hope so, i whant to make a new project with sr-X

Halfdead14
Yes, but I would hope anyone spending a large sum of money on these knows the CPUs need dual QPI links for intercommunication.

In hindsight, statistically, you're right.  But we can't just assume people with money are knowledgable enough; at least some ask questions instead of assuming it is in the case and confused/frustrated it isn't working. 
 
Especially when SR-2 first appeared, many members and review sites assumed that 1366 = i7 900 series therefore kept refering to it as a dual X58, assuming i7's would pair up.  Many criticised the choice for Xeons not realizing i5520 is a server chipset, while the enthusiast point-of-view remained typically narrow-minded.  In the two years I've been following this subforums, approximately once per three weeks we'll see a thread where someone asks why or when i7 support will come to SR-2.  That is why news of SR-X supporting SB-E's in single user is not something I'd push, at least JacobF isn't making it official.

What is the point in posting the release on your web site, and not offering it for sale? 

The CPU's are avalible at newegg Mik3yD.
 
I am not sure, but wasn't this board avalible in small quantity when they first posted it on the website, but now it is sold out.

lehpron

Halfdead14
Yes, but I would hope anyone spending a large sum of money on these knows the CPUs need dual QPI links for intercommunication.

In hindsight, statistically, you're right.  But we can't just assume people with money are knowledgable enough; at least some ask questions instead of assuming it is in the case and confused/frustrated it isn't working. 

Especially when SR-2 first appeared, many members and review sites assumed that 1366 = i7 900 series therefore kept refering to it as a dual X58, assuming i7's would pair up.  Many criticised the choice for Xeons not realizing i5520 is a server chipset, while the enthusiast point-of-view remained typically narrow-minded.  In the two years I've been following this subforums, approximately once per three weeks we'll see a thread where someone asks why or when i7 support will come to SR-2.  That is why news of SR-X supporting SB-E's in single user is not something I'd push, at least JacobF isn't making it official.

was there a piriod of time where a single i7 900 worked in an sr-2? I could be wrong, but I don't think so. therefore a i7 in the sr-x IS significant.

yes it would, but its no worth.
you will same performance.

I don't need the overclocking, but I am also disappointed by the PCIe wiring. If two Xeons have 80 lanes, I want at least 64 of them available for the PCIe slots ... This is definitely a criterion for exclusion. I will have an Intel Crown Pass in a few days for testing, let's see how this is wired.

Brocasta

woodstock23
I don't need the overclocking, but I am also disappointed by the PCIe wiring. If two Xeons have 80 lanes, I want at least 64 of them available for the PCIe slots ... This is definitely a criterion for exclusion. I will have an Intel Crown Pass in a few days for testing, let's see how this is wired.

The CP boards don't have any x16 slots, though, I don't think. If that's the case, it wont use all 80 lanes. I think you have to step up to the CO boards in order to use all of them.

Just searched a bit ... Looks like Crown Pass has a few x16 slots :-)
CO is much less flexible, only two double width x16 cards.
 
Intel Workstation Board W2600CR (Crown Pass):
Slot 1: PCIe Gen3 x16, connector from first processor.
Slot 2: PCIe Gen3 x8, connector from first processor.
Slot 3: PCIe Gen3 x16, connector from first processor.
Slot 4: PCIe Gen3 x8, connector from second processor.
Slot 5: PCIe Gen3 x16, connector from second processor.
Slot 6: PCIe Gen3 x8, connector, from second processor
Slot 7: PCIe Gen3 x16,connector, from second processor
Slot 8: PCIe Gen2 x4 electrical with x8 physical connector, from second processor

Intel Server Board S2600CO (Copper Pass):
Slot 1: PCIe Gen II x4 electrical with x8 physical connector, routed from Intel C600 Chipset, support half-length card
Slot 2: PCIe Gen III x16 electrical with x16 physical connector, routed from CPU1, support full length card
Slot 3: PCIe Gen III x16 electrical with x16 physical connector, routed from CPU2, support full length, double width card
Slot 4: PCIe Gen III x8 electrical with x8 connector, routed from CPU2, support full length card
Slot 5: PCIe Gen III x16 electrical with x16 physical connector, routed from CPU1, support full length, double width card
Slot 6: PCIe Gen III x16 electrical with x16 physical connector, routed from CPU2, support half-length card, Intel designed PCIe riser card
 

Hello.
 
It would be really nice with some "extended" specifications - (aka, a list of CPU's and RAM, that will work, when you keep in mind that e.g. 2 x i7's wont work, etc.)
 
Especially, a list of Memory, that will work - I am considering 8GB or 16GB modules for my server (think the 32GB ones are too expensive atm) -- I know that the spec. sheet only specifies up to 96GB (but alas, the SR-2 spec. only 48GB, but runs fine with 96GB) -- that is, would and could the memory controller handle 16GB and 32GB modules, that is, up to 192GB or 384GB respectively.
 
Thank you in advance.

linuxrouter
I just downloaded the Asus manual to see what the specs look like and this is how they setup the 7-slots with support for quad SLI and quad cross-fire:

PCIE 1 - x16 Gen3 (x8 if slot 2 is occupied) -> CPU 1
PCIE 2 - x8 Gen3 -> CPU 1
PCIE 3 - x16 Gen3 (x8 if slot 4 is occupied) -> CPU 1
PCIE 4 - x8 Gen3 -> CPU 1
PCIE 5 - x16 Gen3 -> CPU 2
PCIE 6 - x8 Gen3 -> CPU 2
PCIE 7 - x16 Gen3 -> CPU 2

This makes the ASUS board superior at this point...
and still to this day i don't understand the logic of using a mux chip rather then the lanes off the second cpu.
 
edit: Jacob said single socket cpus would work unoffically, if that becomes official then the use of the PLX chip will have a purpose.

shadow001

Well i guess it's a pass for me regarding the SR-X, as i still haven't maxed out the potential of my SR-2 since it's only running X5650's, and now that the E series chips are out, the X5600 series is bound to get a lot cheaper to clear inventory and i can see a pair of X5680's or even X5690's and get them to clock to 4.5 Ghz(system is water cooled).....

If i'm understanding the overall situation right, we need one of 2 options:

1: Socket 2011 Xeon's processors with unlocked multipliers since they have the second QPI link enabled for inter socket communication,or:

2: Regular socket 2011 i7's like the 3960k, where the multiplier is unlocked already, but where the chips also have that second QPI enabled and not disabled for single socket systems.

In other words, the future of the SR-X lies with intel exclusively in order to be used to it's fullest potential, and for those wondering all their processors have 2 QPI links by default but it's the firmware within the chip itself that keeps it enabled or disabled at intel's discretion....Firmware that requires special tools to access the read only memory where the data is stored and modify it as needed

 
I'm pretty sure they actually physically sever the traces.

They are not physically severed, they are programmed out by properly configuring fuse data during the production/binning process.  Anything else would be a waste of energy and time (= money).  This is how all modern CPUs/SoCs are made and it is not exclusive to any specific vendor.
 
Fuse data is not typically changeable once CPUs leave the manufacturing facility - usually, it can be temporarily "uploaded" to the CPU for debug reasons, and it would be valid until the next power cycle (cold reset) but you can be pretty sure that the tooling for this would never leave chip vendor's facilities as it represents the most sensitive part of their business models (binning and segmentation) :)
 
Exception to this is when chip vendor decides to make some fuses programmable externally - e.g. AMD's core unlocking.  But this is just an exception and I would doubt very much Intel would do something like this unless they have a valid pressing business reason to do so.

Hate sr-x , doesnt seems to be an nice one from EVGA, but an evga sr-2 does its job in the best way.
If I were to switch from sr2 to srx , I would say its totally useless.

TBH i love the SR-X and will definitely be purchasing it (providing i keep my new job/manage to make decent money from it), however regarding CPU's I will probably wait AND HOPE that the Ivy Bridge LGA-2011 Xeon chips are overclock-able. My guess is that they will be, combined with the tri-gate transistors and die shrink improvements, also unlocked CPU's, so we can truly get the best out of SR-X.
 
Martin

It's mostly because with the SR-X, aside from the price of the motherboard and that of the higher end 8 core Xeon's wich aren't exactly cheap even if they are the best performing ones, don't get the advantage of "free" performance by overclocking the processors, so the SR-X is just like any other dual socket motherboard made by tyan or supermicro or asus.
 
Without that bonus what's the point of it, though i was surprised by the relatively low price of the 16 core AMD interlagos(dual 8 core bulldozers), wich top out at 1200$ each for the fastest version, so if the idea is to use software that can scale to 64 threads, and one can't overclock either the interlagos or the intel E2690 Xeon 8 cores, i wouldn't proclaim a victor in advance given there's 32 extra cores with the AMD option.
 
8 core sandy bridge E Xeon's are good but twice as fast as a bulldozer they are not in order to overcome that much of a disadvantage, and either option is close to the same price, but you get 2 processors from intel(16 cores/32 threads) and almost 4 processors from AMD(64 cores/64 threads)