Well,here is mine

Knightviper

everything shut off like you said and the Higest problem size I can get is 24,768Mb can anyone help on getting me to the 25,000 problem size?

Go into control panel, then to administrative tools, then open services. Scroll down and disable Superfetch and reboot.

I don't think it is the voltage but rather your memory clock, timings, your uncore and your QPI.

Core i7 (980X)   Batch #:) CPU clocked at: 4422.05  Why the heck are my screenshots looking like crap? all the info is there, if you could only see it, lol
Voltage: 1.575 without vdroop
Idle temps: 25-33
Load temps: 80-84
GFlop peak: 81.8119

This is wHat I get after I've done what you's have said on the msconfig bit

Knightviper

This is wHat I get after I've done what you's have said on the msconfig bit

 
First, I can't read what you are posting...the images do not "blow up" or magnify when you post them like you are doing. Either go to a hosting service, aka imageshack.com, or put the image you want to post through Photoshop or Photoshop Elements, whatever graphic editor you happen to use for photos and images, and BLOW IT UP to at least 10" x 14" size, then scale back the resolution a little with the "Save As" dialog to make the thing fit to less than 500KB in size. If you do that hat trick the images will scale properly when you import them into the EVGA.com robots and computer sizing machinery, and we can see them. 
 
As for your problem with the memory, you have some programs which are installed that are taking RAM, and they can be *hidden* programs within applications you have installed or they can be simply OS solutions, anti-virus solutions, or other start-up mechanisms which are taking  a couple thousand MB of RAM and giving you this headache right now. Let me help you out, the easy way. 
 
First thing is open Task Manager, click on the Applications tab, and see if there's anything there that is "On" presently which are taking RAM from your allocation, something that you are not using openly. That's the first thing to do. Second, click on the Processes tab, and simply go down the list of processes that are running line by line, slowly, carefully, and see what is taking RAM for its uses that isn't necessary to the safety or operation of the OS itself, and you *will* find the culprits there, I almost can guarantee that they will show their ugly little faces in that panel of TM...it's simply a matter of knowing what you need to keep using vs not using it.  
 
For example, if you have iTunes installed, there's "iTunes helpere.exe" which will be running; the "AppleMobile devices.exe" process will be running; and even "appleupdatesvc.exe" can be running, all three of which will take up several thousand KB of RAM. Are you using AnyDVD to RIP movies? AnyDVD will have its "helper" application running also, called "AnyDVDtray.exe", and that is unnecessary also. You may have browser executions running there also, ie "chrome.exe" tabs which can take many thousands of KB of RAM. When you *find* processes that you don't need, simply go to the tab at the bottom of that TM window where it says "End Process", (highlight the process first by clicking on it once), then click the End Process tab, confirm the operation when asked, and BAM! The answer will be right there frankly, and it's up to you to find the culprits and end their processes.  
 
This will *not* hurt the computer, nor the processes themselves, as they will come back the next time you reboot, as all you are doing is temporarily killing them so you can run LinX, which is very sensitive to other running processes and the OS processes, and will feint its own needs for other programs, which is what you are experiencing now.  
 
Now if all of that is too complicated, there's another angle to use: "Whatinstartup" is the program's name. Google the term, download the application, and use it! I have it on all my PCs and it works a treat. In that program you can readily identify all running start-up processes and enable or disable them at will, kind of a hat-trick in itself, and you can get all of the things that are running optimally under control 100% using Whatinstartup, reboot after indicating which programs you want to have killed the next time you start up the computer, and BAM! Handled! 
 
Those are the obvious things that come to mind right now, and they are simple things which you can take charge of the OS processes with and end the startup programs one way or another that are taking RAM from LinX 0.6.4. Good luck, and I hope this primer helped you figure out the dilemma you're presently in. 
 
rexrzer77 
 

Yea i had to go in MSCONFIG and get rid of all the startup apps...

jethc9

rexrzer77

I used the LinX in the download on Page No.1 of this Stability thread, and did
it all by the book-ie the LinX 0.64 with the new binaries on the 1st page of
this thread. If the GFlops don't look high enough for your purposes you can
fault the LinX on Page No.1 here...not my run.

Gflops can be affected by voltage. Not everyone with the same OC will have the same Gflops. I think as long as you are in the ballpark no one would call shenanigans.

I know that I have had varying degrees of Gflops depending on voltage. I have seen anywhere from 78-82 with my same 4.2Ghz OC

 
I simply found it "odd" that my GFlops were quite low, respectively, compared to my Core i7 970 CPU using the same LinX program posted on Pg. No. 1 in this thread, and I didn't want anyone to question which LinX I used...that is why I made that comment above. 
 
GFlops can be affected by many factors, as you all noted: temperature, voltage, RAM timings, RAM itself (its configuration, ie 6 x 2GB, 3 x 4GB DIMMs etc), and the version of LinX being used obviously affects GFlops. As for my particular test last night, that version of LinX is quite a treat, isn't it?    Hah! I never had such a temperature situation happen before with any LinX run prior to that run with that exact CPU and computer configuration, ie it normally can push 4.5+Ghz ON AIR ONLY and survive the exam, but NOT with these new binaries, No Sir, it could not survive such a test with *that* particular LinX. 
 
I started out testing at 4538Mhz and saw the temperatures per Core bounce off of 100-degrees rather instantly, which floored me, frankly! "Hmmm, that is most unusual!"--I thought. I thought something was amiss, because I swear on a stack of Bibles that I have passed 4538Mhz with that particular CPU and rig many times with previous versions of LinX, including the LinX 0.6.4 prior to the one posted here, so I really thought the worst at that point.  
 
I took my CM Advance II 690 case's Right side off, which is a real pain because I have a single CM 140mm fan mounted there and it attaches to the fan header below my last video board, which I have to attach by "Feel" as there is no way to actually "see" what you are doing there otherwise with the video board mounted in PCI-E slot No. 4, and looked for anything obvious; ie a fan stopped, wires off their points of attachment, anything unusual at all and I could find nothing happening of note. (I did that with the computer running, grounded with my ground strap) So I shut the thing down and got aggressive and quickly pulled the Megahalems and checked the TIM and tension etc on the CPU, and that looked fresh, still, but I re-did it with new MX-4 and re-torqued the cooler carefully with my mini-torque wrench (just for this sort of thing about 5lbs tension, for you interested in such things), triple-checked everything else in the computer's bays, and it all looked great! 
 
I tried again at a reduced Mhz setting, 4438Mhz to be exact, lessened VCore significantly to about 1.33750 (down from 1.36875 where that CPU normally can run 4.5Ghz without issues), and again, BAM! CPU Core temps shot to 95-degrees instantly, and higher yet again, and I am not going to run 20X iterations with the CPU bouncing off of 95-97-degrees for anybody's competition or recognition, I mean THAT CPU is valuable to me (it's a Batch# 3849B056 920 D0 and runs normally incredibly high Mhz settings with ease with low VCore and VTT settings, in fact I've benched it with Vantage/3DMark06 at 4726Mhz (!) ON AIR without any stability issues at all...it's by far my best CPU of any sort).   
 
Puzzled, I pulled out the last LinX 0.6.4, the one prior to the new one here, and ran the CPU back at 4.5Ghz settings, and lo and behold the temps never got higher than 90-degrees in 3 iterations, so now I found the culprit, it's the new LinX itself! Holy Crap! What in the world did Intel do with those binaries, anyway, I thought to myself! That is a net change or temp capture ratio of more than 20%, going from sub-90-degrees at 4.5Ghz, ON AIR only, to 100-degrees, and higher maybe (!), with the "new" LinX binaries, so I dunno guys and gals if any, I guess it's legitimate and I'm not complaining at all about this situation, I'm just noting the fact that this is a HUGE variation in Core temps with this new LinX variation, and frankly my CPU's, even my best one, cannot contend with the results it produces...that's simply just too much heat and will either damage the CPU, or negate any run because of throttling or other bad things happening as the CPU tries to run without burning itself up!  
 
Thus it seems that Intel, in its greater wisdom and software configuration prowess, has thrown those of us using AIR ONLY a left-curve that just can't be hit! I'm not going to risk damaging my CPU's, or having generally poor results overall, at high Mhz settings with the new LinX in the house as the test mechanism, end of story, sad to say. If this is the result they wanted to have, then more power to them, but it certainly makes my day a little less "nice feeling" frankly speaking, ie to know that my previous high Mhz configurations and settings make no difference with the new utility and that there's no point to doing overclocking extremes ON AIR only when you have results like that.  
 
Even at 4315Mhz settings, with VCore minimized and VTT and everything else the same, I found there's a 10-15-degrees Core temp difference using this LinX vs any other LinX or any other test mechanism that I am familair with, including PRIME95, Intel Burn Test with the latest binaries, so it's just simply very frustrating to me, personally, to contend with something that produces results like this particular LinX does with good CPU's, even at one's best possible optimized settings. Hitting the temps I did in that test I posted here today at a mere 4315Mhz sort of shocked me a little bit, and I found myself watching the test from beginning to end, monitoring things *just in case* there was some other bad thing that was going to happen...luckily it never did happen, but that CPU worked very, very hard to pass at that nominal 4315Mhz setting, and the temps never varied much from 86-93-degrees, Core to Core, for the whole testing session.  
 
In other words that test is brutal for people on AIR ONLY, and produces results that are far and away the most extreme Core temps I have ever seen in any test, at any time, bar none. I would venture to say that it also produces the same sort of result for people using H20 or freeze, whatever sort of cooling is being used, relatively speaking anyway, so it definitely "levels the playing field" and gives any Intel Core i7 CPU a workout unlike anything before these new binaries started to be used in LinX testing, this is certain.   
 
That is my take on the new LinX with the new binaries in place, so rest assured that anyone ON AIR only who passes this test legitimately at higher than 4300Mhz or thereabouts is going to have their hands full with extremely high Core temps, and a word to the wise folks, "keep an eye on your equipment" during the whole test regimen because this LinX is a little dangerous if you ask my opinion about it, versus the LinX of the past. Anyone who passes this LinX's exam, ON AIR only, at higher than 4300Mhz will have definitely earned their stripes so to speak also, it goes without saying.  
 
I am glad I could indeed pass at this relatively "easy" 4315Mhz setting in my case, and even though this particular CPU is probably the best 920 I've ever seen from anyone's stable, it seems that even my special CPU is not capable of going higher than the 4315Mhz area with the new LinX ON AIR only...maybe I could hit a 207 multiplier and pass at 4338Mhz or so, but I think that's the "WALL" and I wouldn't want to see the results, again, of trying to go higher than that somewhat "normal" sounding Mhz with the new LinX as the  test medium.  
 
Like I wrote above, I've actually tested this particular CPU as high as 4726Mhz ON AIR only, with Vantage, 3DMark06, so it is an exceptional CPU with a knowledgeable tester at the helm doing the right adjustments to do such things, but reality is simply much, much lower settings with this new LinX are the rule of the day. All that being written, I honestly don't know *what* the genius's at Intel had in mind when they concocted this test but they certainly got their due, and then some.  
 
rexrzer77 
 
 

So rexrzer77, do you think the 25000 problem size is enough to call your system stable?

rafale

Rexrzer77. Wow. I just finished reading this post. How long did it take you to type it? 

 
You're new here rafale, so you're not "used to me" yet...Hah! 
 
I'm a technical writer (for a living, my job) and I type at some ridiculous speed, perhaps as high as 80-90WPM when I am flyin' high on an assignment, and even faster when I can afford a mistake or two while typing something like the above missive, so I dunno! Maybe it took me about 10 minutes to do the whole deal? Yes, something like that! 
 
Before I started tech writing I was a Journalist for various news organizations also in the 1970-'s-80's, (I'm an "older" person), and before that editor of my high school and college daily and weekly newspapers, so writing has been at the forefront of my being since ever I can remember. The story goes like this: I learned how to type when I was barely 8 years old, it turns out, because I *hated my handwriting so much* that I never wanted to have to write things out by hand ever again, or so my parents told me when I had grown up.  I rode my bike over the our local high school and took a typing class in summer school when in 2nd  grade, they tell me, and that started the whole thing. 
 
I also play the piano for some 50+ years, since I was about 3 years old and used to pick the notes out from songs on the radio at the keyboard, thus my parents thought I ought to have some piano lessons since I liked it so much! All in all my fingers are pretty talented, shall I say it in a humble way!? My apologies for not telling all the nice people at EVGA.com this whole deal before now, but you asked sort of, since you queried me about my typing speed, so there you go folks! That's why my posts are so long, and sometimes they are toooooo long for this place, as I know that people are not used to reading such things in an online format, where brevity is usually the standard by which people are judged and understood.  
 
I actually have to "throttle myself back" some here, believe it or not, otherwise I'd go on forever about an idea or line of thinking, or so it seems anyway. I guess I have "an active mind" to boot, and never stop thinking about this or that when I'm awake, and you know what? It gets demanding and somewhat disturbing at times, because I find it hard to "relax" easily...I'm usually thinking deeply about something or other, you know conceptual thinking (?) about things, when I'm not writing about them. So my apologies to the group here, in general about writing *so much*, but that's where it all comes from! 
 
What do all of you think about the new LinX anyway, that is my question to the group as a whole? I am still kind of shell-shocked about it, frankly! Why in the world would Intel's engineers figure out a way to literally "burn up our CPU's" if left to its own means? That is one HELL of a test, you know that? I'm just curious about how the rest of you who are active overclockers are actually thinking about this new set of binaries, or so they are explained to us laypeople (normal folks, not engineers like the Intel people are). 
 
I mean I can see a rigorous test to root out the deepest, darkest and meanest weaknesses that a CPU could possibly have, and I'm all for that. But why this emphasis on the heat-cycle of the CPU's, because that is exactly what this test administers, you know that? Have any of you actively watched and observed the test antics, the actual "character itself" of this test, just sat back there and watched the numbers bounce back and forth between the Cores themselves, have any of you done that simple exercise? And if you did, what did you observe, see? 
 
I did just that the last couple of iterations of the test during my stint with it last night, and I saw this: the Core temps seemed to have a an actual cadence to them as they would shift from one to the other on a 4-Core CPU. That is what I observed when I just sat there and stared at the numbers on-screen as the test proceeded, an actual method or sequence of things happening over and over again, a sort of repetition of the same pattern of Core temps shifting back and forth between the individual Cores.
 
In other words it didn't seem random per se, like you'd imagine such a test would generate in terms of what we "see" when the test proceeds on its "mission" of trying to tear apart our CPU's! I wish I could explain it better, but there seems to be a mathematical pattern to the whole deal--if you just watch it long enough the pattern seems to develop over time, and it evolves and has sort of a rhythm (sp?...I never get that one right, hope it is this time) to it, over and over again. It's NOT random is what I mean to write! 
 
Even the most random motion has patterns to it if you analyze something over a long, long period of activity, and I think that LinX 0.6.4 as we see and use it today has a structure to it, it isn't simply random figures shuffling back and forth with the Cores, rather the whole test seems to have a "life of its own" the more you watch it, an order or sorts, it's *not* random 100% of the time, although it would take something very special to nail down what, exactly those features of the new LinX have within themselves...a recording device of some kind would have to be used, a very intricate device that could annotate the actual numerical sequences that take place during a long CPU test. Or so it seemed to me last night as I watched! 
 
I may be getting daffy or am simply struggling within myself to "find" an order to the whole deal of a LinX test, but I honestly felt it wasn't just random Core use throughout the test: it seemed to have rhythm, cadence, and specifically had definition of a sort. I know, I know, crazy! But nonetheless that was my feeling as I watched things develop during the last part of my LinX 0.6.4 CPU test last night: I saw a pattern to it all, however esoteric that might sound or seem, there was some kind of rhythm to the whole pattern of the test. I can't tell you what it is exactly because the mathematics involved are tremendously HUGE to say the least, but somewhere in there is a structure of some means in a LinX 0.6.4 test with the latest binaries. 
 
OK, I've said enough I think! The floor is open for comments, but man, what a topic!
 
You know I am writing this note to the Forum at EVGA using the keyboard for PC No.1 in my studio, as I work the keys I am listening to some very nice music by a group known as Thievery Corporation, all instrumental with some voice now and then, but mostly just rhythm and drums, bass and a melody of some sort thrown in there for continuity. The album is called "DJ Kicks"  and the album itself reminds me of the LinX test last night, early this AM, whenever it was. Yes, it's free-form electronica music this is true, but as a song evolves and progresses there's a definite pattern to it all, even if it is "free-form" electronica, and things eventually begin to repeat themselves over time. Not exactly repeating themselves, not exact, but definitely there's a pattern to it and you begin to pick it up the longer a song goes. 
 
THAT is what it was like watching those LinX Core temps generate last night, and so it was noted in my memory banks, and that's what I am attempting to explain right now. Or so it seems...maybe I'm just confused about the whole deal after watching my best CPU almost come undone last night with that test, like WOW! What a workout that CPU got for a measly little 4315Mhz overclocking! That machine has never, EVER gone up into the mid-90's for Core temps at 4315Mhz before last night, and I'm still sort of in awe about the whole sequence of events that occurred.  
 
OK, 100% done now, I am dusted, busted, and definitely finished with this post! The floor is all yours...
 
rexrzer77 
 
 
 
 

Well, despite your experience, my Gflops are slightly lower and temps are essentially unchanged with these 'new' binaries. You may have been using the original binaries for LinX which did not properly stress the i7 cores.

Yeah, I don't know about the temp increase. It looks the same as my IBT temps. I really haven't observed what Rexrzer77 has.

I used the msconfig way and killed everything that was in there the prob that I was having was that LinX ran but never showed any results in the passes

 
My office room was a bit colder tonight so I attempted another shot at getting a better clock.
With a room temp of 22C (3C lower than last night)
I am just a hair below 85Gflops by raising my bclock by 1MHz.
My 980X got to 4.36GHz
Vcore of 1.38125
Idle Temp of  24/14/25/33/30/20
Load Temp of 75/73/71/75/80/75
 
1MHz more on the bclock and I get the disappearing RAM issue...