Printable View
Quote:
Originally Posted by ChrissTi
it dosent work if tested it several times. It makes no diffrence if i enable or disable LLC, i get the same vdroop. It dosent work at 3ghz (stock) nor at 3,6 or 4ghz.
Again, who cares about vdrop? It means jack:banana::banana::banana::banana: if you set 1,34V and get 1,32V idel in windows....
Asus can deliver a working solution gigabyte not, period.
Got everything set and started OC'ing. Got it stable at 362x9 @ 1.45V with turbo on (had some memtest errors with ehanced on so I later turned that off). Memory at 2.66D with CPU-Z reporting 5-5-5-15 @2.1V. Passed memtest and got what I think is a decent score of 14,800 in 3DMark06 with a little OC to the 8800 GTS 512MB video card. :yepp:
I did get 1 error during one of my benchmarks and feel it must be related to the video OC, since the system seemed pretty stable before I OC the video card. I have not increased the voltage to the PCI-E control. Does anyone here have a similar video card, OC'd? Would be curious what kind of voltage would be safe to add.:shrug:
Thanks!
panzer948: Try running orthos for one hour, to see the stability of the system. Other benches may work but orthos will realy test the maximum the cpu and memory can give.
Hornet331: What mobo and cpu do U have and at what freq is the CPU (overclocked)?
What's the highest FSB you fellows have seen with X38 on a 9x multi using a Q6600 G-0?
Happy New Year ! :) :toast: :banana: :party: :bday2: :cheer:
Quote:
Originally Posted by ChrissTi
Happy New Year as well :D
I just ran Prime95 (already had that installed but havent run it yet) instead of orthos. Ran for over 30 minutes with no errors, even with my video card OC'd at 785/1961/1098. So I'd say that was pretty stable. Guess I don't need to up the voltage in my PCI-e.
But I think I can do better with my MB OC. I couldn't get it stable at 400x9 and I think it is because I don't know how to OC the memory.
panzer948: at 400x9 rise the cpu voltage, and mem set to 2.0B(800) or 2.4B (1066) multiplier factor, and the rest auto. Memory overclock it's a bit trickyer.
well since it is the x38-dq6 thread i have a dq6 (thought its the ddr3 version).Quote:
Originally Posted by ChrissTi
cpu is a Qx9650 but that shouldnt affact it, frequencies: i have tested several, ranging from 3,2 to 4ghz and droop occurred every time.
mine highest stable was 425x9:yepp: , but could get 445 x 9 but to pass prime took to much:D (volts and temps):shrug: . with an 8 i was able to pass with 475:yepp: . good luck and have a good new years:up:Quote:
Originally Posted by Blacklash
Hornet331: I have a bad news about ASUS which you susteained that it doesn't have vdrops. Search the net on the forums, there are lots of people complaining about that :D. Vdrop rises with bus frequecy, but rises exponentialy, like the power need by the CPU.
http://www.anandtech.com/cpuchipsets...spx?i=3184&p=5
Till a certain freq vdrop is very small, and after that it grows very fast.
Vdrop can be controlled by bios by increasing the voltage of the cpu when under load. Gigabyte is working on that and in the F7H bios implemented a voltage control of the CPU, and it's activated by pressing CTRL+F1 after entering bios.
U will find them under MB Intelligent Tweaker (M.I.T.) menu
CPU GTLREF1 Voltage Ratio [Normal]
CPU GTLREF2 Voltage Ratio [Normal]
For now there is not much documentation about it but ... we'll wait and see when the final release of the bios will be released :)
uhn... for the 10th time... i dont care about vdrop, i care about vdroop.
This two drops are completely different...
vdrop is when the voltage is lower than its selected in bios and is different on every board.
vdroop is a a specification from intel, and the voltage drops while the cpu is under load. (normaly it protects the cpu from voltage overshoots while the cpu switches from load to idel).
But when i am benching i dont need that feature, and asus allowes you to disable this feature and it works (no or nearly no droping of the voltage under load). Gigabyte just provided a fake option, just to look not as bad. :rofl:
stop sounding like a gigabyte marketing monkey who trys to blandish a missing/borken feature.... :down:
Uhh, Hornet..
I suggest leaving this thread alone if all you are going to do is bash the board. All you do is shove the word "Asus" in our faces, and we don't quite care for it. Either start being productive towards the thread, or a take a hike and preach your Asus fanboy-ism elsewhere. :slap:
Hornet that stupid name invented by intel is just a name, the technology exists for a long time in other electronic equipment, just with other schematics.
and read here:
http://www.thetechrepository.com/showthread.php?t=126
end of vdrop or vdroop subject.
good point LiViD I worked 4 years with A... :upset:, and it's a big s..
I thought they fixed something and I even bought one. I did a big mistake is still a s.. Don't want to say anything more.
Here are some screenshots, of a benchmark and the mem settings for 2 completely different sets of mem modules on X38-DQ6. next will be some bios settings.
Have you considered trying F3 to see if you can get some higher FSB/tighter timing combinations? I couldn't do squat with the F5/F6/F7 BIOS'.
I'm experiencing a difficult to pinpoint problem someone here might possibly know something about.
It's simple: when I overclock the bus to any great extent, over maybe 350mhz, the whole thing locks up within 2 minutes of running Memtest or Prime95, mouse and all. With the multiplier dropped down a notch, it still freezes too. I can reproduce the problem exactly no matter what settings I play with - lower the bus some, drop down to 800 5-5-5 memory speed etc. Very important to note is this happened when I put in new ram I got back from RMA, some G.Skill "HZ" - a set of two 1GB sticks. I'm not sure it's the ram, though it may well be mainly because I can't bring any errors up in Memtest or Prime for that matter, only total freezes, which normally (for me) indicates a FSB too high or memory running too fast.
Strangely, by adding +0.150 to MCH, I can make Memtest at least seemingly stable (Memtest and Prime95 for quad cores are what I'm using to test stability).
My machine is in my sig and I'm running F3 bios but I just changed it an hour ago, coming from F4. I havn't tried any other bios' since this started happening. My settings are standard really, locked PCI-E bus, ram running 800 MHZ 5-5-5, +.4v to ram. I'm really stumped with this one but I suspect it's the ram at this point. Anyone have any ideas ?
There have been some instances where users with quad-cores have better success using the newer BIOS' than the older ones. My rule of thumb is that if you are using a Quad, run F5 or higher. If you are running a dual-core, stick with F3.
Hi happy new year everyone.
LiViD is right frostbyte F5G or H are far better than the earlier bios for the quad.
Get Felix's Memset
Assuming your subtimings are set to Auto run Memset to Bios translation What Performance Enhance : setting have you ??
try a lower setting.
Or note the Performance Level timing, enter bios & manually try a higher value.
luck:)
The older bioses have problems in detecting all my hdd drives. :(
F7H is the only one till now which doesn't have that problem. with this bios also, if you have done bad settings, or your mem modules are not good when overclocked, the mobo does not start. You can oveclock but with good memory, like one that spuports 3 3 3 8 at 800 like Team Group TXDD2048M800HC3DC.
Could not rise more the FSB because of the CPU Temps ... under full load, with orthos, it may reach 72 celsius degrees... I don't have a Quad yet. QX9650 is too expensive, maybe if it will drop the price alittle I'll buy it, but for now waiting for a Q9550.
Frostybyte Try F7H bios, CPU bus 400, Standard Performance, mem 2.0B, Auto timings. If you set 2.4B rise MCH first at 0.100. If it doesn't work rise the CPU voltage slowly till you get it stable.
Happy New Year everybody! :)Quote:
Originally Posted by ChrissTi
ChrissTi, do you know what's the DIMM Clock Fine Delay for each memory module? you have them 6-7-7-7 and I have them 7-7-8-7 ('though DIMM3 and 4 are empty). I have G.Skill's 8000PQ (2x2Gb).
:rofl: i bought nothing but gigabyte boards the last 5-6 years, can you say that for yourself? (so fanyboy my ass... :slapass: )Quote:
Originally Posted by LiViD
I like the board, but it lacks features other competitor (Asus, DFI) offer and it they work... i dont pay 200€+ for a board and than only get half of they joy. :rofl:
But i think it fits Gigabyte, just looks the 965-DQ6 (rev1) oc issue with quads (65nm)... it was advertised as quad compatible... yeah thats about it, it is "compatible". :ROTF:
nice that you have found an articel that explains it for you, but you still dont get the point that vdrop and vdroop are totaly different thing and have nothing to do with each other.Quote:
Originally Posted by ChrissTi
Its also nice that this articel form fcg explains how you can calculate the exact vdroop, but im not interessted in this. I want a option to compensate vdroop. :yepp:
I don't know they were automatically set up by bios, Have U tested your system with Orthos and Everset open to see what your CPU temps are, and if is stable ?
I want F7j bios. Does any of you know where do I get it from ?
Hornet331: I've worked in a computer service for 4 years, so I think I know what I'm saying. ASUS many RMAs ... Dam there were entire lots I have to RMA. DFI is really good, I agree with you.
When I worked in service I liked more ABIT, they were more stable, even Epox and Gigabyte were more stable then ASUS :D.
Orthos not yet. I've tried Everest stress test and OCCT. they both crash after like 5 minutes or so, but I know it's the vcore because the CPU is with the stock cooler and I don't wanna push thermal resistance too far. currently it's set at 1.39 (everest) and it's very stable on 24/07, photoshopping and normal everyday tasks. Even games run without problems, only stress tests bring it down (possibly because of vdroop dropping to 1.34v under load...).Quote:
Originally Posted by ChrissTi