Printable View
no problem, for whatever reason these Venice runs much cooler than the Winchester and they can take loads of vcore, my under load temperature is only 42'C running 317x9(2.85Ghz) @1.71v, though only required 1.58v @2.80Ghz :toast:Quote:
Originally Posted by HousERaT
I see you're running your LDT at 1.2v and your NB at 1.5v..... did you have to up those to get stable? I'm still running mine at the lowest and I don't know if raising them will help much?
raising them didn't help any at all, it didn't hurt either but if you feel more comfortable to have extra juice going through them you can bump up one notch on each parameter, 1.30v for LDT and 1.60v for NB. ;)Quote:
Originally Posted by HousERaT
first of all im on air.
I hit 2808 but its unstable. Well I did 250X12 on accident. It lasted like 1 second. Of course that was no vltage mods.
Can bench with 2786 or something like that. Lowest 1mb pi is at that and was 31.458. That may be 2800. Durnign benching temp goes to about 48c. Folding right now its at 50c, i think the voltage is till at what I had it set to bench so its to high. I fold at 2700.
Fans, 1 120, 1 80 and the vent on the side that lets my heatsink, the thermaltake one, draw in room temp air. CAse sits at 29-30 c. These are acording to uclock on my an8.
Quote:
Originally Posted by H2OGun
I assume this is water cooling?
That is incredible! Wonder which cooling it is also. And the guy who made 32M with 2.6ghz 3000+, could you run it with less vcore?Quote:
Originally Posted by GoodOmens
howdy. got my 3200 oem ADA3200 AAA4BP
LBBLE 0516EPGW
from monarch on monday. swapped out my 3200 winny and tossed it in at default just to see if it worked then rebooted to 9x300 @1.6v, ok, but a little shaky so bumped it up to 1.66 and then ok. so wtf, lets go for 9x320, nfg, reboots at desktop. bump it up to 1.73, same again. take my ram out of the equation using dividers, try again still nfg. anyway, chip maxes out at around 305x9 stable, 1.66v (check out cpu temp in screenshot).
i am reading the venice oc thread and seeing guys with 3000+ getting better results than me with much lower vcore. thats the way it goes. but, at post #32, i see a guy with same chip as me, serial number less than 100 from mine,
getting much better results. kind of a big difference with same batch.
jumped on the 3200 nc s754, got 2700mhz. jumped on the 3200 winny, got 2700mhz. jumped on the 3200 venice, got 2700mhz. maybe a pattern here?
2700 is not bad, but after 2 months of hearing "dropped it in and 3gig!" i feel a little foolish. my bench results are about the same as the winny. thats ok. not supposed to be any significant performance increase. but what there was supposed to be was "much higher potential frequency". have a 3700 sandy in the mail. will demote this to my kids box. maybe.
http://www.flickr.com/photos/33178210@N00/12499540/
cpuz validation = 15508 (cant link to it somehow)
very truly yours,
politenessman
xp90 w/medium panaflo
x850xt
ultra d
2xraptor raid 0
2x256 g.skill la 440
ps as you can see i am noob at screenshots, etc. any help appreciated.
Same chip as you in SLI-DR using 4/14-3 BIOS and ocz rev2 TCCD. Won't go prime stable above 10x255 at 1.66v, haven't tested lower multis yet. Right now it is prime stable at 1.56v 10x250. Surpirsed it isn't scaling better with voltage.Quote:
Originally Posted by politenessman
Over at DFI street AngryGames just gave me a new BIOS that is supposed to help out the venice's and San Diego's...he was able to run well over 3ghz on air cooling with it. I'm probably gonna try it out and see if it helps, but I think this batch of venices isn't the best. Lots of tweakin still to do.
Quote:
Originally Posted by politenessman
ADA3200DAA4BP
LLBLE 0516 EPGW
Z639006D50906
That's me with the 3200 Venice that has the exact same stepping as yours, even my heatsink, mobo, ram, and hard drives are also identical...except the video card, I'm using the 6600GT instead. What PSU are you running them off of? and Which BIOS are you using?? Maybe it's the PSU that has been holding you back? though don't be surprise that these chips OC differently, couple months ago I bought 3 Winchesters from local store, all three of them have the exact same steppings even the last 4 serial numbers were very close, but none of them OC'ed the same, 1st one can only managed 2.4Ghz, 2nd one 2.7Ghz, and 3rd one only 2.55Ghz. My 3700 San Diego will be here tomorrow so I'll see how well it OC against this Venice :D
howdy.
i am using the antec 550 trucontrol, new version (30 amp 12v rail) using a 20/24 converter. also using 414-1 bios. swapped from 310-p bios with no discernable change. i had some slight suspicions about the psu, but it has adjustable rails - i set the 12v to 12.4v and the 5v to 5.3 and tried again, no difference. i realize that voltage readings taken from software do not show the current running throught the components, but only other psu i have around here is an enermax 465watt. i do have an evga 6600gt laying around here somewhere, i will dig it up and check it out. what dimm slots are you using? do you have the 5v dram enabled? what bios are you using?
very truly yours,
politenessman
ps my 3200 winny runs at slightly higher speed (2770 +), using lower vcore (1.55) and scores higher on sandra and evergreen mem tests at same htt. venice is lower power, better mem controller, and higher potential freq (stretch silicon, etc)
i must be doing something wrong here. wish i had another power supply.
this may be total noobishness, but has anyone heard of pci cards being a limiting factor in overclocking on a nf4?
very truly yours,
politenessman (grasping at straws)
It should not be - you can lock PCI bus to 33 MHz. To give you another data point I have a Venice chip same model & stepping as yours. I am using an Epox 9NDA3+ mobo and TCCD ram. I get essentially the same results with this chip versus my Winny - 2.5 GHZ stock, 2.6 with a little voltage, and 2.7 with a lot of voltage. The amazing part so far is how cool this chip runs - I let Stress Prime run for 2 hours at 2.5 GHZ/1.4 volts and the temp leveled off at 41-42 C ! This is 7-8 C cooler than my winny. I have an XP-120 with a 76 CFM fan. I will keep this chip in my main PC since it runs so much cooler. Like you I didn't get any more speed for my 200 bucks.Quote:
Originally Posted by politenessman
yea. mine is idling right now at 1.73v @ 83F. didnt even go above 100F during 2 hours 3dk05 loops @ 1.66. about 74F ambient down here now. 2 80mm in, 2 80mm out. nothing fast or loud. nice. sure would like a couple hundred more mhz tho.
very truly yours,
politenessman
ps really loved my 9nda3+. the 2.8v vdimm didnt matter (tccd). but after a month or so it would not warm restart so i rma'd to newegg (had bought the extended warranty, a must when buying epox). they sent me a refund claiming board had been discontinued. i dont know if this is true.
psps i checked bios. actually, you cannot lock the pci bus to 33mhz. it is just assumed, i guess.
Sorry, I should have explained - if your PCI-Express is set to 100 then your PCI which is derived from that is 33 MHz. That's weird about the Epox board - NewEgg has them available now!Quote:
Originally Posted by politenessman
http://www2.newegg.com/Product/Produ...82E16813123222
:confused:
Quote:
Originally Posted by politenessman
20/24 converter is not recommended to run these DFI ultra-D, many instability issue has been reported at dfi-street forums in the past so If I were you, that's the first bottleneck I would get rid of, get a decent 24pin PSU such as OCZ, a bit costly but worth the investment IMHO. I don't think the video card has anything to do with it, as for ram I use only the orange slots, you can run only 1 stick on slot 3 (the orange one further away from cpu), try 414-3 BIOS instead, then try again, before you go buy another PSU, what's your memory timing look like? get a A64 tweaker and post some screen shots of your timing and I'll see if I can help...
heres screenieQuote:
Originally Posted by Rubescen
http://www.flickr.com/photos/33178210@N00/12552907/
now i have been reading stuff on several forums re 20/24 pin psu. most agree that the spec says the extra 4 pins are strictly for powering pcie slot. other people swear to much higher cpu and gpu overclocks with 24 pin. now i am not fully convinced i need a 24 pin yet, but i do think the x850xt might sucking up some power i need for cpu. i have also seen a thread that explains why if you dont have pcie vid card plug from your psu, you need an adaptor to use 2 molex connectors to power the card (1 molex=72 watts, pcie vid card draws 100+, blablabla). what i get out this info is that i should see a difference when i get home and swap out the xt for the 6600 gt. any help you can give me tweaking the ram timings greatly appreciated. thanx
very truly yours,
politenessman
unplugged any unnecessary devices like USB printers, DVD Burner, Soundcard and all that craps, flash your bios to 414-3 if you haven't done so.Quote:
Originally Posted by politenessman
try these settings:
Trc 7
Trfc 14
Trcd 4
Tref use only 4708
Bank Interleave: Always Disable
Max Async Latency 9ns
Read Preamble 5ns
Dynamic idle cycle counter: disable(always, for stability)
drive and data strength set them to Auto
Disable your 32byte Granularity.
Your vmem is too low to run 300+ 1T, prime95 and superpi 32M will crapped out instantly if your memory doesn't has enough juice so go in the Bios set it to 2.90/enable and leave it alone for now.
Now, keep the HT ratio at 2.0 and don't change this, try 10 x280, and vcore 1.55v x110%, leave the vcore alone for now and don't touch it, go run the superpi 32m and if it passed, congrats...go back to the BIOS and bump straight up to 10 x285 and leave the rest of the settings alone. If you passed SuperPI 32M again, then try 9 x317, if superpi 32M failed, bump up vmem to 3.0/disable, but if you get BSOD, then the vcore need to be 1.55 x113%, try again. Once you get everything stable and if you want more memory bandwidth...let me know and I'll help you tighten up the rest of the setting, but for now stability is your main concern.
thanx man. i will do this thing now.
very truly yours,
politenessman
ok. these settings got me up to10x278 stable, 2.5/3/3/7 but only @ 1.71 vcore. anything less gives me constant reboots. i think thats all this chip has in it, although there are at least 3 guys here and at dfi street with same chip, from same place, very close serial numbers doing close to 2900 @ 1.55 vcore. power supply maybe? using truecontrol 550 (30amp 12v rail) with 20 to 24 converter. i cant think of anything else. temps are still great - 27c @ 1.71v idle, ambient around 22c. thanks for your help.
very truly yours,
politenessman
3ghz@29.5 sec ... Why ? your memory :slapass: :slapass: keeps you back
29 sec comes with 2.76~2.78 ghz and 2-2-2-5
Anyone ordered the retail 3200+ from monarch? Do post some results :D
Quote:
Originally Posted by Troman
SuperPI 1M 28 seconds?
i just saw h2ogun and thesaucier's 3200 and 3500 venices doing 2.9ghz at 1.5-.59 vcore.... jesus they landed on chips better then the ones cherry picked for review sites to use.
Hey dudes, just got my OEM 3500 from Monarch:
LBBLE 0515FPAW
Also seems everyones getting LBBLE stepping and around post 0515 packing dates right? I'm dismanteling my old P4 2.4c setup as I write this. :clap:
You need to use the above vid control as well to get higher vcore.Quote:
Originally Posted by av0