There are several tests that prefer HT on, 3D Mark 06 CPU test is another.
Printable View
There are several tests that prefer HT on, 3D Mark 06 CPU test is another.
Is there such thing as a QPI speed WR or such?Quote:
Originally Posted by jurky
What do you mean it starts throttling over 4.2ghz?
Quote:
Originally Posted by deachus
Thanks for that quick test. I'll be sticking to HT on for sure, because if it helps with 3d then that is mainly all I care about..gaming performance. Benching is 2nd priority in that sense lol.
My testing was for video encoding performance and not gaming performance but noticed the big difference in CPU score in Vantage when doing some benching one night a few weeks back. That led me to do the testing posted above. Now if only I could get that magical 4.5 HT ON LinX stable.
I dont think that competition for QPI speed WR exists. But higher stable Qpi link speed means little more performance in 3D benching.Quote:
Originally Posted by fl0w3n
Throttling means that CPU speed start to go up and down by itself, rapidly changing CPU multiplayer or bclk when you pushed over the limit. In my situation 4200MHz on qpi link speed is the limit on air cooled CPU and the board.
New Rampage III extreme black edition >_<
http://www.hardwareluxx.de/index.php...ig-donner.html
http://img211.imageshack.us/img211/6126/naamloosbg.jpg
Whaaaat!!! Really???? No Way!!!Quote:
Originally Posted by jurky
I probably use the wrong word (throthlig), but it is close enough to describe the problem.:shrug:Quote:
Originally Posted by Ian[Frozen]
sorry but my english is little bit limited.....:rofl:
Oh I see. Ill be resetting-up my loop and giving the old cold weather oc runs a go again to get some ideas of whether my setup can do 4.5 ht on or not.Quote:
Originally Posted by deachus
Okay. Thanks for clarifying. Great results on air none the less.Quote:
Originally Posted by jurky
Question:
Is increasing PCIE freq for preventing GPU driver crashes?
I have a problem where my GPU driver crashes under heavy load(prime + furmark) and sometimes in long gaming sessions. The PC doesn't crash and prime never errors out. LinX can also some times cause a GPU driver crash, but not always.
in my case, i believe i increased to 105Mhz PCI Freq to make my boardQuote:
Originally Posted by zoson
boot on (Bclk) 215x21 and i brought it back to 200x21 and i didnt change
it back.
i too had gpu driver crashes, but i think i fixed it with more cpu power.:shrug:
I just bumped my PCIE freq to 102mhz, my CPU voltage up one notch from 1.325 to 1.33125v, and my qpi up to 1.33125 from 1.325v.Quote:
Originally Posted by Andreezy
I'll report back if it stops my gpu driver crashes. I'll also try slightly higher settings if it doesn't.
good luck:up:Quote:
Originally Posted by zoson
when testing in LinX, what i have learned to do is try and set the lowest
stable point for QPI, then i run LinX/IBT all threads and i begin with lets say 128Mb---> then i increase to 256MB--->512MB so on and on untill system
start crashing or stops on error, depending on the crash (BSOD) 124/101 i
adjust untill i make it all the way to lets say 4GB, after i can run 10/20 passes
on 4GB then i try and go for testing under full stress.
P.S.
i have a i7 970 coming in the mail, should be here tom.
anyone test if the ln2 jumper if it helps any on ln2?
I don't get a bsod or cpu crash. It is only a gpu driver crash. Screen goes black for 1 sec, then comes back and a little message says that my gpu driver was restarted because it had a problem. It doesn't cause any problems with linx - I am linx 200 loops stable.Quote:
Originally Posted by Andreezy
i did a lot of trial and error and not in the safest way, i too remember myself going through this where my screen would dim then driver crash and recover.Quote:
Originally Posted by zoson
dont remember if it had something to do with bad voltages or more Vcore needed.
in your case you dont seem to need more volt to cpu.:shrug:
try a clean install of the drivers and update what you can/have not updated.
i was also crashing some drivers before and i believe it was related to IOH Voltage.
try and check your settings with other similar setups..
didnt help me at all, i wonder does it do anything :shakes:Quote:
Originally Posted by Dom7184
Zoson which driver are you using
Whatever the latest driver is I updated recently. I've been updating each time hoping that it would solve the problem. I started at 176.xx series and I think i'm on like 250.xx now. None of them make any difference.Quote:
Originally Posted by fmsam
Hey ,
got my RE3 a while ago and followed this thread ( thanks god for this awesome thread) but now i got som problems... with my past mainboard Evga Sli Le 20x200 24/7 on 1.28Vcore 1.45 qpi 1.8pll
now to my 2 real questions, 1. i have problem with getting a stable 24/7 clock at even 3.3ghz ... so i go with stock and xmp profile for memory thatīs all and im not very statisfyed with it
and for some reason its more stable when i go higher and bench, BUT voltages are to high to run on 24/7
so what is recomended voltages for a 24/7 setup 20x200? (Vcore-Qpi-Pll-?) using bios 1102
question 2 . when i oc my max stable clock is 4388mhz, 1.5Vcore~ 1.82pll ~ qpi 1.45 and just raising ioh/ or ich to what feels fine . thats max what i can get stable . read lots about pll on 1.4-1.6? and Vcores around 1.38v?
what is yourīs opinion on voltage on oc? ( i read about ppl having problems with "higher" clocks on W3540? any ideas?) oc bios 0099 ( most stable i have found yet, need other ?)
setup: RE3-W3540 D0,Gskill Trident 2000mhz cl8 1.6v,5970, corsair 1kw , cpu and gpu watercooled by 3x120 and 2x120mm thermochill rads
Greatfull for help!
soulblade
Registered User
Join Date: Oct 2008
Posts: 1 <---------wow how did you manage to stay :sofa:
to get back on topic, i too have a xeon on this board and only thing i can tell you is, be carefull when using T-Balance/X-Power distribution profiles if you dont have your MOLEX CPU Cable plugged in..
i O'C to 4.2Ghz and then was trying to up it and while doing that i changed the profile from T-Balance to X-Power and it brought nothing but headaches for me since i dont have that Molex plugged in. if you do good for you.
RIIIE w/ Bloomfield = sucks
RIIIE w/ Gulfy = rulz
http://www.unknownteam.cl/wp-content.../01/310421.png
haha i donīt know really ! on XS 90% of your diffirent problems are found out in a thread and dont really need an reply :DQuote:
Originally Posted by Andreezy
just wounder, molex? ainīt that just for making pci-e lines more stable when doing CFX or sli setups?
canīt really see why is should help but ofcourse i will give it a try, i been using both T-Balance/X-Power and i donīt really feel any diffirent on oc.
what bios are you using at 24/7 and oc bios ?
but thanks for a fast answering :)
and for other reply , yes i would like to own a Gulftown to but my wallet aint that big at the moment ;) bloomfield have to last a little while more and i problebly get a 980x when ppl uppgrade to sandy bridge ;) Ty anyway
i am using the 1102 bios, and i meant the molex plug called "EZ Plug" which i
believe supplies power to cpu.
but if its no difference for you between T/X power balance then good.
in my case if i go from T_Balance to X_Power i would get bad memory residuals in Linx no matter what qpi/dram voltage used, if i set it to T balance then everything is good.....:up:
maybe i should just plug that darn ez_plug myself...:yepp:
i cant hump over 4333mhz.....bad memory residuals if i go over that:down:
:shrug:Quote:
Originally Posted by TOOL_CHILE
i would be more awed to see your 980X ES moving 85-->93 GFlops in LinX
which it can do very well, you seem to have enough hardware to pull it off.
good luck on tuning that chip.:up: