first i have to make my windows bootup again, wrecked it by ram oc..
first i have to make my windows bootup again, wrecked it by ram oc..
And another update! Getting more excited now:p:.
Though my only stability testing atm is a few hours of Prime long FFT's, Ill use some other applications later on as well. Is the AOD stability test worth anything or just like Prime small FFT, aka crap?
Good results, glad to see a little NB clocks indeed help stability then.
I just quit a Prime which ALMOST ran 6hours at 1.3125Vcore and Vnb. With ~20C ambient (Im guessing), it only became 42~43C. I also just noticed, it runs (thus far) stable at only +0.0125V from default:).
Thinking quickly what Ill test now, as Im going to bed in a few minutes. Thinking to drop Vcore and Vnb at a little below 1.3V and see how it does that.
I also thought, now Im searching for lower Volts, I must be burning the CPU nicely in now. At least, I hope:ROTF:. I could jump and try to get 2.8Ghz again, but now I got a stable point I want to know first what Volts it really needs. I decided one thing though, just run Vcore and Vnb 1:1. Maybe the Vnb might get overvolted for only 2.2Ghz a little, but as it was one of the parameters I changed which got everything stable I don't see anything wrong with it. Also DFI BIOS recommends it, although as I said, I dont know it it's really needed.
And oh yes, I might order a 3Ware 9650SE RAID controller tomorrow since I found a shop which has it in stock RIGHT now with a nice price, about 170Euro's or so. And me and my wallet saw the announcement of WD VelociRaptors too:eek:. 300GB, 2.5" but in custom passive cooled 3.5" housing and they claim it's 35% faster than current Raptors. 299$ MSRP though.
Might I make a smart suggestion and advise you to go with a Areca 1210?Quote:
Originally Posted by Rammsteiner
Areca Bar None has the most Compatibility for people that change Boards frequently trust me on this!
On the second total waste of money and advise that you go SAS if you have that kind of money to dump on a Raid 0 Array beside they are 300gb which means storage drives not O.S Drives.
OuchQuote:
Originally Posted by Oese
KTE tried the auto tune feature a week ago. That brought him above 3GHz but it was not stable. I assume that process uses AOD stability test functions, so they are not usefull.Quote:
Originally Posted by Rammsteiner
Do you use watercooling, you have no info about that in your sig?Quote:
Originally Posted by Rammsteiner
If those overvolts i measured are correct and common to all boards your real vcore is at ~1.355V. :rolleyes:
With NB speeds below 2200GHz and NB voltages at ~1.3V i see no need to keep vcore and vnb in sync. Is there a bios manual from DFI available? Only have the sapphire descriptions.Quote:
Originally Posted by Rammsteiner
On the other hand that will shorten the lifetime of the cpu without doing something productive. :)Quote:
Originally Posted by Rammsteiner
Anybody running Vista 64 bit?????
I just gave it a whirl, wouldn't stay up for more then a few minutes without BSOD, not overclocked and overclocked....
Just wondering if anyone else having issues with 64bit Vista.....
LinkQuote:
Originally Posted by Rammsteiner
:yepp:Quote:
WD promises for the record "Killer Speed": According to data sheet provides maximum transfer rate of 125 Mbytes / s and the average access time is about 4 milliseconds. . This might make the drives in about as fast as SAS drives with 15,000 min -1.
While you all are sticking your probes into your boxes, can you get a reading off your 12V rail(s). I would like to know how close the VCore is tracking to the 12V source and if compensation is added for rail droop. We can round up these number and know about where it will track based on the 12V DMM reading. Thanks.
For now all I can do is drool until over this massive thread until my board arrives on Wed.
Hi AC,Quote:
Originally Posted by AlabamaCajun
I measured the rails with 4pin and 8pin connectors.
http://www.xtremesystems.org/forums/...postcount=1453
Cool thanks. It looks like these regs are tight where you are only a few mils over. Interesting enough with the 4Pin droop at load it was slightly higher than the 8Pin at load.Quote:
Originally Posted by justapost
:welcome: AlabamaCajun good to see some more folks joining the ranks of the D.F.I 790FX-M2R Board:clap: :yepp:
Brother, the Areca is very nice although it's a lot more expensive which I dont feel like paying for a RAID card. Whether it's worth it or not, the extra money isn't just a few %, more like 300%. In a review I saw the one I'd like to order is very fast as well. Im not swapping systems often so it will be fine. I dont mind re-installing Windows or so, but I'd like to get the performance from my Raptors which get ruined by SB600.
Justapost, Im using a TRUE with 130CFM 120mm fan. And thx for the comment about AOD stability test, reduces some time waste:D.
There's no DFI BIOS manual to be found, or I didnt look good enough. Sapphire BIOS manual does it for me for the few questions I have here and there. I know more Voltage might reduce life time although default is already 1.3V. Also keep in mind I dont run the NB < 2200, but AT 2200:).
And another update, Prime for 9hours at 1.2875Vcore and Vnb. Yay, this is going awesome. Getting a bit worried though since Im not used to this amount of succes with Phenom yet. Ill let the PC play some music, browse the internet a little etc to see how it does in idle. Although Im typing this from the PC so first forms of idle testing go well lol.
Rammsteiner: Tech Report has a good article with some serious testing on the VelosiRaptors. Here's the link:
http://techreport.com/articles.x/14583
They even gave it an Editors Choice Award, which is pretty impressive coming from TR, IMHO....
They seem pretty pricey to me, but If I had plenty of extra cash I wouldn't mind having one for my system drive! :D
Not sure it would be worth $600 to Raid 0 a pair of them, but if you do I'd sure like to see some benchies.... :p:
Oh, missed the reviews lol. Results look very nice though:).
But Ill save my cash for MOAR SPID0R POWAH!
http://www.techpowerup.com/58365/AMD...ost_Ready.html
:yepp:
2 hours into 1.275Vcore and Vnb Prime... I ran WinAmp for a little hour and browsed the web a litte, took a screenshot and saved it in Wordpad before this Priming. No freeze/BSOD's occured.
Last night I also noticed the decimal Mhz's from HT/NB DO NOT run 1:1. At some point HT was 2000,3 where NB was 2200,1... I know the NB runs at a higher multi though, but that shouldn't result in a +0,2Mhz for HT. (0,3/10)*11=0,33. This is such a little margin we shouldn't even notice what proves for me personally HT/NB speeds have a little latency between them where the little 'waving' HTT might give the HT a higher clock for a very short time than NB.
Anyway, now things go a lot better I can only make the conclusion that my poor Phenom OCing in the beginning was caused by a messed setting in BIOS. Be it Vcore and Vnb not running 1:1 or HT sometimes overspeeding the NB bus, I dont know. Thus far it works!
Yes as Achim said, AOD AutoClock just puts high load on CPU components [found in stability test] and then changes clocks 1MHz HT higher but drops loads, and quickly ramps up the load until it freezes or crashes. Then you can use it to stop at the highest stable clock you found.
If your system does not suffer from the WISE/idle instability issues, then you can use AOD AutoClocks highest stable clock and it should be perfectly stable for use since it tests more CPU components than most other stability testers we use.
If your system does suffer from the WISE/idle instability issues, then no matter what AOD/EVEREST/P95/OCCT/Orthos/Sandra finds as your stable clock, your processor will still be very unstable - hence no point using them for oc stability testing.
Hmm I use the Xigmatec HDT-S1283 with the stock fan (53cfm) and an Xilence Xtreme 74cfm. Those 130cfm fans are hard to get here. Can it be that temps will drop ~10°-15°C if i use an 130CFM fan?Quote:
Originally Posted by Rammsteiner
Thank you for the update, you still use an 13,5 cpu multi?Quote:
Originally Posted by Rammsteiner
I'm priming at 2.8/2.2GHz with 1.3125V/1.25V since ~6hrs. 0.00625V less on the nb failed after 4 hours, i'll let this prime for at least 20hrs.
There are still a few HT settings i did not yet play with. I tried different CPU-HT Gap values with no luck.Quote:
Originally Posted by Rammsteiner
Thanks tye,Quote:
Originally Posted by KTE
So you'd say AOD stability testing is as reliable as prime95?
Im not entirely sure anymore if the fan was 130CFM or the metric one...
It makes about 30~35dB(A) noise for 120mm fan. So if 130CFM sound reasonable, then it's 130CFM:ROTF:. But I dont know by how much your temps will drop with a better fan.
Im still on 13.5multi yes. I keep everything except Vcore and Vnb as it is.
Thx for the little better input about AOD stability test KTE. In that case Ill just leave the PC Prime and do some semi idle stuff. At least it shows a lot more stability now compared to when I just got it. Back then it would almost insta freeze after Prime Small FFT.
Im still under the impression at least Prime95 with Small FFT's is useless. As noted before selecting a test, it stresses L2 and within. I dont know how the CPU works together with the NB, but since Large FFT made Small FFT stable settings almost insta crash I think Large FFT's do test everything and Small doesnt.
Ive some HT settings as well which I ddint try out yet. But I think I read earlier the adjustment gap is useless. The 4 options in the HT-link meny, no clue what it is. And as long as I dont know what it is, actually almost no one, Ill leave it disabled (default). I guess DFI implemented it just to have more :banana::banana::banana::banana: options lol.
Not sure Achim, I'd need to have a setting fully stable before testing both and comparing the two - but AOD test is testing more CPU sections than P95 does, definitely. Also, it fluctuates the load, i.e. ramps it high and low, which is also another good test addition for stability. :)
I have two 12cm +100CFM and two 152CFM Deltas here, use them on my Xig and any cooler I can - they're both very loud at full, nothing I would like to run in normal home builds. You only want to use them if having a fan controller to run them from 5V to 12V as needed, in which case, my GPU fans are always louder at 45% so their noise is drowned in the background.
I've not looked into it much but at work we use these air cleaners and they're rated for 3 set speeds, 1/6-horsepower motor moving 550/702/1044 cubic feet per minute of air through a small section at the front of the box [extremely powerful, cautioned where you point it]. I've thought for a while that they should work very well for when system cooling is needed, even at a specific overheating location (?) :D
Looks nice KTE. Although I'd tighten the heatsink very well before it gets ripped off including half of the mobo lol. Also make some air vent so all the air gets pushed trough the small surface of the heatsink. This might make the air stream even faster too lol.
Just manually stopped Prime for 1.275Vcore and Vnb @ 13.5x200. Ran 6.5 hours, only became 39~40C with 17.5~20C ambient.
I could try even lower Volts, but I already went down 0.075V on both. So I set multiplier at 14 and it's testing now. If this turns out to be stable... Yes, then Im happy with my findings and nothing is lost yet for me. Maybe 3Ghz is on sigh in that case:yepp:. After this 2.8Ghz test Ill try and reduce Vcore and Vnb around 1.2625V or even 1.25V. But I think this is a bit looking too good to be true. If it works, you won't hear me complain though:p:
:) If my prime95 test is finished, I will mount an vacuum cleaner with an funnel on the xigmatec.Quote:
Originally Posted by KTE
http://www.abload.de/thumb/vacuummsf.jpg
:rofl: Gained 1°-2°C
Lowest voltages for 2.8Ghz/2.0Ghz
CPU-VID: 1,3125V (Bios) 1,351V (DMM)
NB-VID: 1,15V (Bios) 1,144V (DMM)
CPU-TEMP: 63°C coredamage, 60°C max prime95
Prime Time: 22hrs
Lowest voltages for 2.8Ghz/2.2Ghz
CPU-VID: 1,3125V (Bios) 1,354V (DMM)
NB-VID: 1,15V (Bios) 1,144V (DMM)
CPU-TEMP: 63°-64°C coredamage, 62°C max prime95
Prime Time: 12hrs
I use the following procedure for stability testing.
1. Sandra Memory Latency (that triggers a freeze here very quickly)
2. CoreDamage for a few minutes to see what temps i will get
3. Prime95 together with 3dmark06, to generate a very high power consumption
4. Prime95 longtest
5. Application testing (Surfing, Openoffice, Antivir) for a while.
Lowest voltages for 2.8Ghz/2.4Ghz
1.3125V: passed 1, power off at coredamage
1.325V: passed 1, power off at coredamage
1.35V: passed 1 and 2, failed prime95 immediate
1.3625V: passed 1 and 2, failed prime95 after 3mins
1.3875V: passed 1 and 2,3 power off during prime run after ~30mins.
1.4V: passed 1 and 2,3 failed prime95 after 48mins
1.4125V: passed 1 and 2,3 failed prime95 after 4:50hrs
1.425V: passed 1 and 2,3 failed prime95 after 5:30hrs
Trying with 1.35V CPU-VID
1.35V: prime failed immediate
1.375V: prime failed immediate
Back to 1.3125V CPU-VID
1.4325V: passed 1 and 2,3 failed prime95 after ~9hrs
:coffee:
go get PCmark Vantage, run the full test 90mins or so...passes that its stable ;)Quote:
Originally Posted by justapost
Great testing there, thanks.
That's the sort of testing which is very interesting and helpfully conclusive. :)
My 740G just booted 9850, randomly -> gotta ask MGR if he'll let a girfriend deliver the system to work now so I can test it at break :D
Well wtf,
This is starting to make zero sense. I passed 1.275V. At that point I decided to go fro 14 multi. But whatever I tried, even adding Volts, it was everything but stable. Played with HT multi, NB multi, nothing.
So I tried previous settings which ran 9h+ and it randomly restarts or BSOD's... The restart makes no sense at all since Ive the auto reset OFF for when a BSOD occures. Also have the shutdown thing in BIOS off at xxC.
I just noticed though I ran this time with unganged where as, IIRC, I previously ran with ganged. Maybe Im an uber n00b and never caught that, but does ganged vs unganged has any effect on stability? At DDR800 mode it is.
Im really considering, how annoying it's starting to get, to throw it all out and plug everything on my DQ6 and see how that goes because this makes no sense.