Printable View
You sold your board because of FSB holes and because your CPU didnt like the mobo? Wow... but as long as the user is happy, everybody is happy :)Quote:
Originally Posted by webwilli
Well, I guess you can use CrossFire :) , So whats the amazing OC you achieved now?
:ROTF: Good Luck with that :rofl:Quote:
Originally Posted by webwilli
I have 0 issues & love my Striker II...:up: :up: :up:
I have completely opposite feelings ... 460 MHz FSB on a quad core, 4.14 GHz on air cooling, 22691 3dmark06 with one 9800GX2 and stock cooling, memory at 7-6-5-18T-1T@1840MHz linked sync ...Quote:
Originally Posted by webwilli
I am waiting for the WC set ... all voltages are in the "safe harbor" ... 1.35V VTT, 1.39V CPU, 2.0V memory, temps could be better, of course, and will be better with WC :yepp:
People who have no issues seem to be running 2gb.
Yes, but that can be said for overclocking on any mobo.Quote:
Originally Posted by Broly
Thats true, but most people running 4Gbs here are running 2x2Gb so I dont think there is an increase of stressing the NB
Apparently the 2x2's are massively OC'd
My patriots are JEDEC 1066 I think, and they're rated for 7-7-7-20 at 1600 MHZ which is a little too good to be true. I think I am going to sell these and buy the new Cellshock or dominators.
Can anybody verify that this corruption issues has something to do with SATA HDD instead of IDE HDD? Or the chances of corruption is higher on SATA HDD as compared to IDE HDD??
Grateful for ES, where we can warn folks when things like this go on.
:(Quote:
Originally Posted by webwilli
of course it makes a difference!Quote:
Originally Posted by slim142
whether the chipset has to drive 2 single sided or 2 double sided sticks makes a huge difference. its double the chips the memory controller has to drive... and the board has to power those chips too and keep vdimm and all the other voltages stable... the more chips, the more load on the vdimm circuicy...
i dont think the amount of memory makes THAT much of a difference when it comes to instability though... sure, more mem is more stress so higher chance the nb craps out... but it rather sounds to me like some boards clock great and are stable at high speeds, while others are not. there seems to be a big difference from one board to the other. which is exactly what happened with 680 btw.
from everything people who had a lot of boards told me there was a rough 1:2:2 ratio of 680 boards. 1 great, 2ok, 2 real bad.
it also happened when people used raid controllers...Quote:
Originally Posted by memory
but iirc yes, people seemed to have less corruption problems with IDE hdds...
[QUOTE=saaya;2956012]
of course it makes a difference!
whether the chipset has to drive 2 single sided or 2 double sided sticks makes a huge difference. its double the chips the memory controller has to drive... and the board has to power those chips too and keep vdimm and all the other voltages stable... the more chips, the more load on the vdimm circuicy...
i dont think the amount of memory makes THAT much of a difference when it comes to instability though... sure, more mem is more stress so higher chance the nb craps out... but it rather sounds to me like some boards clock great and are stable at high speeds, while others are not.
[QUOTE]
Sorry for not clarifying my statement. I meant what is in bold :up:
LOLz!Quote:
Originally Posted by Broly
Sorry but all DDR3 performance kits are just MASSIVELY OC'd Micron D9's :p:
My Patriots will do DDR3-2000 8-8-8-24 on just 2.0V just the same as any other leading kit only they cost less.
Around the EVGA forums Patriot has the fewest issues on the 790iU, Actually "NONE". Been reading all leading 790iU threads around the interwebs since launch and yours is the first Patriot issue I've read and I wouldn't count it as an issue since you seem to not understand DDR3 much in the first place.
Gskill has the most issues and OCZ comes in second.
EDIT: I just realized thats you Broly... ROFLOL... I still remember your tantrum on the EVGA forums followed by all that BS you tried to feed everyone about your tech knowledge....... should have known it'd be you making a comment like that :P: - No offense or anything..
Saaya
add strikerII nse ?? no corruption yet. :)
longsiew!! a lot of times due to GTLRef / VTT Ratio... try 0.584x for ref 0 and 1. i saw a ref mobo's VTT move and those spikes always gave me corruption. if u can monitor vtt throughout a period of time to see how stable it is...Quote:
Originally Posted by memory
I second this theory.. Ive had one case of corruption that was specific to my CRYSIS install a few days after I got the board I know the issue which triggered my corruption was caused by bad VTT/GTL. Since my own experience I've noted several times seeing posts around the web(here aswell) that people could pinpoint the exact time the corruption occured and those times were always after a "freeze-up" exactly the same way 790i's freeze when their VTT & GTLVREFs are bad.Quote:
Originally Posted by LardArse
i know, just do an experiment with variable resistors to change voltage on the fly you'll see straightaway.Quote:
Originally Posted by 5010
Is this for quad. Going by your reference to both 0 and 1, I assume that it's for quads. Also, I though it was .66 times the VTT is the reference. So if that's the case, do I have to go down on the GTL(-xxx mV)??Quote:
Originally Posted by LardArse
Heh, nice to have someone finally confirming my suspision. Far to many folks are having issues, 99% have more money than brains and just can't handle VTT/GTL theory, yet its so necessary w/790iU even at stock speeds the auto VTT & GTLVREFs are totally borked and just cant go unadjusted. Hence we have 1000's or cry-babies who cant or wont help themselves to solve the problems their having with the board posting around complaining everywhere and spreading false fixes w/ driver reinstalls etc etc.. A real mess!Quote:
I know, just do an experiment with variable resistors to change voltage on the fly you'll see straightaway.
.67x is 00mv(standard) in the 790iU BiosQuote:
Originally Posted by tekjunkie
65nm Quads like it abit higher; .68x - .70x
Duals + Yorkies like it lower; .66x - .63x or even less (as low as .58x according to LardArse)
Quote:
Originally Posted by 5010
Thanks. I went with what AT had in their S2E and always had it in the postive range for my QX9650.Unless I read it wrong, which is quite possible these days with me staying up all night playing around with my s2e, they had said that the quads require more than the reference,i.e, postive GTLs.It didn't do squat for my stability. As a matter of fact it was detrimental to the stability. I am SPI 32M stable @450FSB with my QX9650. I am gonna run prime now. I will prime with GTL @ auto and @ 0.63 times VTT and post back with results.
I had corruption issues with P03R2 BIOS before with the same 2x1Gb, it's not a point. The voltages idea looks more appropriate ... the auto correlation could be corrected in P04.Quote:
Originally Posted by slim142
All but the new and NYA Corsair 1800 4G kit which uses Samsungs.Quote:
Originally Posted by 5010
ref: http://www.xbitlabs.com/articles/mem...-1800-4gb.html
I would really like someone to get them (b4 me) and play with them on the 790i boards.
I decided to try the 790i after abandoning nvidia chipsets after the 680i problems and yes I did have corruption problems, in fact it led me to mac pro after years of building my own systems. I just got tired of all the incompatabilities and stuff just not working.
I have finally completed Striker II Ex rig (fully watercooled by two Swiftechs), Q9650, 2x Ultra in SLI, 4GB DDR3 1800MHz and Raid0 from Raptors and with BIOS 0601 in last week no problem ... no corruption, no issues ...
Great stable system on 450MHz (RAM 1800 - 20-7-7-7) CPU 4050MHz ...
dont you think this is simply based on coincidence and hence the more widespread a brand is the more people will report issues using this mem?Quote:
Originally Posted by 5010
ocz seems to be very widespread and gskill as well while patriot is used by very few people.
your the only person i know using one... and i bet its not a retail board, or is it? :DQuote:
Originally Posted by Rol-Co
yeah, the 790 problems all seem to have to do a lot with vtt.Quote:
Originally Posted by LardArse
that explains why some boards have this issue and some dont, and why some people report a board works fine with one cpu and doesnt with another. the problem is that 65 and 45 and quads and duals have different preference for vtt, and even the same type of cpu has different versions of cpus, some that like higher and some that like lower vtt...
ideally the board should set a certain vtt automatically depening on whether a cpu is recognized as dual or quad, and whether its 65nm or 45nm.
OBR, why do you keep posting that you dont have any issues?
first you said all was great, again and again, then you said you had instability issues too and didnt post for some time, and now your back posting again and again how everything is rock stable :P
instead of repeating how your system is stable, which doesnt help anybody, could you please post your detailed settings or maybe some tweaks you found to improve stability? that would actually help people... just telling people that have stability problems that you dont have stability problems... isnt really helping anybody :D
I assume that the value you calculate is then applied to both dies on the data bus..so why are there different values in the bios for REF0 and REF1. Using LardArse's .584 at 1.36VTT I get a -117mV adjustment, but this value is not available for REF1 adjustment (-120mV is the closest), although it is available for REF0. Am I missing something, or do I just get as close as possible to a value in between?Quote:
Originally Posted by 5010