HWMonitor reads the r3e temp probes as OPT_Temp 1/2/3.Quote:
Originally Posted by Kosmic71
Printable View
HWMonitor reads the r3e temp probes as OPT_Temp 1/2/3.Quote:
Originally Posted by Kosmic71
Toasty!! Surely you can push for just another 28MHz :D :yepp: But that voltage is very impressive. :up:Quote:
Originally Posted by CryptiK
Thanks Z, I will try that out.Quote:
Originally Posted by zoson
Nice run CryptiK. Did you tune VCore yet? Seems like you have finally decided to push your 990X :)
PS double posts can be deleted by editing it and hitting Delete. I feel your frustration though..
Yeah I could have deleted but wanted to whinge publicly :p:
I think I assed it with that setting, tried a couple of steps lower in bios and passed first loop then BSOD. It is super hot though, my rad is clogged and the chip seems to run hot, might check my mount wondering how the paste is, its been on a few months now. Ran out of mx-2 so I got some antec IC-7 diamond stuff to try cos it was on sale when I went to buy some other stuff.
I taught myself to tune so Ive been busy with that and no time for OC lol
I'm testing my new X980 on R2E i don't put it on R3E.
4655 OCCT 1H30 max temperature 59°.
http://img444.imageshack.us/img444/9062/46551h30.png
OCCT is a joke.
LOLL I agree so much it's not even funny.Quote:
Originally Posted by zoson
@CryptiK - Yea that seems to be the downside to the low voltage/high overclocking CPUs. They are high heat leakage.
Have you tried Shin-Etsu X23-7783D? That is the best TIM imo. Considering your 990X ran 4.5GHz at 1.33v you should be able to bring VCore down from 1.429v for 4.67GHz. Is your motherboard on WC?
It's true - I think although my 990X needs slightly higher volts than your uber-golden chips, it still runs remarkably cool on air. The 970 I had before had a better IMC than my current chip but it ran hotter, and that was with the copper TRUE and very powerful fans compared to my current vanilla regular TRUE. I'm amazed that I get "only" 85C load @ 4.5GHz with quiet fans (Gentle Typhoon AP-15) on my TRUE with 20C ambient. :DQuote:
Originally Posted by spoolindsm127
Yeah for a gulfy it seems to run hot. Ive never tried shin etsu at all, nowhere here really stocks it from what ive seen but ive never gone looking for it either. I was hopeful I could get it down, and I think at 'normal' temps ie: 20*C cooler, I could. 89*C is way too high for my liking, 'd never let it run that hot for any more than a brief test. Yes, my mobo is watercooled with the EK block. I have to clean my rad and see what that does to my temps.Quote:
Originally Posted by spoolindsm127
Its true, youre lucky yours seems really cool, I was thinking your ambients must be very low or just the chip itself. The 89*C run of mine was in ~22*C ambient so it's pretty hot running from what Ive seen.Quote:
Originally Posted by LennyRhys
The heat curve is definitely exponential though - whether it's because the CPU gets a lot hotter or the TRUE reaches its efficiency threshold I'm not sure. The 500MHz jump from 4GHz to 4.5GHz accounts for a "mere" 20C jump in load temps from 45C to 65C above ambient; past that stage though the temps hit the roof, and I reckon another 100MHz would easily be 10-15C extra, as I already know that 200MHz is at least 20C extra: 85C load at 0C ambient...with a 5500rpm server fan LOL. :D
I can run 4.7GHz daily no probs as the CPU never gets stressed like it would in LinX, but the temps are just crazy, hitting 70C even in Photoshop. Much happier with 4.5GHz. :up:
And nowQuote:
Originally Posted by zoson
http://img33.imageshack.us/img33/462/47740.png
Much, much better! That looks like a good chip!Quote:
Originally Posted by Le Titi
You're only a little lacking on performance. Here's something for you to shoot for:
http://i144.photobucket.com/albums/r...g?t=1327261293
I'm surprised he's not getting 90 GFlops with his mem/uncore - I got about 90.6 with lower uncore of 3700 and looser memory timings (8-9-8-24), with c-states off. Maybe needs more QPI or CPU volts, or run LinX for 5 mins to clear the memory before doing the proper run?
Nice run :up:.Quote:
Originally Posted by Le Titi
But how the heck are your load temps SO low at 1.5v VCore?? 90CFM fans make that big a difference? What are your ambients? Or possible LinX isn't fully loading the CPU, hence the low GFlops?
@CryptiK - I'm not saying TIM will resolve your temp issues but you never know, it may give you that extra headroom you're looking for. Try to find X23-7783D, if not I'll send you some ;)
I've just done a system rebuild into a Coolermaster Cosmos 2. It's running great, cool and quiet.
I'm frustrated that there are no usb3 headers on the R3E board, and I really want to connect the front USB 3 on the Cosmos 2. Can anyone suggest a decent PCI-E card that offers USB 3 headers?
Also, this might be a stupid question; I'm running 3 x GTX 580's in the R3E. Would putting a 4th card in (the internal header USB 3 card) somehow change the bandwidth switching on the lanes?
Cheers :)
His GFlops are a bit on the low side, however he only used 4.8gb ram instead of 5.12gb. there's a small difference there... As far as load, i'm guessing he didn't use diagnostic mode, so he has some services eating up a few GFlops. It would also explain the fluctuating GFlops. He might need a little more volts, but he's pretty close. We'll see what he can do when he puts up a 20 linx run. :)Quote:
Originally Posted by spoolindsm127
Also, I use X23 7783D. I find the best method is the + method. I also put thin dots near the corners of the heatspreader. I heat the syringe a little with hot water from the tap before I apply it. I think I've got it down pretty good, but I have been looking for tips. Skinnee doesn't really like it. I was kind of surprised by that. He likes prolimatech pk-1 the most, apparently.
x58 has 36 pci-e lanes. Right now your board is running 16x 8x 8x. That leaves 4 lanes to spare. If the card is pci-e 4x or lower, you should be unaffected.Quote:
Originally Posted by Toritechie
However, if you try to add something that'x 8x in there, you'll drop to 8x 8x 8x 8x.
Anyway... The performance difference between 16x and 8x is like 1% at most.
Sory now i'm changing my rad for a TFC 480 with 4 X 72 CFM Silent X.Quote:
Originally Posted by spoolindsm127
I run a TFC480 be interesting to see where your temps end up. I run 1600rpm scythe s-flex's usually at 900 rpm hence my temps but I can crank them up if I need to theyre just too loud to sit next to like that though for any length of time.
Temps are at 62° C max for 4712 OCCT.
What is the diference of the Rampage 3 Extreme an the same in Black Edition.
I by a Black Edition because the first R3E as a big problem of pins on the Socket.
For me not to much.Quote:
Originally Posted by Le Titi
Don`t use the Thunderbolt, as I have a seperate audiocard, icm SLI.
Performance: Using 6x 4Gb, it doesn`t differ much from R3E.
It`s more stable using higher multi`s and Hypers, or 3x 4Gb icm the same core.
Looks: I`m pro Black, though when it came out, R3E was best looking atm.
enjoy, and post your findings..:D
Quote:
Originally Posted by Le Titi
I don't think there is any difference worth mentioning at all *except* Thunderbolt card.Quote:
Originally Posted by XII
Nobody with R3BE can do better than my overclocks on R3E.
The Thunderbolt card isn't that good and the SW didn't even work for me properly. Bought a Xonar Essence STX to replace it. Thunderbolt's sound card actually connect via USB.. :)
My regular R3E ran way too hot and the sound card went bad, sold it after RMA and got the R3BE. I couldn't check the paste since the retailer would have denied my RMA. No difference in OC between the two.
Well, the R3BE is the best looking board so far. :D
Sorry the hijack, but I really don't want to litter the forum with miscellaneous posts to be able to see the FS forum.
Anyone have a R3E with EK Nickel/Acetal block they are looking to sell?