http://i447.photobucket.com/albums/q...P32_Ob12_2.jpg
Printable View
OCZ3P18002GK
--------------
Overall I'm quite pleased with the way these sticks have worked out - apart from being 2N only and seeming to need quite high VDIMM and VNB to stabilise, they've not caused any trouble. For a cheap kit (mine were £75 or $135 approx), just to get up and running with on this board, I think they were pretty good. For what it's worth, I'm not sure I could have hit 1000Mhz any cheaper.
However, when pushed (a bit) both the NB and the memory needed skews to remain stable, and a lot of time since getting the RE has been spent finding and testing for the 'right' DRAM Skews for my kit - even for these modest clocks and timings.
With a 'better' kit, I would also expect decent results (eventually) from finding the best skews, so the last six weeks spent staring at the blue MemTest+86 screen feels like time well spent :ROTF:
1000 / 8-8-8-24 is as far as I think I'll get them without going over 2V, and memory performance (with PL7) is about the same as 900 / 7-7-7-20 (with PL6), which also needs a notch under 2V. Neither is quite what I'd like for a good 24/7, so I've decided to retire the OCZ and try another kit. Over 2V, I don't think any gains would be significant, and no doubt I wouldn't stop until I fried them. They'll do as decent backup sticks now :)
Unfortunately it's just down to trial and error to find the best combination for individual kit/board/clocks. Over time patterns can emerge for 'good' skews, but to make sense of it, you need to keep notes, change one thing at a time, and test and retest.
Found Memtest86+ very useful for DRAM skew testing and came up with this simple 'method':
If you run a deliberately low VDIMM, so that you get freezes in Test 3, you can see (quite) quickly which skews have the most benefit - for given clock and timings, certain skew combinations will pass multiple Test 3 runs, while others freeze instantly. I found the results were repeatable, and with the correct VDIMM (and VNB etc) applied later, memory was stable for full Memtest runs and then Prime etc, where it hadn't been before.
Trying, say, 25 skew combinations, I might get four which appear good and don't freeze - these tended to follow a pattern. The good skews all had the same offset between A and B, and differed only in the overall Advance or Delay from Normal (example: A2 DEL 25 / B2 Normal might work similarly to A2 DEL 75 / B2 DEL 50). At this point you then have a few 'good' skews to test properly and find which works best.
Don't know if this is really meaningful but it worked well for me. Adjusting CTL Skews gave me best results and I left all the others on AUTO.
Also, as setting DRAM SKEWS requires a full restart (IE your hard disks go off and on), I disconnect mine since Memtest86+ runs off a floppy or USB stick. That way you save your disks plenty of unnecessary spin ups :)
It may be apples and oranges, but it sure seems like the Intel chipset is working towards/obtaining that same objective much more efficiently -- unless those probes are way off. If the latter is true, why bother with probes? If they are that inaccurate they do more harm than good. Speaking of which, where exactly are the probes that provide CPU temp measurements on the 790i boards (say, using Everest UE)?Quote:
Originally Posted by 3oh6
I think most are finding this board has problems with RAM compatibility, period. LOL. One of the first kits we found to work well on the board was the G.Skill 2x2GB PC3-12800. I've had good luck with Cell Shock Blue, but others can't get the board to boot at any clock speed with this RAM. I think this will improve with subsequent BIOS releases, but 0601 didn't seem to offer much relief.Quote:
Originally Posted by Bullshooter
Just got these today :D
2x1GB Transcend aXeRam 1800MHz PC3-14400 (8-8-8-24) XMP (TX1800KLU-2GK)
http://i447.photobucket.com/albums/q...Transcend1.jpg
No idea what ICs they are and don't know if they'll even POST, but hopefully...
Intel burn no problems, Vantage no problem.
Super PI Mod 1.5 passes all the to 16M (16M also fails sometimes) but on 32M gets either a total system freeze or program stopped working appcrash message, the most loops I made it to was 14.
Any suggestions? (Vista 64)
http://i512.photobucket.com/albums/t.../Capture-1.jpg
What Vdimm and Vnb?
Freezes are NB related, try more volts on that.
32M failing is mem unstable 95% of the time. Does it pass HCI memtest?
Thanks for helping,Quote:
Originally Posted by Zeus
I raised the NB voltage to 1.643 -- still fails :eek:
"Does it pass HCI memtest?" I just down loaded this test and will run it later, but I did run Memtest86+ V2.01 and passed.
Also I have another observation when looking at CPU-Z 1.47 I’m seeing a slight fluctuation in the core speed field between 4199.9 and 4200.2 and in the rated fsb field I’m seeing movement between 2099.8 – 2100.1 is this normal or should this be rock solid?
Here are all my current settings, do you see any glairing problems?
Current settings:
CPU Type: E8600
Cooling: H2O in the CPU, GPU and NB
OS: Vista Ultimate 64
Graphics card(s): One Evga GTX 280 over clocked: 710, 1458, 1280
Memory Brand and speed: Corsair: TW3X4G1600C9DHX 9-9-9-28 @ 1.9volts
Tweaker Menu (Bios version 0601)
Ai Overclock Tuner: Manual
OC From CPU Level Up: Auto
OC From Memory Level Up: Auto
FSB Frequency: 525
CPU Ratio Setting: 8
>> = sub menu
>>CPU Configuration
>>CPU Ratio Setting: 8
>>C1E Support: Disabled
>>CPU TM Function: Enabled
>>Max CPUID Value Limit: Disabled
>> Vanderpool Technology: Disabled
>> Core Multi-Processor: Enable
>> Execute Disable Bit: Enabled
CPU Clock Skew: Auto
NB Clock Skew: Auto
FSB Strap to North Bridge: Auto
PCIE Frequency: 100
DRAM Frequency: DDR3-1681
DRAM Command Rate: 1n
DRAM Timing Control: Manual
1st Information: 9-9-9-24-4-60-8-4
CAS# Latency: 9
RAS# to CAS# Delay: 9
RAS# PRE: 9
RAS# ACT Time: 24
RAS# To RAS# Delay: Auto
REF Cycle Time: Auto
WRITE Recovery Time: Auto
READ to Pre Time: Auto
2nd Information: 10-4-5-4-7-4-7
READ to WRITE Delay (S/D): Auto
WRITE to READ Delay (S): Auto
WRITE to READ Delay (D): Auto
READ to READ (S): Auto
READ to READ (D): Auto
WRITE to WRITE (S): Auto
WRITE to WRITE (D): Auto
3RD Information: 19-6-1-10-10
WRITE to PRE Delay: Auto
READ to PRE Delay: Auto
PRE to PRE Delay: Auto
ALL PRE to ACT Delay: Auto
ALL PRE to REF Delay: Auto
DRAM Static Read Control: Auto
DRAM Dynamic Write Control: Auto
>> = sub menu
>>DRAM Skew Control
>>DRAM CMD Skew on Channel A: Auto
>>DRAM CLK Skew on DIMM A1: Auto
>>DRAM CLK Skew on DIMM A2: Auto
>>DRAM CTL Skew on DIMM A1: Auto
>>DRAM CTL Skew on DIMM A2: Auto
>>DRAM CMD Skew on Channel B: Auto
>>DRAM CLK Skew on DIMM B1: Auto
>>DRAM CLK Skew on DIMM B2: Auto
>>DRAM CTL Skew on DIMM B1: Auto
>>DRAM CTL Skew on DIMM B2: Auto
Ai Clock Twister: lighter
Ai Transaction Booster: Auto
EPU II Phase Control: Auto
CPU Voltage: 1.45
Load-Line Calibration: Disabled
CPU PPL Voltage: 1.51106
FSB Termination Voltage: Auto
CPU GTLVerf (0): Auto
CPU GTLVerf (1): Auto
CPU GTLVerf (2): Auto
CPU GTLVerf (3): Auto
NB GTLVerf: Auto
North Bridge Voltage: 1.643
DRAM Voltage: 1.89511
NB DDRVref: Auto
DDR3 ChannelA Vref: Auto
DDR3 ChannelB Vref: Auto
South Bridge 1.5 Voltage: Auto
South Bridge 1.05 Voltage: Auto
CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
Under advanced menu:
North Bridge Chipset Configuration
Memory Remap Feature: Enabled
Initiate Graphic Adapter: PEG/PCI
PEG Port Control: Auto
Still testing, but a few results for my QX9650. Would be nice to see some more quadcore settings/results in this thread
First off 410x10. Seems a pretty good, stable and cool OC. With CPU/NB/GPU in one loop with the PA 120.3 temps in blend stayed between 48-54c over the 2 hour period. Ambients of about 26-28c
http://img139.imageshack.us/img139/2...d410x10mo3.jpg
Note Vcore at 1.53v = about 1.48-9v in CPU-Z idle. 1.41v @ load.
http://img363.imageshack.us/img363/803/410x9biosgj0.jpg
Moving on to 460x9. Need to do some stability tests. A few thoughts though. cheapseats comments about using low vdimm in memtests is spot on. Certainly the way I've worked for a while. A quicker way to see what's working or not.
Next and I think I've mentioned this, but findings so far it may well be advisable to look at DDR3 Chanel A and B Vrefs, before struggling with clock and dram skews.
At 460x9 with Mem at 1843mhz 8-7-6 memtest would freeze at the opening screen. just the title top left. Sticking +25mv into DDR3 Chanel A and B Vref, and memtest is happy as Larry.
http://img376.imageshack.us/img376/5...t460x10ky7.jpg
Note these aren't really tested yet. Will Blend and OCCT
http://img56.imageshack.us/img56/8607/460x9biosku5.jpg
Oh and one last one. Will have to work at this a bit, but shows promise. Unfortunately 3dmark 05 blanked and crashed mid Canyon test.
http://img357.imageshack.us/img357/4...est4500tt3.jpg
RLM
A little chillier in Chicago so I tried booting up at 500x9 (4.500 GHz) and she is stable.
But my ram cannot run at 2000 do I changed the divider and ran it at 1668 with 6-6-6-15 1N Timings.
http://service.futuremark.com/compare?3dm06=8396690
http://img177.imageshack.us/img177/8021/23905t2qb6.jpg
I gained a whole 6 3DMark06 points to 23,911.
I moved the ram to the A2 and B2 slots and skewed them differently.
Got them to DDR3-2000 with 10-9-9-27 2N timings.
Barely Beats out the Ram running at DDR3-1666 with 6-6-6-15 1N Timings.
Been trying to hit 24K but No luck, unless I push up the voltages. And Do Not want to do that. The sun has been shining and the house is warming up. Wait till December and it is Ice Cold in here, may hit 25K then is the water in the system isn't frozen.
I think I'm going to EZ-Flash to 0601 and see if that may squeeze out another 100 points.
http://service.futuremark.com/compare?3dm06=8399358
Well I flashed to 0601 and it's only a little better.
Didn't help in 3DMark06 but it shaves almost 0.10 seconds from SuperPi 1M.
4.50GHz is done in 10.250 seconds now.
Good progress RLM! I've settled on my current setting for some games and get some value from this board back... will start more tweaking no doubt very soon :D
Cheers man:D Look forward to seeing how you get on.Quote:
Good progress RLM! I've settled on my current setting for some games and get some value from this board back... will start more tweaking no doubt very soon
Kensek, I saw your thread. I misread first time and thought you were using a chiller, not it's chillier.lol
That's very impressive. Bl00dy hell. No way I could achieve 4.5ghz on H2o for a 24/7. Those Q9650 are doing very nicely. A 10x multi would be nice though:D
RLM
@RLM
It's just chilly cold Chicago air cooling my regular water loop going through a ThermalChill PA120.3.
The Temp Forecast for tonight is a Low of 36F.
It's too early to start snowing!
Nice stuff Kensek!!!
I might have to consider one of those Q9650's one of these days.
Okay an OCCT run. Temps are a bit higher than I would have liked but then OCCT isn't app of choice:D
Bios settings as above, except had to bump vcore from 1.537 to 1.55v
http://img395.imageshack.us/img395/7...t460x10xv4.jpg
I know what you mean, good aren't they?Quote:
Nice stuff Kensek!!!
I might have to consider one of those Q9650's one of these days.
RLM
Yeah!Quote:
Originally Posted by RLM
I still haven't plumbed the depths of my E8600 yet, and the price on the Q9650 needs to come down a bit (which I figure it will with i7 on the cusp of release) then I'll be on the lookout.
RLM, I see that you have your vCore quite high but after droop it is not that bad. Do you leave it like that for 24/7? I now have my RE set up with the same cpu, but on air. I haven't begun pushing things much at all as of yet, just running 9.5x400 @ 1.3v. Have you tried a lower vcore with LLC?
Alright, I've been searching but can't seem to find.
What blocks work on the SB with Crossfire?
On my Maximus the Ek NB/SB Asus 4 was fine. Will it still work on this board with CrossfireX 4870 X2 with EK blocks?
and here is what I can test
http://www.abload.de/img/prime390015stunden-1f35.jpg
Rampage Extreme + CSX Diablo 2x1GB + E8600 seem to love each other :) :
According to EK the EK NB/SB Asus 1 will fit the southbridgeQuote:
Originally Posted by zlojack
For the list of compatable blocks try here...
http://www.ekwaterblocks.com/downloa...et_04-2008.pdf
I've been there.Quote:
Originally Posted by rddave
I'm pretty sure the mounting pattern is the same for the two blocks. My concern is more for the fittings when the block is mounted, and whether they clear the cards in Crossfire.
I've been communicating with Eddie and will post the confirmation one way or the other here so people know.
Samsung IC's in there .Quote:
Originally Posted by cheapseats
Very nice Zeus -Quote:
Originally Posted by Zeus
Let us see how you have your bios set for this run. Good sp1 score:up:
Hey Guys!! Some very nice numbers being posted. Congrats to all. Me, I've been planting a 6 meter tree in the front yard and getting ready to put plinths (base boards to us Canucks) in the rest of the house. Everybody should build a house once..after that you'll know what I mean!!:rofl: Especially in a foreign country where you don't speak the lingo.:eek: Anyway, its late..and thought I would have a play. So hooked up the phase for the first time, then decided to cool the NB & Chipsets with Wa Wa Tusu and see what happens.
Here's a question for you... When under phase is the only place you can see the real temp of the cpu in the bios?? The last time I rebooted it didn't even read correctly, said 39 C instead of the -49.5 it said earlier. Core Temp and Real temp just report one core @ 24C & the other @ 37C. The phase unit is hooked up to the CPU fan connection. If it wasn't so late I'd be having a BIG COLD DRINK right now as this 5 ghz. has been on my mind for a very long time!! First try and YAHOOZ:party::party:!!:D