A64 OC Database & Overclock Guide

the first line is unique to the model... which you didn't mention ;)
(basically, i know that a 3200+ venice will be ADA3200DAA4BP, i was just using the first line for those who aren't really sure, or things like the 3500+ san diego)


as for the ram thing.. 2T will be ~10-20mhz slower than 1T (meaning you'll nee 210mhz to 220mhz at same timings to beat 200mhz at 1T)
but, if you have stability issues.. bump up the ram voltage! :D
 
well, it's a dfi nf4ut board, I looked in the dram settings menu, but I didn't see anything that looked like command rate, the ram is running 2.9volts right now, so I guess I shouldn't worry about it too much, I just tend to like to run the ram as slow as possible when I'm doing the CPU. It's a 3000+ venice, I'd check the first line, but I think I finally got the insulation just right (or very close) by putting a thin open foam (same stuff the mobo sits on in the retail box) around the tubes going onto the fittings, though the middle one still has a little metal showing around the hexagonal part you would grip with a wrench. Anyways, I'll probably try some more and see what it will do, and what the condensation does, going to take a shower, then comes windows overclocking and hopefully some screen shots of stable speeds.

If I could do it all over again from the start, I would have just bought a radiator and a few panaflos, or I could still set this chiller up to cycle (but it'd take some effort, and usually the heating element would cycle, not the freon loop). Either way this thing was zero'd in at 2.7 without much effort, which for all intensive purposes is pretty good for a 1.8 ghz cpu, or average for a venice.

One more question about the DFI board/Nf4 chipset. What is a sufficient aftermarket passive chipset cooler for replacing the stock unit as it is loud as hell, especially compared to my 1 panaflo 80mm, 120mm seasonic psu, and eheim pump. I'm amazed by how quiet the seasonic is, the first time I fired it up I thought I'd left the fan unplugged. Much better than my old one (enermax).
 
ah, in the dfi, it's called command per clock (CPC), enabled is 1T, disabled is 2T.

for the chipset heatsink, i do believe there is a nice zalman one. i don't know how well it fits under the video card (probably be pretty tight), but with a small fan on it, it should keep things pretty cool :D
 
well, I'm going to have to mess around with finding a good way to insulate the block, as there was just too much condensation after the 6 hour run (I did monitor it) to call the system ready to run without my attention. Interesting though, my idle temps are around 13C under ntune, and as soon as I fire up prime, they jump to 30C, where they only fluctuated +- a degree or two no matter what prime could throw at it. Looks like 2.88 is a lock at 1.75 volts (I needed the juice!), 2.9 was a no go within 3 minutes of bumping it up with clockgen (did this around the 6hr 30min mark). However I quit at 2.9, maybe with antifreeze or just more time/juice 3ghz could be attainable. The thing I don't like is that most would knock higher vcores than 1.75, which is too bad, because with temps holding steady at 30C (under 6 hours of load) I bet even more juice would scale this at least a tad bit higher. I did need to go from 1.7 to 1.75 to step from 2.8 (maybe more headroom, no more than 2.83 with 1.7) to 2.88.

Somebody tell me it's ok to run more juice! As a 3000+ at 3ghz would be very impractical (so is this lab bath) given the vcore I'm guessing it'd take, but it would be a very cool accomplishment.

On another note, my mofset temp (dfi) rose over the 6 hours from 50C all the way to 69C. Is throwing a fan on these likely to yield any results such as higher speeds, or lower minimum vcore?

Screens to come when I decide to hook it up to the network. Yeah I'm lazy.
to sum it up: *work in progress* 3000+ venice retail box, LBBLE 0517BPAW , @ 2.88ghz with 1.75 volts, and something in between water and phase.
 
had another tweak... and 2.65 seems to be the max. no matter what i do its not stable above that.

Just working on the RAM now. seems it can either do 189mhz cl2.5 1t or 220mhz cl3 1t.

Boosted the ram to 2.8. more volts doesnt seem to have any effect.
 
i personally wouldn't go much over 1.65v, but you're hardcore so it's cool. just don't plan on keeping the cpu for too long :p

i think the difference between idle and load is interesting though.. with my watercooling, i've never seen a jump larger than 7-8c.. i guess the chiller is having trouble with the heat load? :D
 
I don't think it's that, as the temp drops right back down once prime is stopped, showing that it removes the load very quickly. The temps probably just aren't as cold as they could be with antifreeze.
 
Anyone know the stepping on tmy 3500 Venice

ADA3500DAA4BP
LBBLE 0519FPFW have not seen this line posted anywhere
 
the stepping is LBBLE, the rest is just denoting which year, week, day and batch from that day ;)
 
ADA3200DAA4BP
LBBLE 0516EPGW

*sniff* You dont like me anymore Eclipse! You said you'd put me on the list *sniff* but i'm still out in the cold *cries*

;)

Seriously tho, where's my name on teh list? :D
 
mtnduey, seems eclipse is a bit busy, I asked him to update my oc 3 days ago, so just be patient :) I'm sure it'll get posted. *pokes eclipse with a stick*
 
Hey ecilpse you have me marked down as 2850 @ 1.52V and while that is what the monitoring software says the bios says [email protected] and if you calculate it out it is 1.275 * 123% = 1.568. So you have you decide which number is valid, cause I'm not pulling my MB to put a DMM on the backside ;)
 
sorry about that guys, i've been crazy busy lately, and a tad neglectful :(

also, instead of writing up a story about your oc, it would be amazingly helpful if you just posted according to the rules, and post your story after it ;)

all right, this time i think i have everyone, let me know if i don't.

also, a big congrats to billr for being the first one with a 3800+ up, shall we have a toast for both being one of a kind? :D
 
(cf)Eclipse said:
sorry about that guys, i've been crazy busy lately, and a tad neglectful :(

also, instead of writing up a story about your oc, it would be amazingly helpful if you just posted according to the rules, and post your story after it ;)

all right, this time i think i have everyone, let me know if i don't.

also, a big congrats to billr for being the first one with a 3800+ up, shall we have a toast for both being one of a kind? :D
Click to expand...

A toast would be good. One of a kind, well, lol that can put one in many categories ;)

I think I’m board limited at this point; I had it to 3.0 in software at 1.65 volts at 48c but not prime stable. I love this MSI board, but maybe it’s time for a change. I want to stick with AGP for now though, I just don’t game enough to go PCI or above. Water-cooling might help too, who knows; I’ll just keep playing.
 
FINALLY!!! Got my new system up:

Venice 3200+
ADA3200DAA4BP
LBBLE 0516EPEW

Max Stable OC 2.7Ghz

Core voltage 1.62
Mem voltage 2.80


MSI K8N Neo2 Platinum
1 Gig Patriot PC 3200 w/XBL 2-2-2-5

watercooled
(29C idle, 43C full load)


So far, max stable setup seems to be:

HT x 3
FSB x 270
CPU x10
Memory 2.5-3-3-11

 
Just as a warning, there is a small risk you guys will end up like me and get a POS 3000 Venice. My Venice could do 300fsb with a 4x mult but nothing over 240 with a 9x. I sought the advice of these forums and Mr Eclipse and it was deducted that getting higher with that multiplyer or any over 6 with a fsb higher than 240 was not possible. I tried 1.5v for 245fsb witha 9x mult and it wont get into windows. Same goes for 1.55v. I even tried upping the chipset voltage and having all the extra power plugs plugged into my mobo(nf4 dfi lanparty). I just ordered a 3200 off newegg and I'l post my results when I get it.
 
beelz, can you add the 3000+, just for the interests of science? i think a lot of guys aren't posting due to their oc's being "lower than average", when the average is higher than it should be purely because of that :mad:
 
(cf)Eclipse said:
beelz, can you add the 3000+, just for the interests of science? i think a lot of guys aren't posting due to their oc's being "lower than average", when the average is higher than it should be purely because of that :mad:
Click to expand...

you can bet the farm on it.

I'm going to try to see how much I can get at all stock voltages except for the memory which won't run at 2.6V ;) running 2685 @ 1.4V and stock everything ATM. Memory is at 3V. temps are cpu 30C pwmic 37C NB 44C @ 28C ambient, fans off or running 5V or less.
 
Ive been running my Venice now 24/7 at 2700 (9X300) with the ram at 245 stable as heck 1.55 volts for a month. Its passed both memtest 86 and Prime 95 for 12 hours. Temps are near stock idle, about 10c over at load.
 
just an idea, but lets say CPU X does 2.6ghz at stock, but CPU Y only does 2.5
Is it still possible that with more juice CPU Y will scale higher in clock speed than X with more juice?

If the chip that goes the furthest with stock voltage is always the best chip, then why not force everyone to post what their CPU does at stock voltage as well, I understand cooling is a variability issue, but it could give a decent view of what steppings are better if enough users post their chips. The way it is now, I could be running a venice that only would do 2.3 with stock, but because I threw 1.7 volts and phase on it, I can hit 2.9 ghz, but another guy with stock cooling might get his venice up to 2.5 with stock vcore, but because it's just stock cooling, he tops out at 2.6 ghz with air and 1.5 volts. If the second CPU would have been in the 1st guys system it might have done 3 ghz. Might have been a little confusing, sorry.
 
I'l get the info from the heat spreader on the cpu on friday and post it. Thats when my new 3200 is coming, egg is FAST.
 
ok guys. you're ignoring the first rule :(

1) Try not to post if you do not have a new cpu or an update for a previous one. All comments/questions to me should be taken to PM's. If it's a general question regarding your cpu, post here!
 
arnemetis said:
Looks like I have the worst yet, but it is o so much better than my old 3200+ cbbid :p
ADA3700DAA5BN
CABGE 0517UPAW
270x10=2700mhz
1.616v according to cpuz
xp90 with 92mm tornado, AS ceramique
In the middle of priming, a few hours in now. Tried [email protected], no good, 2750 no go at 1.6, so I'm settling for 2.7. I may be able to get the voltage down a bit, will edit this to confirm these settings or to make changes as needed. btw bought retail from zzf.
Click to expand...
update to above old post, 10 hours prime, several hours memtest, and several hours 'blend' setting in prime. 250x11=2750mhz, 1.61v. Quoted above so you could see it is an update. Thanks eclipse.
 
(cf)Eclipse said:
all right, this time i think i have everyone, let me know if i don't.
Click to expand...

Dont see mine up there still ;)

ADA3200DAA4BP
LBBLE 0516EPGW

2680 mhz @ 1.6v on Air

No luv for the Dew :( hehe j/k ;)

tks again for the A64 OC Guide man, that got me started in the right direction, it OC's WAY different than any other CPU I've owned, ever.
 
oh yeah, i meant to mention you. i put the stuff in, but then couldn't find how fast you got the core :p

silly me :eek:
 
well, I finally got to 2.908 with my venice @ 1.73 vcore and temps of ~28C load. It doesn't seem to want to do anything more than that. I went as high as 1.78 vcore, but it won't get me 2.933 stable under prime. I probably won't consider this for long term everyday use, as the chiller is impractical for how much headroom it gives me.
 
ADA3200DAA4BP
LBBLE 0517BPDW
XP90 Air
1.55v

2770 MHz

I'm sure this thing could go much higher but my max vcore is 1.55. Time for a new board. :D
 
updated ;)

and i should work on finding a new top end other than 2.76ghz at 1.45v now that i have water.. hopefully sometime soon, while i'm waiting for my ram :p
 
Well, if this is still being updated, then it looks like I've got the nicest 3000+ on the list....

ADA3000DAA4BP
LBBLE 0517APJW
2,961MHz
1.600v
Air - Zalman 7700ALCU

I'm off, kept my ram in slots 2 and 4 but reversed them and now my current BIOS isn't stable for shit when I'm OC'ing...gotta start back at 200x9 and scale up the HTT 5 at a time to find my new "right" BIOS for this setup...should be fun to see how low I can keep the voltage during testing, as I'm SuperPi 1M stable at stock clocks (1.8GHz) at 0.992v...and 300x9 (2.7GHz) at 1.36v.

If/when I have an update I'll let you know...
 
hey! a fellow junkie :D

i'll get you in corn, but before i do that, this oc is prime95 stable, yes? ;)
 
(cf)Eclipse said:
hey! a fellow junkie :D

i'll get you in corn, but before i do that, this oc is prime95 stable, yes? ;)
Click to expand...

Maybe you should hold off until I finish my 32M testing....'cus I'll have 'er burnt in by then...and I'll get you a Prime stable clock...but thats where I top out for daily use and 1-8M testing...so yeah, maybe I'm not doing AS well as I thought...but I still had to post. :p

I'm convinced that even with an XP-90C and M1A Panaflo I'd be doing 1M SuperPi runs at 3GHz...so I'm pumped to get something more running on this thing for cooling.

Buddy of mine gets his dryice containers in a little while and we're throwing this chip of mine under that for some -70C action...hoping for 3.2 - 3.3 for SuperPi 1M and 3DMark2001 anyways. Venice world record is like 3,480 or something, so wish us luck for a nice suicide shot for the world record...I can hardly wait. :cool:
 
damn, if you can do >2.9ghz with 1.6v, you should definitly be able to break 3.5 with DI

and i've got you into the excel document, i'll just wait for that magic prime95 number :D :p
 
(cf)Eclipse said:
damn, if you can do >2.9ghz with 1.6v, you should definitly be able to break 3.5 with DI

and i've got you into the excel document, i'll just wait for that magic prime95 number :D :p
Click to expand...

Yeah man, I'm ridiculously excited, 'cus it does 2.96GHz on 1.600v on air cooling with 42 or 43C load temps...even plain h2O would have me cracking 3GHz easily...I'm very hopefull that we should be able to take the WR. Once we've got that behind us, I'm waiting until I get ahold of an X900XT, get it voltmodded....then we're throwing both my Venice and teh X900XT under dryice to see if we can crack into the Top10 3DMark2001 on the ORB. :cool:
 
i'll be there to throw a party when you do it, cause from how things seem now, it's very definitly do-able. just be sure not to kill that awesome cpu in the meantime :p ;)
 
(cf)Eclipse said:
i'll be there to throw a party when you do it, cause from how things seem now, it's very definitly do-able. just be sure not to kill that awesome cpu in the meantime :p ;)
Click to expand...

No problem...we're good to go, my 3000+ is naked, but I've got the IHS to put back on...so we'll be good in terms of the chips safety...cus this is my second Venice and I'm sure as hell not losing this gem.
 
That is a very nice overclock with very little juice. I'm curious as to what it did at 1.4v or around stock? I'm sure it could have been binned much higher, seems like the chips that will do 2.9+ on <1.65v and air cooling are relatively uncommon (as opposed to ones only doing 2.7, or 2.8-3.0 with much better cooling/higher vcore), and a nice find for sure. What kind of temp improvements (if measured) did you get by removing the plate?
btw, my cpu is a 3000+, not 3200+ listed
 
I'm curious as to what it did at 1.4v or around stock?
Click to expand...

All I've done so far is SuperPi 1M and Sandra runs...but these are the voltages it was RUNNING at for the following clocks...still no burn-in done. I'm gonna be starting at 200x9, moving up 5 HTT at a time, raising the voltage as I need to, for 32M stability...these are the results I've got so far though for 1M:

200x9 - 1,800MHz - 0.992v
300x9 - 2,700MHz - 1.360v
305x9 - 2,745MHz - 1.392v
310x9 - 2,790MHz - 1.440v
315x9 - 2,835MHz - 1.440v
320x9 - 2,880MHz - 1.488v
325x9 - 2,925MHz - 1.536v
329x9 - 2,961MHz - 1.600v

Hopefully I at least see some lower voltages at the above OC's...I remember that after I burnt in my first 3000+, the voltages I needed for most of my OC's dropped between .025v and .050v.

Wish me luck. :cool:

EDIT: Argg...I think I'm gonna go back and do 205 up to 295 for 1M stability just to see the votlage range that's required as I step things up bit by bit.

I'll be keeping track of all of the info and probly just posting up my own thread about it, so stay tuned. :)
 
try an everest write benchmark heh heh ;) nice number BTW

edit: oops sry eclipse I forgot

here, update 2750mHz @ 1.44V, 3.0V mem, stock all else
 
You must log in or register to reply here.
Back
Top