Nice background :p:
Printable View
Nice background :p:
I didn`t read all 10 pages, but my SD 3700 has a cold boot issue, it will not boot in a Vapo with iether a DFI NF4 Ultra or a A8V deluxe. In the DFI the MB LEDs stall out at CPU detect. Both MBs have the latest BIOSs.
I'll include a screen shot and more #'s within a few days :) Results so far:
3700+ CABGE
Watercooling with Swiftech wb and Bonneville heater core
dfi nf4-d
pqi PC4000 (TCCD)
250*11 / 1:1 memory
cpu 2.75Ghz mem 250Mhz
spi 1m ~ 30.122s
vcore 1.5, vdimm 2.7
superpi stable for only 2 minutes at 2.75Ghz but 1hr 40 min at 2.7Ghz. I'm doing a burn-in at 2.75 right now. Will try for max OC with 1.6v vcore.
Eric
Water should be enough to keep the temperatures tame. Mine only shoots up to 31-32C load after overclocking to 2.75 and bumping to 1.5v.Quote:
Originally Posted by joe2004
Eric
my bad, I bent a couple of pins insering it in the socket full of dielectric grease, it`s back up and running on air again, I`ll retest frozen later.Quote:
Originally Posted by Rukee
I used air-conditioning last night to drop the temperature in my room to about 15-16C and voila I was able to overclock SD 3700 to 2,830 MHz with 1.62V.
I run 32 MB superpi bench without problem breaking my record with 26 min 28 sec. However I did side by side comparison running Deep Fritz Mark (chess) and my Venice at 2.81 GHz was doing actually a little better than SD at 2.80 GHz. Hard to explain that but again I think it is a drop in performance because of generated heat.
This chip almost certainly requires better than air cooling and can fly with phase change, no doubt.
Quote:
Originally Posted by stealth17
keep the IHS on, last thing you need is crushing the core especially this is going to be your first phase change, unless you've swapped the cpu in and out and mounted the tube on phase change for hundred times otherwise I wouldn't risk it.
True, however if you look at the results posted, the difference between air and water cooling is minimal (~100Mhz) and the difference between water and phase is huge (400-500Mhz).Quote:
Originally Posted by ewitte
@ easypanic
What voltage are you using, I see CPU-z says 1.808 and it reports low so are you using 1.82v??? Man I didn't dare to go over 1.7 with H20.
easypanic : Thats seems an awfully high vcore to be running, any idea of how its scalling ie. what voltage per 100mhz gained?
I was referring to the comment about it throttling because of heat. You would have to have a pretty poor wc or screw something up for that to happen on water. Really even with air. I'm still at 1.5v. With a room temp of 70-75F my load reading was 28-29C running 2.2Ghz at stock voltage and 31-32C at 2.75Ghz 1.5v. Thats a pretty big difference for only a 4C increase in temps.Quote:
Originally Posted by Eldonko
Eric
It needs an awfully high vcore to reach that speed :(.
I can go to 2.7 GHz on 1.45V, beyond that I need lots of Vcore....
I guess it can do 2.9 with 1.7x ,but I haven't had the time to test more...
I am absolutely positive that at least with the one I have the main issue is heat. I can easily overclock to above 2.8 with 1.6+V if the temperature of my room is low, meaning using air-conditioning full force. But when I have normal room conditions, say 25-26C then this is not possible. The system starts crashing before 1.6V.
I can also see the slowdown in performance after running benchmarks for a while, at the 100% CPU utilization. This is almost certainly due to heat, I cannot see anything else. This SD CPUs have a great potential but a standard air-cooling won't do it and I am using XP-120 with a Delta 120 fan. Venice is much better choice.
By the way your 31-32C temperatures do not sound realistic to me, unless your room conditions are fairly cold.
Mine too likes vcore, also dosen't mind temps. Last night i was runnning S&M at 2.8ghz 1.65vcore and it was hitting 65 degrees. Cooling xp90 but it didn't crash? is that weird or what?
Those temps sound real low to me too, unless its chilled water or something. I run mine at 1.65v and have a good h20 system (big heatercore with 4x120 cooling it) and temps peak at 45C. At 1.7v I'm guessing I would hit 47C.
easypanic do you plan on running 1.8v 24-7? I dunno if that's a good idea heh.
@peter lu - how did you manage that clock at default voltage? I can't seem to get past 2600Mhz at default. If you got 2800Mhz stable at 1.4v you should be able to hit 3k with it with some volts. What stepping? Cabbage batch?
Going by idle temp I'd estimate those temperatures being about 5C off. That would put it really around 36-37C. I'm running 1.5 versus 1.65. That will make a good difference. Also what WB are you running it on? The waterpump and flowrate seem to matter the most. Also my GPU is not included in the loop. Are you watercooling your GPU as well?Quote:
Originally Posted by Eldonko
My temps don't change even 0.1% turning my 120mm fans up to 12v or even adding another 2. Since it does not help I just run 2x120mm fans at 5v off of a Sunbeam Rheobus controller.
Eric
65'C = way passed the Winnie crashing point :D :D I don't know about any other Winnie out there, but mine will be in crash city if it was anywhere close to 57'C. It looks like most of these Venice and SD can take a better beating and dig higher vcore :D :DQuote:
Originally Posted by Da[]San
Hey guys... New member here but not to the OC community :)
I just picked up a 4000+ Diego which came in yesterday and I've been working on it's overclock, yet it seems a bit on the weak on side... I'm actually trying to figure out of it's just my PSU taking a dump on me (which is suspected at this point)
Here's system specs:
4000+ Diego (CABHE 0515)
EPoX 9NPA+ Ultra
1gb OCZ Rev. 2 EL Plat.
eVGA 6800GT
2 x 74gb Raptors
Fortron 500w Blue Storm PSU
Watercooled (Thermochill HE120.3 [6 x 120mm Pana's in push/pull], 2 x Laing D4 pumps, Swiftech MCW6002-64 CPU Block, DD Maze 4 GPU Block, DD Maze 4 Chipset Block).
I'm having a helluva time getting this beast over 2.8ghz 100% solid. At 2.8 it's solid as a rock and nothing can stop it.
2.8Ghz OC Numbers:
FSB: 255x12
HTT: 3x
VCore: 1.75v
Timings: 2-2-2-5-1T (6:5 Divider)
Idle Temp: 34*c
Load Temp: 41*c
Some strange things that have been happening... If I run at 1.75-1.8vcore the idle and load temps are lower than if I run at 1.6v? wtf? Also, the vcore seems to be drooping pretty low, if I set it up for 1.85v in the bios in reality it's about 1.78v idle and 1.73v load.... Comp won't even post at anything higher than about 2.92ghz no matter what I do to the voltage, timings, dividers, cpu multi, fsb, etc... Edit: I tried touching up the 3.3v rail etc, on the psu but nothing helped.
Just curious as if there's a secret to this proc :) Thanks guys, glad to be a part of XS!
Largely untested since I finally got everything up and running very late last night, but I'm currently at 2.9ghz on (rather extensive) Air Cooling with my 4000+ SD
No screenies or anything since I'm posting from work, and nothing interesting to SS yet anyway
Windows stable for games and downloads...will start torture testing tonite with some Priming and get a few SPi runs in
haven't seen temps over 40c yet...but like I said it hasn't really been pushed yet
Most details listed in my Sig
:toast:
I wonder what is the max safe temp then? i'm taking a break from clocking this puppy until i put back my watercooling gear i just don't trust >1.6vcore on airQuote:
Originally Posted by ben805
sounds like your PSU is total :banana::banana::banana::banana::banana: or it is dying. also, your mobo could be causing the droop. my bet is the droop is whats causing you to stop at 2.8ghz!Quote:
Originally Posted by FragTek
list your PSU brand and specs (rail amperage)
I've seen this board do some amazing things on other cpu's w/ a different (much better) psu so that's where my suspicions currently lie.Quote:
Originally Posted by stealth17
I'm using a 500w Fortron Blue Storm rails are as follows: +3.3V@30A, +5V@28A, +12V1@15A, +12V2@15A, -12V@0.5A, +5Vsb@2.0A.
I don't know of the 12v rails are strong enough for all of the crap I have running on them to start off with... lol
the 12s arent too bad considering its a atx 2.0 psu... are you sure the 5v is 28A? the 3.3v is good, -12v is good, +5vsb is good but i just can not believe a forton 500watt has a 5v of 28amp!!! :slapass:Quote:
Originally Posted by FragTek
you need closer to 50amp to really stabilize things.... :fact:
what all do you have running in the computer?
oh and welcome to XS! :toast:
what was the difference between max ocs with or without aircon ?Quote:
Originally Posted by joe2004
About 80 MHz. I got it today to 2.86 GHz 32MB SuperPi stable using 1.7V. Without air-conditioning I cannot apply anything about 1.6V.
Thx for the welcome stealth... And yeah the 5v rail is pretty craptastic and I think that's where my problems are coming from.Quote:
Originally Posted by stealth17
The biggest drains are off of the 12v rails from the radiator rans, water pumps, and case fans. Ive got 6 x 120mm Pana M1A's on the rad and 6 x 80mm no name uv orange case fans. 2 Laing D4 pumps, 2 hd's, and 3 legacy drives. And the vid card as well. There's a crapload of stuff going on in there :)
What's the public XS opinion of the 600w PowerStream? That's what I have my eye on to replace this one with and it's got a 46a 5v rail plus some beefier 12v rails.