custombuiltcomputers We have run some tests with the program LinX 4 core 4 thread. This is in no way trying to show how stable our processor is with this program. We do not have hours and hours to run this stability test just to satisfy a few people in the forums. What we do want to show is the heat removal capacity of our modded OCZ unit. There was a post in the beginning of the thread of a screen shot of a W3540 @ 4.818 @ 1.41 volts. This individual stated he had a custom built phase unit from someone over at XS and that it was calibrated to 220 watts of heat removal. His average running temps @ 4.818 @ 1.41 volts was in the mid 30c rang with a peak of 34c. We have set our i7-975 to the same clocks but with much more voltage than he was running. All and all we feel that the test results we are posting should be creating more heat than his run at the settings he listed. We like to show temps with Everest for it can show the rise and level off of temps throughout the test. We have also run Real temp for that seems to be the program used throughout this thread. As you can sfies some of the skeptics here slamming our modded unit.
 
Oil change allows the use of different refrigerants, hence our blend of three gases works well. 

Link listed below of test runs with temp.....
<a href="http://s1006.photobucket.com/albums/af190/cbgpcs/?action=view&current=2.jpg" target="_blank"><img src="http://i1006.photobucket.com/albums/af190/cbgpcs/2.jpg" border="0" alt="Photobucket"></a>

 You ought to read this thread completely before you attempt once again to place yet another fraudulent marketing claim about this thing of yours. Your post shows 54GFlops. At 4.8GHz with HT off you ought to be doing closer to 70. This means you are running at about a comparable 4GHz rate, which in a cold room a good air cooler would achieve similar high temps. If you want to make claims like you do, first you need to put a load on that chip of yours so folks could reduce their skepticism.

overclocker333

johnksss

yeah...like all of a sudden...the cpu doesn't have enough power to say...."im done"

LOL... I'm sure you inow... it's not the cpu... it's one of the many LinX bugs.

yeah..i was just using that as an example, because that's about what seems to happen. if you look...you'll see the temps drop to supper low while the cpu is idling and right as it kicks in to finish the linx test...you get the famous error. that is the time the cpu is powering backup (so to speak)
im in the 4.7 range now...but it wont complete!

trs32505

The reason his GFlops are down is due to his low QPI and high multi. Now if he were to do the opposite then his gflops would increase.

Yes, but even with a low QPI, he should still be getting more GFlops. I have the same cpu, and was getting well into the 60+ GFlop range using only multi to overclock.....with HT off, and was only at 4.5GHz compared to his 4.8.

With HT on, at 4.5GHz again, GFlops dropped to 58.

On the other hand, when I changed my overclock to put more emphasis on QPI instead of multi, GFlops increased, marginally. They went from 58 to 59. Definite increase in performance, but not enough to really make a ton of a difference or match what he's got going on. Something else is wrong. Could possibly just be an issue with LinX.

I can run LinX with a particular setup, and get 58GFlops. The restart my computer, and run the same thing again, and get in the mid 59's. It's very unreliable.

johnksss

low uncore & possible ram at lowest state. (speculation of course)

Very true, this could have a huge impact on GFlops. He may have the memory all set very low to make it easier to get to 4.8GHz.

After all, he didn't want to spend even one hour to run LinX in its entirety, so it's probably safe to assume that he didn't spend a lot of time putting together a top quality overclock either. (more speculation, just trying to give some ideas as to why the GFlops are so low)

Boy, you guys are like slamming the hell out of me...glad I could add some spice to the thread........

I'm done , I'm out, nice haveing the evga community welcome someone like you do..............

trs32505 The reason his GFlops are down is due to his low QPI and high multi. Now if he were to do the opposite then his gflops would increase.

Any way you look at it, the processor is idling.

To SimC33: I have a i7 Club listing for HT Off-----below is my very first HT On submission ----

Corei7: 920 DO Batch#: 3937A806
CPU clocked at 4002.1MHz(CPU-Z)
Voltage: 1.33125
Idle Temps: 37, 34, 40, 31
Load Temps: 81, 78, 79, 75
HT On

              Well guys I was able to get what I was after. Ran this one with the fans blowing cold outside air into my back room. Ambient temp around 26F. I knew my vcore was high enough and the Vtt so I have been messing with the other voltages.
              Actually I think I could lower my vcore. Anyway I have become a believer in with vdroop. CPU runs a lot cooler that way if nothing else.                                                                                                                           
                                                                                                                                                                                                                       Giving overclocker something to shoot for with his new phase.

Thanks. I am thinking overclocker will take it in the fun it was intended.

awalleyeguy Well what I noticed was voltage of course goes down at load with vdroop, and on my last run varied between 1.588 to 1.621 and pretty much stayed there even if upping setting a couple notches, now lowering a few nothches could bring load voltage down, but not the same amount as what voltage was lowered. In other words. If I lowered voltage say 4 notches the load voltage may have only gone down 2 notches. So I am guessing that load voltage has something to do with how much voltage the CPU needs for a set clock with a particular voltage amount available. So it is not etched in stone that with vdroop and 1.5v setting cpu will get 1.42v under load.  Of course without Vdroop that's a different ball game all together, sorta opposite.

The *Club i7 Stability Continuation* thread is probably one of the better repositories of i7 OC information and statistics, but the data is often contaminated or otherwise under illuminated by guesses and outright mystical insights, and in that vein I am mystified how so many of us can work so hard and expend so many resources in this Herculean effort but yet operate with so little technical guidance. Or, maybe it's simply just more fun to work in the dark.
 
There is available to us one critically important document which it seems many of us are totally unaware of. It is, the expressive but dreaded, "Intel Core i7-900 Desktop Processor Extreme Edition Series and Intel Core i7-900 Desktop Processor Series Datasheet, Volume 1," October 2009. You can pick up your very own free personal copy at:
 
http://download.intel.com/design/processor/datashts/320834.pdf
 
In it you can find many wonderful and glorious things, e.g., descriptions of Vcc (Vcore) and Vtt and VID, and their interrelationships. In this supernatural and transcendental tome are revealed all the secrets of the universe, and if that is true, you ought to read it.
 
One of the more interesting insights into the soul of the 900 is the disclosed and confessed intimate relationship between VID and Vcc. From this one small but significant book you will learn first hand from the sacred mouth of Intel just how it is that VID commands the value of Vcc as a function of Temperature and Time. So yes, boys and girls and ladies and gentlemen, right there in holy scripture you can be enlightened, you can learn just how it is that when the internal thermal sensors of the chip achieve a critical set point, Intel will induce VID to tell you and your motherboard just how to adjust Vcc for optimum thermal performance, and yes, from this little known event in the life of your processor you can see and empathize with it as it lowers its Vcc to keep itself cool, and as we all know, cool is cool.
 
So, therefore, mystery of mysteries, with datasheet in hand we will no longer be forced against our will to acquiesce to the forces of darkness which would have us guessing about truth. Rather, liberated and enlightened we can bravely and confidently assert just how it is that that pesky and uncertain "with Vdroop" so thoroughly and completely controls the known and unknown universe.
 
 
But what I still don't understand, why does and how is it that "without Vdroop" Vcc rises?

@ awalleyeguy
 
Your latest overclock is just as beautiful as grd003's post!  Great job on a great chip!!!

I think it is now time for you permanently to cross over to the cold side since the cold bug has bitten you... it seems to have that affect!  Wouldn't it be nice to have a chiller instead of just that cold room?  I see a phase unit and 5ghz LinX run in your future!

expedision 

DrNip 
  

 I believe the B027's are the best 920's out there with the B026's coming in a tight 2nd place.  Either way they are both great batches. 
  
DrNip 
Core i7 920 D0 Batch#: 3845B027    
CPU clocked at: 4420.0 MHz      
Voltage: 1.270 idle/1.282 load   
Idle temp: 37,38,36,35    
Load temp: 71,71,67,64 <- Priceless  
Air  

expedision  
Core i7: 920 Batch #3845B026 DO 
CPU clocked at: 4420.1 MHz 
Idle Voltage: 1.25 Wo/VDroop 
Load Voltage: 1.27 
Idle temps: 45,45,44,42 
Load temps: 84,85,79,80 

Are you sure about that  

 


Yeah I'm sure about it.  In general they are a little better I believe and there will be those out of each batches that will shine a little more than the others, like yours.  I would pick mine still out of the two of ours as it runs the same speeds with just a slight bump in volts but way cooler temps.     

Of course that is probably just to do with ambient temps in the room or is it.   Why is your result posted in the "HT On" category when yours is actually off? 

custombuiltcomputers 

We have run some tests with the program LinX 4 core 4 thread. This is in no way trying to show how stable our processor is with this program. We do not have hours and hours to run this stability test just to satisfy a few people in the forums. What we do want to show is the heat removal capacity of our modded OCZ unit. There was a post in the beginning of the thread of a screen shot of a W3540 @ 4.818 @ 1.41 volts. This individual stated he had a custom built phase unit from someone over at XS and that it was calibrated to 220 watts of heat removal. His average running temps @ 4.818 @ 1.41 volts was in the mid 30c rang with a peak of 34c. We have set our i7-975 to the same clocks but with much more voltage than he was running. All and all we feel that the test results we are posting should be creating more heat than his run at the settings he listed. We like to show temps with Everest for it can show the rise and level off of temps throughout the test. We have also run Real temp for that seems to be the program used throughout this thread. As you can sfies some of the skeptics here slamming our modded unit. 
  
Oil change allows the use of different refrigerants, hence our blend of three gases works well.  

Link listed below of test runs with temp..... 
<a href="http://s1006.photobucket.com/albums/af190/cbgpcs/?action=view&current=2.jpg" target="_blank"><img src="http://i1006.photobucket.com/albums/af190/cbgpcs/2.jpg" border="0" alt="Photobucket"></a> 

 

OK that is fine and dandy but this is a"Club i7 Stability" thread not a "Promote My Product "thread.  Start a new thread but to be perfectly honest and no disrespect I don't think anybody cares anymore.

HAHA I hear ya.  Prolly never get rid of mine unless my X980, when it drops, is of the same caliber.

Hello guys,

I've finally gotten somewhere. I'm not pleased yet, because the time and the GFlops are way to low i.m.h.o. So when my ram is back from RMA, I will try to push a little further. One of my sticks won't show on cold boot. That's the reason for RMA.
(For the record: I didn't restart LinX after an error or stop on demand.)

Forum name: Seaquest
Core i7: 920 D0   Batch #:  3910A326
CPU clocked at:  4210Mhz
Voltage:  1.34375V Without VDroop / HT on
Idle temps: 40, 37, 40, 38
Load temps: 80, 80, 76, 76

OC 4210MHz


 

I hope I'm good to go for the nice badge....

WOW... 20 pages... 2nd only to the "Get It Colder II" thread!!!
 
Gratz to all of the new arrivals and hard work on those stable overclocks!!!  For all of you with low GFlops and high run times... that is part of the dreaded LinX bug... I would keep working on it till the GFlops and time come in line with the others you see posted.

This is turning out to be a great thread with a lot of very helpful info!!!  Lets keep up the great work and "Keep On Pushin!!!"