Printable View
Well one kit will, the other is giving errors very early (4% in) at DDR3-1704 MHz 6-6-6-18 @ 1.65v. It does better at CL8, seems to hate CL6 and CL7. Second kit is much better but one stick not as good as other 2, so I have put the 2 best sticks of the kits I bought with the best stick of my kit and thats what I was using to do that test and the HCI Memtest run.
Just had a look at my old table. 1.65V set gave me 1.645V at one of the central VDD pins on the slot closest to the CPU. This was on one of the earlier BIOSes though.Quote:
Originally Posted by CryptiK
How much is the average clock with the Noctua NH-D14 cooler?
Im at 4 ghz now, and when i run Linx it gets hot...
How hot is too hot? 70? 80?
And what software do you use to measure the temp? RealTemp?
New toys :)
http://img197.imageshack.us/img197/5131/dsc00831xb.jpg
http://img847.imageshack.us/img847/3131/dsc00835u.jpg
http://img146.imageshack.us/img146/3852/dsc00840hi.jpg
http://img828.imageshack.us/img828/999/dsc00851if.jpg
Neither kit that I got will run at Cas 8, but all 24 Gb does 9-8-8-20-100-1T @ 1600 Mhz and 1.5v. They wont do any better with voltage increase at all.
http://img831.imageshack.us/img831/2061/32m.png
Completely overkill, and my geil modules did Cas 6 up to 1700 Mhz, but I couldnt refuse getting these 12 Gb kits at £119 each. I'm going to wait for the current G Skill 2133 Mhz kits priced at £50 to disappear or go up in price before selling my Geils though.
I can get a little better from them at 1.65v, 9-8-8-15-90-1T:
http://img847.imageshack.us/img847/6216/32m2.png
Cas 8 simply wont boot, even at 1.4v QPI and 1.7v DRAM.
this is my new pc i built last week but how come my gflops in linx higher when my HT off anyone could tell me why
HT off
http://img687.imageshack.us/img687/8...inbio1005h.jpg
HT on
http://img707.imageshack.us/img707/1...linbio1005.jpg
Awesome pics Cornell. I see that you have the 7D. I'm wondering what lens your using? Because no matter what even with all my "L" lens I can't get that type of sharpness. I'm assuming you edited your pics?Quote:
Originally Posted by Cornel
last i remember, something to do with LinX not utilizing HT properly and the work load being more efficient with HT turned off across all cores. i get the same thing, roughly mid 70's with HT on and about 95-100 with it off.Quote:
Originally Posted by lioncool
Real temp seems pretty reliable, Ive been using it for a while now,Quote:
Originally Posted by Cornel
As far as temps, the bios default for overheat is 90 so for safety I reduce that to 80, Ive seen ppl with a 4ghz o/c around the 70s with that cooler so aim to get around that under load.
I just got my 930 up to 4.5. I have the volts at 1.41 it's the lowest I could go to keep it stable. My temps after full load for 3 hours or so get to 77c . What you guys think? back it off a little or rock on?
Rock on man. That's a solid overclock!Quote:
Originally Posted by VengeancE
hey zoson, just so ya know i have a kit of 6gb ocz platinum lv 2000mhz 9-9-9-30@1.65v
They dont work very well with my i7 970 (GT) neither:down:.
with 2gb it can go 70Gflops in linx but <3gb and its like hitting a wall @ 54Gflops.
so, issue its not relative to 12gb.
they suck donkeybutt like yours:idea:
:p: no offense
according to intel 100c is top rated temp, 95 being your -grind teeth in fear- and 98 being -stop-. mid 80's to low 90's is about your safest HIGHEST possible temp... just don't push it if you don't feel comfertable.Quote:
Originally Posted by Cornel
Thats odd behaviour, whats your QPI Data Link Rate set to?Quote:
Originally Posted by Andreezy
i've tried diff blclk 200mhz and under and just about every qpi.Quote:
Originally Posted by CryptiK
i use the slowest speed besides slow mode.
its a dead case i pulled the sticks already.
they work good on other cpu, just not on gulftown...:shrug:
This is taken with an L series. How's it compare for sharpness? If the point at which focus is achieved is not in the sharpest focus when you review the photo, have you tested to see if your lens is front or back focusing and adjusted for this if so?Quote:
Originally Posted by Cookiesowns
http://i720.photobucket.com/albums/w...rrammacro1.jpg
Just did some 12GB testing - I feel your pain its insanely hard to do.Quote:
Originally Posted by Andreezy
Low gflops when all ram is selected. Think its most likely just the LinX software is unable to work properly rather than actual hardware issues.
you tested with 3 sticks 6GB and had normal results and now you saidQuote:
Originally Posted by CryptiK
you tested with 6 sticks 12GB and had low Gflops results?
Yes that's correct, went down to 57 GFLOPS with 12GB (6x2GB).
wow thats a surprice:shakes:
this one really does have me smh.:p:
its suposed to fly with more memory allocated.
did you try to uncore @2x
My uncore was and is usually more than 2x. I was running DDR3-1600 6-6-6-18 1T and 3600MHz uncore.
I was using a Canon 100mm L f2.8 IS Macro lens.Quote:
Originally Posted by Cookiesowns
Yes, i edit them, but not much, just adjusting some white balance, color temp, cropping etc.
Maybe we should start a new thread... "G33k p0rn" and post all our nice pics of our computer/parts.
This weekend I'm going to a friend's house who has a r3e, and has an i7 950. I'm taking my memory with me and we will test the GFlop issue with the same memory on same model motherboard, but with a different cpu.
I was also wondering if it was Linx causing the problem, but others don't seem to have the issue...
just so you know zoson, my 6gb kit, performs with no problems on bloomfield:), the problem is only gulftown hexa chip.
so, i assume your result will most likely be similar to mine.
on a side note, how do you rma memory which only problem is not being able
to scale with O'C Speeds??:shrug:
Cryptic, thats just amazing how this memory sticks are not scaling proper.:down:
Not sure if you are describing the same issue - when using 6GB, my GFLOPS are as normal. When using 12GB (of the same model ram) the GFLOPS are at a lower level.
You seem to be saying you have a single 6GB kit that when used alone is coinciding with low GFLOPS.
In all honesty I dont care what LinX says, it isnt coded the best and has always had problems with hexacore, also HT, now this weird lower GFLOPS issue some are seeing when using 12GB. I am almost 100% certain its just buggy software, and proper apps can and do make use of 12GB when its installed. Im only focused on real world performance I just use some synthetic tests to verify stability under high stress.
Verify your bandwidth with AIDA64 or everest, do some SPI32M runs, some Prime and some LinX (no matter how poorly the latter works) and you'll notice everything is fine except LinX. That doesnt indicate a problem with hardware - just that program. My bandwidth is good with 12GB, and everything else runs beautifully.
yes, its a single 6gb kit (3x 2gb).
i dont really want to test them again, i feel that the problem is in the memory itself, somehow its developing a bottleneck when moving data.
on my 6GB kit, the RTL are set very, very tight, dont know if maybe loosing them up may help isolate the problem.
sighs...i hate testing 12GB cause of the waiting time.
forget the post above, i turned HT OFF,ram set @ 1694mhz, uncore 2x @3388.
im getting 89GFlops
http://i121.photobucket.com/albums/o...h_89Gflops.png