• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

First real OC, everything look OK?

U

Uncle Humjaba

Limp Gawd
Joined
May 6, 2006
Messages
440
Been running prime95 for about 15 minutes with no problems so far...
Ran it for 30 minutes previously with core voltage set at 1.3125 in the mobo and had no problems so I dropped it to 1.27500.

wbtaut.jpg
 
Alright... I have it downloaded... What do I do? I turned on error detection, stress level 1 (max stress), run 10 times, and it fails after maybe half a second. Is this an unstable overclock or is it something else?
 
The program said "test failed" almost instantly. It's now running doing 1/2 stress. Could it be that I have 4gb of ram in the system, and the 32bit version of windows can't utilize that much?

EDIT: It just passed 5 tests at 1/2 stress... Memory timings are locked at stock speeds and voltage so I would imagine using half memory vs all memory would not affect the results/reliability of the test? Temps on the first two cores were 57*-60* with the other two 54*-57* during the test.
 
well i dunno about that i usually pass 20 time of maximum stress for stable qualification
 
Uncle Humjaba said:
The program said "test failed" almost instantly. It's now running doing 1/2 stress. Could it be that I have 4gb of ram in the system, and the 32bit version of windows can't utilize that much?

EDIT: It just passed 5 tests at 1/2 stress... Memory timings are locked at stock speeds and voltage so I would imagine using half memory vs all memory would not affect the results/reliability of the test? Temps on the first two cores were 57*-60* with the other two 54*-57* during the test.
Click to expand...

FYI, 32-bit Windows handles 4 GB of ram, but when you have 4 GB part of it must be reserved for hardware purposes which is why you lose about 500 MB of memory.

In short, that definitely wouldn't be the issue.

(2^32 = 4,294,967,296 addresses)
 
Well I'm not really sure what the issue could be then. It literally fails as soon as I press enter. No spike in CPU usage, temp, or anything... just a steaming pile of fail.
 
Uncle Humjaba said:
Well I'm not really sure what the issue could be then. It literally fails as soon as I press enter. No spike in CPU usage, temp, or anything... just a steaming pile of fail.
Click to expand...

I'm not an expert in overclocking by far, but I believe that instability can frequently be caused by underpowering processor. Have you tried going back to 1.3125 volts yet to see if it would make a difference?
 
NExUS1g said:
FYI, 32-bit Windows handles 4 GB of ram, but when you have 4 GB part of it must be reserved for hardware purposes which is why you lose about 500 MB of memory.

In short, that definitely wouldn't be the issue.

(2^32 = 4,294,967,296 addresses)
Click to expand...

Not entirely correct.

Windows XP can't address all 4GB, however other 32bit OSs can.

The problem here is that a single 32bit app can only address a maximum of 2GB of memory which is why IBT fails immediately when you try to go over 2GB.

If you want to do a full memory test with IBT you will need 64bit windows.
 
I'm in Vista 32bit, but still in dxdiag it says I have like 3350mb ram. I will take 2gb out tomorrow and test it. Again, I'm not overclocking the ram, so I'm only interested in the stability of the processor. If that doesn't work then I will bump the voltage back up.
 
I'm running Windows XP 32bit with 4 GB of ram and it passes IBT v1.8 just fine (I just can't test the full 4GB).

I would bump the voltage back to 1.3125 at least. You're overclocking and undervolting at the same time and expect it to be stable??
 
Deimos said:
Not entirely correct.

Windows XP can't address all 4GB, however other 32bit OSs can.

The problem here is that a single 32bit app can only address a maximum of 2GB of memory which is why IBT fails immediately when you try to go over 2GB.

If you want to do a full memory test with IBT you will need 64bit windows.
Click to expand...

no, windows XP can adress exactly 4.0GB of memory (2^32/1024^3), like all 32bit OS's, but vga ram, L2 and L1 cache, hdd caches, and various other (some virtual) caches consume a couple hundred MB, as such you can see anywhere between 3.0 and 3.8GB of your motherboard ram.

There is no such thing as a 32bit app. X86, which is a standard all hardware in PC's adhere to, it states that the bus between the northbridge and the system memory should be 32bits wide. The updated version of this standard (aka X64) states that this bus should be 64bits wide.

Technically if programmers did everything by the book using windows memory management then any program designed before X64 was round would work fine on a 64bit operating system, but because of little wierd programming tricks used by alot of programs, unless the program identifies itself as 64bit aware, windows 64 runs it in a (very effective) emulator.

back on track:
With a Q6600 its harder to push the memory to the breaking point than the CPU, but theres an easier way to test it than to run hours and hours of testing programs, simply go into bios and change your memory:FSB ratio to something that runs the memory at a lower speed. Changing said ratio to 1:1 and turning the ram to cas 5 means at no point should you have to worry about your ram unless your setting record breaking OC's
 
Also your video card RAM is mapped on the first 4 GB which is where most of your RAM is lost usually.
 
Deimos said:
Not entirely correct.

Windows XP can't address all 4GB, however other 32bit OSs can.

The problem here is that a single 32bit app can only address a maximum of 2GB of memory which is why IBT fails immediately when you try to go over 2GB.

If you want to do a full memory test with IBT you will need 64bit windows.
Click to expand...

Windows addresses all 4 GB of space. 2 GB for the kernel and 2 GB for applications.

Also as noted in the SS above, IBT self-limits itself within the 2 GB limitation for Windows 32-bit.
 
Well I took the second 2gb stick out and it ran the maximum stress test 5 times with no problem at the same voltage. Barely got above 52*C. It's running 15 times now.

So there's that :) Keep in mind it also has to play with 896mb of video ram as well. And, as noted before, my ram is timed manually at 5-4-4-18, 2.1v (factory recommendation) and at its current ratio is running at 400mhz.
 
So it passed 15 tests with no problems. Played a bunch of CoD 4 last night, watched The Departed (xvid decoder ;)) played GRiD for a couple hours today... temperatures all stayed in the 40's according to the coretemp logs. What determines the longevity of a chip? Assuming I don't encounter any further stability problems, if it keeps operating at a safe temperature like it is now, is there any reason it shouldn't last as long as a stock chip?


I do apologize to those of you with small screens
2eq50qw.jpg
 
a chips lifespan is determined by any defect in the doping of the silicon, its suseptibillity to electron migration and jupiter and saturn lineing up, stuff like that, beyond your control. The general wisdome is that Conroe (and therefore kentsfield) is perfectly fine up to 1.5V.
 
You must log in or register to reply here.
Back
Top