no difference. Whatever?!