DFI 965-S is here! For sale @ ewiz

[RealTelstar]

hey can someone explain the jumpers between the pci slots to me? what do they do?

As written on the manual, those jumpers are to set the third pci-e slot to either 4x or 1x.

[XtremeTiramisu]

x-tiramisu, I just flashed using winflash last night. I'm on 2k3 and not vista though, so I can't vouch for its safety. That said, make sure you check flash all and clear cmos. just kind unnerving doing a flashall, but dfi makes it sound like it can brick otherwise.


revv23, I believe they may be to configure having a bunch of x1's vs having an x4 in the x16 physical. I could be wrong though


calcu, have you tried using 6x multi and seeing if you could go higher? Personally, I have an e6400 and just went through a night of 8x438=3500 totally blend stable, as it should be. 7x500 still has issues and crashes blend within a minute, despite settings.

Thanks for your reply,
So for winflash, you just download the BIOS with its associated files (I think 3 of them right? .bin and 2 others?) and put them in the winflash folder?

Also, do I set my HDD as IDE in the main BIOS configuration page since I only have 1 harddrive? There is AHCI or something...should I set it to this or IDE but I know it's not RAID for sure.

I can dissable all RAID configurations like Jamicron and related right?

[ziddey]

if you're using vista, you could easily set them to ahci. no real difference though, although you have a sata dvdrw, so you'd need bios 424 to be able to still boot from optical

that said, there's no way to toggle a boot menu during post?? that's terrible. and there's no way to further skip memory testing even in quick post mode?

[ziddey]

hmm, at 500, I can pass test5 on loop for ever. shows nice 6028mb/s too. but blen will crash within a minute guaranteed. friggin

[Praz]

that said, there's no way to toggle a boot menu during post?? that's terrible. and there's no way to further skip memory testing even in quick post mode?

Hitting Escape during memory count cancels it. Hitting escape during hard drive detection brings up the boot menu.

[ziddey]

i knew about the memory part

didnt know about the hdd part. thanks!

So I wonder what it is I should be going for now.

I've confirmed memory stability at 500fsb with tight timings and low volts. I've confirmed cpu stability in a different motherboard, and again in this one, albeit at higher multi and lower fsb.

But when I try 7x500, no matter how weak I have timings and if I turn off memory enhancements / increase voltages everywhere,

orthos blend will lock up the system within a minute.

The board does squeal like a real though. mosfets are staying cool it seems

[The Nemesis]

hmm, at 500, I can pass test5 on loop for ever. shows nice 6028mb/s too. but blen will crash within a minute guaranteed. friggin

I've been haviing trouble getting orthos to pass more than 2 hours & 38mins using a cpu & ram already proven on other platforms. It's interesting because this chip behaved the same way with the P5B Deluxe. I thought @ that time the chipset couldn't sustain FSB higher than 450 now I'm begining to wonder. Anyway my second DFI 965-S will be here tomorrow & I'll see if it clocks better or the same. I'm starting to think the stock cooler is good enough for high FSB 32mb Pi but not orthos. I'm gonna give prime 95 v2.51 a try tonight.

[ziddey]

I really wonder how high is too high for vmch. You see most people running around with values way higher than max listed by intel. By contrast, most people stay under max for c2d's. But I'm really tempted to just drive 1.7vmch to this thing and see if it changes anything.

I guess in the end, if I have to settle for 450fsb, I wouldn't mind it that much due to no performance drop post400fsb like the other boards. If I could somehow muster out 4:5 and get 1125 ram, then hey

edit: wow. 500fsb with 1.7vmch = blend locks up system in 2 seconds.

back to stock volts. seems like small fft is failing right away now

Test 1, 800000 Lucas-Lehmer iterations of M172031 using 8K FFT length.
FATAL ERROR: Resulting sum was -216745390374912, expected: 366073178927591.6
Hardware failure detected, consult stress.txt file.
Torture Test ran 0 minutes 2 seconds - 1 errors, 0 warnings.
Execution halted.

added vcore to no avail. hmm...

looks like that happens when i switch gtl to mode2.

but fwiw, i was finally able to get 6x500 to work today. not testing blend and it seems to be aok. could very well be heating mosfets after all. hmm.

edit: hmm. blend locked up after 10 minutes even at 3ghz and stock vcore. testing with mode3/mode2 for ddr/gtl now.
if it bombs out again, I might pull a stick and run single stick single channel for some tests

edit: hmm. using single stick in slot2 at the moment. 1.35vnb, 1.6vsb, 1.27vtt, 1.10vich. can't remember if it's auto/auto or mode3/mode2. but so far looking good. the fact that slot1 alone wasn't good but slot2 seems to be working ok seems to suggest that the finetune settings might indeed help.

[Calcu]

calcu, have you tried using 6x multi and seeing if you could go higher? Personally, I have an e6400 and just went through a night of 8x438=3500 totally blend stable, as it should be. 7x500 still has issues and crashes blend within a minute, despite settings.

Yeah, that was one of the first things I tried but it didn't change anything, so that's why I'm thinking it's the memory but then again, loosening timings & adding voltage didn't change anything either, so it may actually be the board. But I just borrowed another e6300 to test today, so we'll see how that goes,

[ziddey]

are you running two sticks of ram? if so, try one stick and move it around trying every slot. i seem to have found stability with slot2 and some bumps in votlages. trying to figure out which ones

[Peter949]

I was wondering which colored memory slots is preferred on this board. Any recommendations?

[ziddey]

people have said the white ones, but it could vary. I believe it all really just has to do with the fine tune settings. hopefully that'll allow me to run the 500+fsb i want. currently, i've tested single stick in slot2 (blue) with bump in voltage across the board and found stability. testing slot1 (white) under same conditions and seems to be going fine as well.

dual channel on 2+4 didn't work out well either

edit: anyone know where i could buy a spare eeprom from? i figure i'm going to be attempting some bios mods and will probably nuke it a few times

[The Nemesis]

edit: anyone know where i could buy a spare eeprom from?

http://badflash.com/

[ziddey]

Once again, I thank you. And I just wanted to make sure you got that email. Thanks again orthos ran 2.5 hrs with single stick and ramped volts. going to have to try two sticks again soon

edit: wow, just checked ebay. ton of services and chips provided. prices really good

[nealh]

http://badflash.com/

do they have one for the 965-S?

[ziddey]

send them the rom file and they'll do it up. or buy a blank eeprom. also on ebay, there's tons too. same deal. send the eeprom. but it's around $10 cheaper there

that said, dq calibration!!!!!! seems to be doing wonders. just wonder if it'd be like high drive strength for dimm and blow up sticks

[ziddey]

thumbs up for dq calibration. at 7, prime would fail instantly with rounding error 0.4 when should be 0.5. stock 4 would cause it to lock up instantly. same with 1. but 6, it's priming now for 30 minutes and going. the question i have now is, is dq calibration related to dimm drive strength, and is 6 higher than 4. if so, what would be a safe amount? i'm back down to 2.1vdimm, which was tested good in p5b, and it's seemingly fine. but high drive strength leads to early death, and i'm afraid of that.

i'll eventually try 5 and see if it works, but i have a pretty good feeling it won't.

[XtremeTiramisu]

Hitting Escape during memory count cancels it. Hitting escape during hard drive detection brings up the boot menu.

So will this method work if I use my SATA DVD optical to boot from the Initial bios? The thing is I would not be able to access winflash without a clean installed OS and without winflash, I can't flash bios to 0424beta from the initial BIOS so I can boot from a SATA optical drive.

I really hope that the ESC method works for the initial BIOS to force to boot from the boot menu...

[ziddey]

the problem with booting sata optical is only when you turn on ahci or raid mode.

so if you keep it on ide mode, you'll be fine.

that said, hitting esc wouldn't help you since in ahci/raid mode on 326, it wouldn't be able to use the optical properly to have it work anyway.

[Praz]

If you're talking about using the SATA optical drive with RAID on the ICH8R it isn't going to work without the 04/24 bios. The optical drive and and hard drives in non-RAID will function fine together.

[ziddey]

Woohoo. Alright, it's done deal

7x500 with 2x1gb d9gkx in dual channel. decent timings too, 4-4-4-9-4-28-2-11-9-2-8-6-6-5-4-4

fast fast memory enhancements

mode2 ddr, mode1 gtl

auto for the fine delay.

And the most important part: DQ Calibration: 6

Now I can have pretty much all volts at stock (1.25vmch, 1.2vtt, 1.05vich). However, I do have a problem. For some reason, I get the same issue by having vsb set to 1.5v. I MUST have it set to 1.6v. I never would have thought to change this value, but I noticed sometimes the drives I have on the jmicron weren't being detected at post, and I'd get a subsequent hang at the end of post.

So the two major pieces were 1.6vsb and dq6 and now I'm rolling . The board is so much less stressful to work with when you switch boot mode to mode1 too. when it needs to do mode2, it will, even when you have it set to 1. And so, I don't really know why anyone would want mode2.


So now I've got c1e enabled again, resulting in spectacular idle temps. It's sweet that dfi allows vid changing still, so at idle, it's 6x500=3000 with 1.32v. eliminates the droop issue where idle vcore is too high


edit: hmm, dq 6 was almost perfect, but failed small fft after 6 hours. did occt cpu+ram and locked up before the 30 minutes. Same deal the second time. Dropped dq to 5, and passed the test.

[XtremeTiramisu]

Thanks a lot for your generous help Ziddey and Praz.

Will I notice a significant performance increase/decrease from IDE to AHCI or AHCI to IDE for only one Seagate Burradica 7200.10 SATAII 16mb cache drive? Sorry, HDD configuration isn't really my experty..:P

Thanks in advance for any help guys

[ziddey]

nah. personally I run ahci just because, but it's more of a hassle than anything else. ahci will allow ncq to be on, which for what we use our computers for, will actualy hurt performance. oops

[M3kk]

Hello, i have a question...

What is the difference between this http://torent.ro/articol/29/3824/dfi...alc882_hd.html
and this
http://torent.ro/articol/29/3823/dfi...alc882_hd.html

?

[XtremeTiramisu]

Hey guys, I'm in the stage of dilemma here. I'm not sure to put the rams in DIMM 1&3 or 2&4?

I recall someone mentioned about overclock-ability or compatibility issue for certain DIMM configuration earlier in the thread but I can't remember clearly what it was.

Any suggestions?

*Nevermind: I saw the picture in the last page and the guy put his HyperX rams in the 2&4 slots. Thanks