Here is a good number to expect at 4.4 clock speed.
26422
 
 

any updates of the delivering time ?

Somehow I missed this part of your old comment:
 
"when i started the build we did not had the numbers on 5960X overclock. turns out 4.4GHz is about is high is it will go, and stability is still an issue. But the new Intel CPU benchs like hell, and 8 cores at 4.4Ghz are more or less equal to my SR-2 with 12 cores at 4.4GHz for chess training."
 
You have the 5960X also? This is what I'm contemplating at a replacement for my 3930K and the SR2 that's impossible to find now. What benchmark did you get in SF again and was that with HT off? What motherboard, memory and coolers did you use? Or am I just misunderstanding and you're simply comparing your SF speeds with what was posted by Sedat for the 5960X?

yorkman
If you're referring to my X5680's then I've delayed it. I paid but I told them not to ship since I wasn't sure if I should buy from alibaba. I decided not to risk it with them and unless I can find a used/new SR2 for sale then I'm basically SOL and will have to give up on my dream pc. However, that won't be easy as I believe it's the best system for me to go with.
 
So main problem is that NOBODY is selling an SR2. For now I'm waiting to see if one becomes available. Or are there are chips I can go with that can be overclocked.

so you did not buy one of the two sr-2 system thad you know thad ar for sal the 1300$ or the 1050$?
hope for you thad you whil find a sr-2 ,so thad you can start to build your dream.and if not thad you wil recive your money back from the cpu's

Looking for your opinion here. If you were me would you go ahead and settle with the SR2 and 2x 5670's instead of 2x 5680's? In addition to this you get 48 GB Dominator ram (2000 Mhz) and an Obsidian 900d case all for $1400? I think the 5670 is basically the same as the 5680 except that it's 400 Mhz slower so while I may not be able to overclock it to 4.5 GHz maybe 4.3 Ghz with 2x Swiftech H220-X's is still realistic? All I'm missing from this deal is a 1000W PSU (and of course the 2 h220's).
 
What do you think guys. Should I go for it (perhaps later upgrade to a couple of X5690's after their prices have come down)?

And is it possible with those coolers (H220-X) wihout any motherboard blocks?

I'll have to wait and see then. I can't buy it for another week. He's 500 km from me but will be in my area then. Once I have the system I will just order the H220-x's and see if I get lucky with my x5670's.  It'd be nice to order it today but I'd get burned if he didn't show or if he sold it to someone in his city before then. Then again, maybe it's worth the risk otherwise I'm in for quite a wait for the stuff to come in from the states. I suppose they'll take it back if I don't open it and I'll just lose out on the shipping fee.
 
The only thing I'm afraid of is if my x5670s won't oc well. Then I'll be suspecting the H220-x's just aren't cutting it and will force me to try custom cooling and I still don't know which loop would be a significant improvement over the H220-x. Needless to say I'd have wasted money on the H220-x's if I was to then buy a custom loop. Question is, for an Obsidian 900d, which loop is best yet cost is respectable?
 
 

I dunno. It doesn't look very impressive and certainly not as good as the H220-X. It just doesn't look very powerful...the reservoir is not visible like in the H220-X. But if it's better than the H220-X then I'm all for that. It's in stock for me too so that's a plus.
 
I just don't know if it's any better than H220-X.

The H220-X is not in stock for me and it'd cost me more if I had to order from the states. Do you think the H320 is better than H220-X for the SR2 and 900d case? It looks like two should fit (one at top the other on side). I'd prefer not to just get my feet wet. The primary goal is to cool everything well and provide highest overclock.
 
What I don't want is to go with the 220 or 320 and find out that I could OC higher if I went with a custom kit instead. That is the primary goal, highest stable OC for 24x7 with acceptable temps. Would a custom loop kit like the XSPC RX-240 be that much better than say the H320? I'd be will to go that route if as long as it's not just some little difference like 1 or 2 degrees which still achieves the same OC as the H320 or H220-X.

Thanks for the info. But if all I could get was 4 GHz with 5680 or 5690 I'd be very pissed off. 4.5 GHz would be my min. target. Same goes for the X5670's...I wouldn't be happy until I got 4.3 GHz.
 
I wouldn't go with H110. They're much louder and not as strong as the H220-X or H320, according to my research so far.

you can connect two H220 for two 5690 ? and make a single circuit ?
the liquid out of the WB 1 does not return immediately to the radiator but joined the WB 2 after going through the rad and 2 rad after 1 before returning to the WB1

Hey guys! I have good news, bad and ugly.
 
The good:
 
I got myself an SR-2 with dual x5670's using 2x Cooler Master 240's, 12x4=48 GB Corsair Dominator GT 2000Mhz (9-10-9-27 @ 1.65V) and a Seasonic 1000W Platinum PSU all in an Obsidian 900d case. I can OC to 4.2 GHz. No video card but I put in an old el-cheapo PCI-E since I don't care for graphics. This rig is for chess only.
 
The bad:
 
1) BIOS and OS won't see my full 48 GB of ram. It either sees, 24, 32 or 40 GB.
 
2) OC is not overnight stable at 4.2 GHz. I set the timings for ram as per specs and tried auto also. I tried Shamino's template but it's much less stable at 200 blck * 21 (4.20 GHz).
 
But with 191 blck * 22 it at least passes LinX test every time but I only let it run the default 20 runs testing just 2 GB of ram. Just trying to get a quick down and dirty stability check. Temps at full load are a bit high. CPU0 temp is about 75c and CPU1 is about 60c under full load with stockfish engine.
 
I've seen that pic that shows the labeling of the cpu0 and cpu1 are wrong or something and it says it's normally for cpu0 to have 5-13 degrees higher than cpu1 but I don't know why there'd always be such a huge differences in degrees between one cpu and the other, even at same cpu voltages. And are those my real temps for both cpu's or the cooler temps more correct?
 
Other than that it rocks. I'll work on the custom cooling once I get this thing stable. I'd like to get 4.20 GHz stable for now. I can live with 40 GB of ram detected but naturally I want the 48 GB, and too often it's either 24 or 32 GB instead. Evga E-Leet is the only thing that sees the 48 GB of ram consistently. Win'7x64 only sees what the bios detects which is 24, 32 or 40 GB.
 
The ugly:
 
I need your help guys with both issues. Below is my current BIOS config:


I have IOH vcore set to 1.40V set in bios (latest). CPU0 and CPU1 are set to 1.35V. VTT is also 1.35V. HT, SpeedStep are disabled. I tried higher for everything (1.40)V but temps get way too high for 24x7 (88-96C in LinX). 



 

Tis ar my bios settings,if not mentiond its on auto or on stock
 
CPU
 
C1E : disabled
MAX cpuid : disabled
Intel VR : enabled
Intel VT-D : enabled
Corency support : disabled
Execute-disable BIT : enabled
Active core : ALL
A20M : disabled
Speed step : disabled
C-state tech : disabled
Acpi T state : disabled
CPU : 200
PCIE : 100
CPU multi : 23
QPI link : full-speed
QPI frequency :
Memory frequency : DDR-1333
MCH strap : auto
CPU uncore : auto
 
Voltages
 
Vdroop control : whitout vdroop
Boot Vcore : 1.41875v
Eventual Vcore : 1.41875v
Boot cpu0 VTT : 1.375v
Eventual VTT : 1.375v
Dimm voltage : i need to set 1.75v ,my bord seems to be need a bit more volt's on the rams. it dont boot stock speed at 1.65v
 
Signal tweaks
QPI 0 : -85
QPI 1 : -16

do you realy need to run IOH vcore at 1.4V?this can lower your temps if you can lower the voltage

Maybe not. I was just trying that to see if it'd help stabilize at 4.2. I have the system, so far, stable at 4.1 as it's been stockfishing for 7 hours now without a hitch. Temps are pretty good now as I've lowered IOH to 1.375 before starting my 7 hr long analysis but that's also likely due to lower ambient temps (it's cooler outside today). I did notice my ram chips run quite hot..and my rads are also blowing very hot air almost to the point that I can't place my hand on top of the case where the air blows so perhaps these coolers aren't the best...or they're just working very well.
 
I'm probably going to let it continue analyzing for another 16-24 hrs to at least be sure that it's stable at 4.1 GHz.