I`m sure I read someone saying a page or two back the only thing that has been added into bios 1203 is memory compatibility fixes and nothing else.Quote:
Originally Posted by zoson
Printable View
I`m sure I read someone saying a page or two back the only thing that has been added into bios 1203 is memory compatibility fixes and nothing else.Quote:
Originally Posted by zoson
0095 is installed. No noticeable difference so far between it and 1102.
the 0095 does not require less tension compared to 1102 (the Vcore?)
this might be an official bios in some days/weeks
http://www.abload.de/img/zwischenablage01qlfo.jpg
ps:
this bios has NO QPI PLL voltage settable and separated from CPU PLL options
when i get the changelog i will post it
pps:
@spoolindsm127 :bananal::bananal::bananal:
Thanks mate for given us the head up.Quote:
Originally Posted by Fire4Wire
Hopefully this one contains more fixes, performance enhancements, better overclockability etc than just memory compatibility fixes.
And may never have, according to @LardArse opinion :(.Quote:
Originally Posted by Fire4Wire
Fire4wire you are behaving really badly! He appologised for making a joke and now you are dragging it on !
Pretty lame
I ended up dumping 0095 and loading 1005 again.Quote:
Originally Posted by tistou77
Randomly only 8/12gb would show up on boot of my mem with 0095.
I needed the same vcore for 0095 once my loop temp stabilized.
Asus needs to get the bench and 24/7 user biosses hosted on different FTPs, it's far too confusing for most...
ok, bios 0011 (and also 0095) requires a little less Vcore than bios 1005 or 1203, I thought it was the same with the 1102Quote:
Originally Posted by zoson
you are back on the 1102 bios?
Most likely it won't..Quote:
Originally Posted by Sylver123
As all of us have probably figured by now that the hype of the R3E is fading away and so is the bios support.
Plus Fire4Wire posted that just for the attention, he's pathetic like that :rofl:
I was going to use another word to describe him but I believe it'll get censored ;). Let's stick with lame :p:. BTW he was referring to me, not PytonOrm :up: Nonetheless he was acting asinine with him as well. Lucky for him PytonOrm has balls and self respect and gave him an undeserving apology.Quote:
Originally Posted by AceBaran
PS
@Fire4Wire nobody cares about what bios you have :) No one expects anything from you :up:
BTW didn't you say you sold your R3E to test a 1155 board? Haha lieing POS
I'm testing with 1005 tonight on BIOS2. BIOS1 has 1102 with my 'failsafe' settings.Quote:
Originally Posted by tistou77
I needed same voltage for 4.2 with both 1102 and 0095. There was no difference. The only thing that changed is 0095 doesn't always see all my mem.:down::confused:
So I got really bored and installed this board again. I got a new kit of Mushkin Radioactive 1600 c6 for Christmas, the bios I was on, 1005, didnt seem to like them, would not run them at default speed or timings which I had tested with no issues on the XPower board. So I updated the bios in hopes it would do better, but now my vcore is waaaaaaaay off, WTF?
edit: still cant boot into windows at just under 1600mhz using 2:8 mem multiplier, but can boot at 1960mhz with loose timings but drop a channel. REX3 FTW :rofl::ROTF::rofl:
I just got my watercooling installed. Now its time to OC. Are there any baseline settings I can try, initially, to run at 4GHz with my 950? I know its a lot of trial and error and I dont mind that, but a little help to start would be awesome.
I've read the majority of this thread and I have to say I'm disappointed in the Rex 3. BLCK seems hit or miss which is not on for a top end OC'ing board. UD7 and Classy seem was more consistent in getting 230+ too many examples of where this board is no better and sometimes worse than the Rex 2 was. Bad memory compatibility too. If you OC with multi and low BCLK it seems ok though.
I think you guys are dreaming if you think asus is going to bring out the option for separate PLL controls. As Peter said, they appear to be tied into the same plane so separating them is not possible. Also when have asus ever brought new functionality to a board post release? Asus asked for my input during Rex 3 development, and I suggested things like ability to lower CPU PLL (could not do this on Rex 2) as guys on Giga's and Classy were doing this with OC'ing improvements. It was taken on board and Rex 3 has it adjustable, but I asked could they do it for Rex 2 also. Not a chance.
Expect a few more bioses with 'enhance memory compatibility' as the entire change log, with more and more time between each release. If you want that nice level of transparency and communication, trying a gigabyte seems to be a good option. I remember when HiCookie was on here every couple of days in the UD7 and UD9 threads answering questions, posting mods, testing beta bioses and handing them out, was awesome to see.
Yep, no point having a QPI PPL voltage setting with a bad IOH chipset.
I don't think this board was really designed around low / locked multiplier CPU's.
Men, what voltage are u using on qpi vtt? i set up tu 1.35v and that problem dissapear =/Quote:
Originally Posted by OC Nub
Seems like someone's feelings got hurt :ROTF:
Sooo why didn't you just send a PM to me and Ace rather than spamming the thread yourself?? :rolleyes:Quote:
Originally Posted by Fire4Wire
It's okay dude I know truth hurts. You don't reply to my PM regarding your thoughts on bios 1203. You lie by saying you sold your R3E and that you are sick when I ask for your help. Don't expect anyone to respect you or honor their word when you can't do the same in return.
Make a New Year's resolution to not lie or be a douche. It's never too late! :up:
PS Funny how you said those were you're last words to me and you came back to edit your post and add more BS :rolleyes:.."coincidentally" RIGHT after I make this post :rofl: I have a life so I can care less about beta bioses. You're pathetic man look at the way you act :shakes: tisk tisk
On to important things, finally got to 4.5GHz whatupppp
http://img20.imageshack.us/img20/732...hyperpi32m.jpg
Cannot get it to boot with the 2:8 multi no matter what QPI voltage I try, it doesnt like this ram, or the Redlines I have.Quote:
Originally Posted by TOOL_CHILE
@Spoolindism127 Great Vcore buddy ... i envy your chip !!!
The new beta bioses coming out for R3E mainly contain 'memory compatibility' enhancements. What was the result with bios 1203 or 0095/0096?Quote:
Cannot get it to boot with the 2:8 multi no matter what QPI voltage I try, it doesnt like this ram, or the Redlines I have.
Thanks chief :up: The IMC on this chip is ungodly!Quote:
@Spoolindism127 Great Vcore buddy ... i envy your chip !!!
I actually just received my W3570 3846A803 today so have to decide now which chip to keep :p:
Quote:
Originally Posted by spoolindsm127
Keep both chip!!:D
I think 3846Axxx batch is magical chip.
The first, not to set high VTT V when you start OCing.
If you set high VTT V, perhaps the chip will very high temp.
Many of 3846Axx batch chip do not like high VTT V.
Yes I am VERY tempted to keep both! Hopefully if financial situation improves I can :)Quote:
Originally Posted by kurumi_Japan
I did research on this batch and I agree, I'm expecting some magic :D
Thanks for the advice on VTT :up:. I will start at lowest 1.2v QPI to determine strength of IMC.
My 3949A372 920 needs only 1.363v to stabilize my sig mem clock :eek:. But it's a pretty toasty chip as you can see. I still have to tweak my WC setup but I'm hoping this Xeon runs cooler and is similar in terms of IMC ;)
This is my first unlocked and first Xeon cpu so I am very excited :D
Tried a couple beta bios's after work today, still no luck. Cant boot using 2:8 multiplier, and 2:10 shows only 2gb of ram in windows.Quote:
Originally Posted by spoolindsm127
I think my board is just a POS, it never did well to begin with months ago even with compatible ram.
Will go back to the XPower until SB is out next week.
How much QPI did you actually use? For my mushkin with a 45nm cpu I needed 1.4875v for 200 bclk and 12gb 3x4gb 2000mhz 9-9-9-25-1t... for 210bclk I needed 1.543v QPI.Quote:
Originally Posted by OC Nub
32nm cpu doesn't seem to have the same problem... Overclocking has been a breeze.
I started at 1.35vQPI and went up as far as 1.5v and couldnt get it to boot. With the 2:10 multi it boots at 1.48vQPI, but even trying up to 1.55vQPI it still wont show more than 2gb in windows with the beta bios's I tried tonight. With the latest official bios windows shows 4gb with the 2:10, 1960mhz but still wont boot at 1600mhz. Increasing vDimm doesnt help either.
The bios Im using on the XPower is dated 8/20/10, it supports this ram, cant understand why Asus board with a much newer bios does not. The same kit performs very well on the XPower.
I say as a last resort you try the previous beta bioses. Specifically 0878 and 0011.Quote:
Originally Posted by OC Nub
I know 0011 has some additional memory settings compared to the official release bioses. And 0878 was mainly geared for memory tweaking.
I'm sure you know but just in case check out the OP they're all there :up:
BTW you may have already tried this but did you try raising VCore? Did you try lowering vdimm? What if the board is overvolting the ram and it doesn't like it :confused:
I checked DRam with a DMM, whats set in bios is pretty close to the actual voltage. Finally got 1 kit working with the 2:8 multi, windows shows 6gb installed now. I used the 21x multi for the cpu, so it required slowmode to be enabled in order to boot at 220+ bclk.
Happy new year for all you guys! :D.
Well, new year, new benchs and new configs to test :).
211 BCLK x 19
1.27 vcore
1.37v qpi vtt
1.65Dram
1665Mhz 6-9-6-18 - TRFC 132 ( crappy :( ) 1T.
5 runs of linX, so i will try it with more BCLK, linX, cinebench and 3DMark later.
Setup:
Dahaka - Unknown_Team
Core i7 950 + Swiftech Apogee XT
ASUS Rampage III Extreme
Sapphire HD 5870 + Swiftech MCW80 + Thermalright VRM-R5
G.Skill PI 3x2gb 2000MHz 6-9-6-24
Pcpower&cooling 750w Silencer
Happy new year for all :D.
Well, new year, new benchs and new configs to test :).
211 BCLK x 19
1.27 vcore
1.37v qpi vtt
1.65Dram
1665Mhz 6-9-6-18 - TRFC 132 ( crappy :( ) 1T.
5 runs of linX, so i will try it with more BCLK, linX, cinebench and 3DMark later.
Setup:
Dahaka - Unknown_Team
Core i7 950 + Swiftech Apogee XT
ASUS Rampage III Extreme
Sapphire HD 5870 + Swiftech MCW80 + Thermalright VRM-R5
G.Skill PI 3x2gb 2000MHz 6-9-6-24
Pcpower&cooling 750w Silencer
Happy 2011 Everyone!! :)
Started off the new year with overclocking in unfamiliar territory; high BCLKs :D
All stock voltage and freq. board is solid at it's BCLK limit of 221.
Once you tweak PCIE Freq. and QPI, IOH, and ICH voltages I got this: (I forgot to select Mainboard tab in one of the CPU-Z instances :p:)
http://img408.imageshack.us/img408/4730/238bclk.jpg
Happy New Year everyone.:up:
Happy New Year!!
Been playing with the board some, trying higher uncore in 32m. So far this has been my best score, using the i7-930 in SLOWMODE.
Nice one Nub. I am in the 7.5 range also though you got some sweet latency going.
Thanks.
Its hard to knock any time off the score at this point. If I go any higher on uncore or ram frequency it loses a channel in windows, Ill have to raise timings to try any higher.
I here ya. I knocked off maybe 5-6 secs jumping from my 920 to a 3570 Xeon at a slightly faster speed with a 6G set of Gskill perfect storm. I really need to dial it in on the ram for a decent score as this board can do it.
http://www.unknownteam.cl/wp-content...-BCLK-linX.jpg
That is my last "stable" config for "high" BCLK.
OC nub, how it's the 1203 bios? it's a good one ore it's the same thing that 1005?
Maybe RMA it mate? Sounds like it has some genuine issues.Quote:
Originally Posted by OC Nub
And for ^^ no way should you need 1.45+ Vtt for only 200 BCLK, 200 is nothing. My 920 on R2E is doing 210 BCLK totally stable with 3575 MHz uncore, DDR3-1682 MHz CL7 @ 1.18 Vtt.
It runs ok with elpida hypers but it doesnt like this new ram.Quote:
Originally Posted by CryptiK
I should have RMA'd the board long ago but Ill probably end up using it to build a rig for my wife.
Max SPI32M BCLK?Quote:
Originally Posted by spoolindsm127
Just did some testing on my R2E, and Got this - it would be nice to see a R3E better it. Do you really need 1.20vICH and 1.40vIOH for 238 BCLK or was that overkill?
Heh, sounds like it just hates that ram, which is odd, asus are usually pretty good with ram compatibility. Should make a good office pc mobo with a moderate overclock.Quote:
Originally Posted by OC Nub
Nice CryptiK!!
I seemed to hit a wall with the rex3 so I installed the UD7 and my tortured 920 and tested 32m. It wont do 1800c6 like the rex3, but does great at c7. Ive been trying 3 boards to see which I could get the best time with, so far the UD7 is doing the best. Im suprised the chip is still working, Its been tortured bad with high vcore on water.
Thanks I was pretty happy with it since it's a Rev2 IOH R2E.
Nice board that UD7, the uncore and especially QPI link speed is where my R2E is lacking, but I see you went to slow mode at 234 BCLK. I can't get to 225 BCLK without using it (at low Vtt, not sure if that affects it but I thought not). I am not prepared to beat on my CPU to see where the limits lie, but I'm quite certain it's the board given the low Vtt my chip requires for high ram clocks and high BCLK. 1.744v on water is :O that chip has indeed copped a hiding hehe :D
EDIT - btw TOOL_CHILE resize your screen capture, the sideways scrolling to read posts is really irritating. You can do it in W7 app or photoshop to name 2 options.
Installed my Yate Loon high speed fans. They are quite loud. :(
R3E only brings them down to 1500rpm on q-fan 'silent' mode... :'(
Here are my results, little older one 240MHz with Slow Mode, and fresh one 230MHz without Slow Mode.Quote:
Originally Posted by CryptiK
IMO, such high voltages are overkill, especially with Slow Mode Enabled.
Interesting, how can you have 1150 MHz for ram but only be running 3685.9 MHz uncore? On my R2E, at that ram speed the lowest uncore speed available is 46XX MHz. Does the R3E have lower uncore multiplier selection or what?
It's nice that your board will run QPI data link rate at ~8300 MT/s, my R2E will not post with anything above ~8000 MT/s, so I have to use slow mode from that point which is disappointing. At 240 BCLK our SPI32M runs are within 3 seconds of each other so with slow mode the performance is essentially identical (we have different ram timings which may account for it).
Gulftowns have lower UnCore to RAM ratio of 1.5x, so if I was using 2300MHz for RAM, lowest possible for UnCore is 3450, but I never use lowest multi cause it is less stable then lowest+1.Quote:
Originally Posted by CryptiK
Ah of course, totally missed you were using a gulfy.
guys, how much voltage are you using on QPI vtt for 24/7 use and around the 200 BCLK on a Bloomfield CPU?, i can't believe that my 950 and RIIIE ask for 1.4v and over that voltage to pass linX =S
I have R2E, but I'm using 1.18v for 210 BCLK. What is your uncore and rams speed though? It's not just BCLK that needs Vtt.
i try it to reach high speed rams right now and BIOS 1203 it's very good for that
http://valid.canardpc.com/show_oc.php?id=1570188
Uncore it's about 3.8 and memory at 17xx cas7-8-7-21 1T TRFC 110. Vdimm at 1.6v on G.Skill pi pc16000 cas6.
I pushed about as far as it will possibly go with slowmode disabled, it just bested my E770, XPower, obviously the Rex3 also, and the board is just over 1 year old. But just for a screen and fast 1m run, maybe on extreme coling someone could do 3D benches. This time I removed vcore from the cpuz screen, needless to say it was rather high :ROTF:. This was booted from bios, no software.Quote:
Originally Posted by CryptiK
Do any of the BIOS correctly control CPU Fan with Q-Fan mode? I set it to Q-Fan and the fan always runs at 100%...
Are the all off the same type ? I can control my Corsair fans with Qfan... don't have any other 120mm ones here
What is the best Bios for 3D Performance ?
My old EVGA SLI Micro Board has over 900 Poinds more @ 3DMark Vantage @ the same Settings as the Asus Rampage Extreme III .:(
Same clocks on cpu, memroy and uncore, more clocks on VGA and i got less points with RIIIE than my oldie UD3R WTH
I have just tested max bclk on my R3E and I m very pleased with it. :)
I hit 259 max bclk on air - with slow mode - PCI-E 115MHz - VVT 1,45V - IOH 1,27V
I think that I could go further, but QPI Link over 4100+MHz is so hard to get on this board. QPI multi in slow mode for 980X is much higher then on i7 920 and I hit 4100MHz QPi link on 259 bclk while on 920 it would be around 3000MHz
And here is 254 bclk 32m Spi run - air cooled - no system tweaks
http://img822.imageshack.us/img822/3...aphsot0001.png
and Everest - trial version of AIDA :(
http://img821.imageshack.us/img821/627/snaphsot0005.png
maxxmem - not optimised
http://img259.imageshack.us/img259/797/snaphsot0004.png
The 0096 Bios seems to be best for me in 3D and 2D.
:shocked: GIGA FTW !!! thats an awesome bclk mate congrats :up:Quote:
Originally Posted by OC Nub
I just installed the Rampage board into my Corsair 800D chassis today and was baffled and just a little bit incensed that ASUS elected to only put one USB header on the board instead of two ot three. The 800D has two cables for four USB ports mounted in the top front of the chassis. A standard practice for todays chassis designs. Two USB ports at the least. Why did ASUS not support those ports with this board? I have a work around but it means sacrificing four ports from the USB ports on the rear panel if I want to use all four front mounted USB ports. I just don't get why ASUS did this. Any body have a clue as to their logic?:confused:
Where can I download the 0096 Bios ?
Thanks Sergio.Quote:
Originally Posted by andressergio
Definitely overkill. I did honestly need 1.36v on IOH in order to boot into bios. Might be because of my beginner BCLK overclocking skills or not so great IOH chip.Quote:
Originally Posted by CryptiK
Sorry, didn't do any SPI runs. Truthfully it took so long to get 238 to boot I didn't bother with validations. Plus I already threw in the W3570 so putting that matter to rest :p:
Nice clocks there by the way :up: A bit disappointing to see a newer board struggle to surpass what the previous model can achieve :rolleyes:
Quote:
Originally Posted by jurky
WOW :eek: Amazing!! That's an unbelievably strong IOH chip on your board! Great results sir very very impressive :up:
Really nice result and board :up: They have been consistently impressive.Quote:
Originally Posted by OC Nub
238 is good, and unless you bench 920's and need every BCLK you can get, it's of no real consequence. Seems some R3E's IOH quality varies quite a bit. Gone are the days of the Rampage Extreme X48's where every MCH was high bin and performed really well, way better than any other X48 at the time (the best ones were just astonishing). Maybe they figure anyone benching these days is using unlocked or high multi CPU's so they can get lazy with BCLK and throw any old IOH on their 'top end' model?Quote:
Originally Posted by spoolindsm127
Now the worst R3E X58's are beaten by low end or previous model boards, which speaks for how Asus is running the show at this stage of the game. Sad really. That's why you see such a variation of results on the R3E, some are great, some IOH's should never have been used for these boards. UD7's appear more consistently good, so do the Classy 2 and 3.
I think Asus may be letting their marketing department run the show. If they make a good board that's consistently good, they boards sell themselves.
Were Can I buy this cooling soultion? LOL just kidding it made me laugh.Quote:
Originally Posted by [200]
"CPUID Reviewer Kit for Intel Sandy Bridge"
Contains CPU-Z 1.56.1
Shows VCore now :up:
Thx, IOH chip is golden one :yepp:Quote:
Originally Posted by spoolindsm127
here is one Spi 1m run with settings I used. I need to raise PCI-E to 117 for this run.
http://img202.imageshack.us/img202/797/snaphsot0004.png
But I m still having problem to past 260 BCLK, raising IOH and PCI-e frekv. more further and no GO.
1M @ 259.5 BCLK, very nice board/CPU. Was this done in windows or boot from BIOS? Tried raising Vtt? It could be your limiter although 1.45v is getting up there for a 980X no point hurting it for a screenie.
I boot from 255 bclk with no problems.Quote:
Originally Posted by CryptiK
I tried everythig for 260+bclk VTT up to 1.55V, IOH 1.35V PCI-E 120 and no go. I dont really need such high bclk, 254bclk was my goal because it is max for 2:6 cl5 on my memory for some competiton.
I m going to stop here on 259bclk to save a chip and motherboard from possible damage, maybe on LN2 sesion I will try to get more. But I m very pleased to know that I can go up to 259bclk on air if I need it in the future.
very nice jurky, could you tell us the batch of the gulf
some information about the 1206 bios?
Thx, 980x batch is 3013A557Quote:
Originally Posted by dejanbinladen
Thank you.Quote:
Originally Posted by CryptiK
The solution is here.Quote:
Originally Posted by David1947
They are all Yate-Loon SH 120MM. Q-FAN works for them on ALL headers *except* CPU FAN header. CPU FAN header is always 100% regardless of temps or Q-FAN setting.Quote:
Originally Posted by Leeghoofd
The RC bluetooth header is just a usb header. Just plug whatever it is you want into that instead of the bluetooth module.Quote:
Originally Posted by vladimir
If you want to keep the bluetooth module, then yea NZXT extender is the best option.
no one uses it anyway cause it doesnt work :shakes: :rofl:Quote:
Originally Posted by zoson
I use the NZXT and it's perfect ;)
Yeah, RC bluetooth has never worked for me.Quote:
Originally Posted by dejanbinladen
I do use the module though, so I have onboard bluetooth support.
The lack of usb headers has not bothered me though. I still have four unused usb ports on the I/O panel, and two from the front panel headers. Each of my two monitors has a usb hub with four ports also.
Quote:
Originally Posted by OráculoŽ
runs smooth thx for asking
That's good to know, any idea if it will be released soon? What are the changes? Changelog?Quote:
Originally Posted by Fire4Wire
I have terrible problems here with the freezing temperature of the cpu
I'm glad you said it yourself. It'd be a good idea to preserve your golden board. I mean 259 on air is absolutely incredible. I had the hardest time trying to get to only 238.Quote:
Originally Posted by jurky
I experienced exactly this and posted about the CPU Fan Header in the Asus Support Forums. That specific header is the only one that requires a 4-pin PWM fan for speed control. I've tested it and it's true.Quote:
Originally Posted by zoson
Hey all, I seem to be having a hard time getting more bclk. I have not worked too long on this new board but hope you can help. I was hoping to get 4400Mhz to 4500Mhz on cpu. If I go lower the multi I dont get the most from my memory. Can you see anything here with this info? I am more of a DFI guy so Asus is new to me. Thanks Gary
sorry here
Try using the 21x multi and setting bclk to 215. Try raising QPI and vcore will need to be in the 1.38v to 1.48v range for 4400Mhz - 4500Mhz depending on your chip unless you got a golden chip. Also, it is really tough to get 4500Mhz 4C8T LinX 20 runs stable. You will probably have to turn HT off to get 4500Mhz LinX stable.Quote:
Originally Posted by thesaucier
Quote:
Originally Posted by deachus
Thanks for the advice. I have lowerd the multi and tried to raise the bclk. I seem to have a hard time at around 200 bclk. Plus it takes my memory way out of its sweet spot. I would like to get there with just getting the bclk up while using the 23 muti if possible. But I will play with 21 multi some more and see it I missed something.
I have just tested max bclk without going to slow mode and I hit 234 max bclk, but after that its start throttling, and everything above that it seems useles. I used this settings 1,35V IOH, 1,5V vtt, and 117 PCI-E.Quote:
Originally Posted by spoolindsm127
I concur:Quote:
Originally Posted by Fire4Wire
http://img207.imageshack.us/img207/50/19307010311t2.jpg
Oh so you were getting 250+ with Slow Mode?Quote:
Originally Posted by jurky
I also had to set similar volts to achieve 238 BCLK. The ss i posted hitting 238 was W/O Slow Mode. When I get bored I'll see how far I can get with it :p:
:clap: Once I get back to my home and can setup my loop again I want to test my max bclk/qpi speed as well.Quote:
Originally Posted by jurky
Previous owner said 226 was the max, I want to see if I can at least replicate or pass that. 234 is quite nice.
Some one has offered to trade me a Classy 3way that does 245 bclk on water, it's tempting me...
You guys ever test a higher QPI speed than core speed? I have a friend who told me a while back his 950 and P6T WS was faster with 4.0 QPI and 3.8 Core for most things, but I forget exactly what he said.Quote:
Originally Posted by spoolindsm127
I haven't gotten the chance to test this yet, but I'd only do it if I could get both Core and QPI over 4.0
Also, what type of cooling are you guys using for the chipset with those voltages?
Actually this is the first time I've seen that mentioned, higher uncore than core. I have not tried it yet but soon will now that you mentioned it :)Quote:
Originally Posted by fl0w3n
CPU was on water but board on air. I have the optional heatsink on the IOH with the fan bracket modified so the fan is mounted on top, rather on side. Temperatures stay around 40c and never exceed 50c :up:
While we're on the topic what is safe voltage to push on IOH/ICH for air/water? I had my IOH up to 1.45v I hope I didn't degrade it.
Quote:
Originally Posted by spoolindsm127
Ah, yeah I've not seen it mentioned much either other than when my friend told me over AIM a while back. He isn't a "bencher" so he didn't show me any tests that proved what he said, but I trust his word based off the work he does.
Be sure to post your results if you do.
Okay, that is my setup right now, but I'd like to put my board under water for assurance. I may attempt to mod my fan like that as well if I don't do water.
Do you have thermal paste between the "fan sink" and the sink on the board? I think mine gets much hotter than that.
I think that is pretty high, but I have no idea the "safe" voltage spec.
Will I have got to here but still not stable for prime.
When I do test I sure will post :up:Quote:
Originally Posted by fl0w3n
My ambient is around 23c. When my fans are all set to 60% duty ICH/IOH temps are 43c/46c respectively. With fans at full speed temps drop to 38c/40c.
Yes I did use TIM when mounting the "fan sink". I put a small drop where the IOH chip would be located underneath. It's not exactly in the 'center' of that fan sink area.
I use Shin-Etsu X23-7783D, best TIM IMO. Thicker than some others so actually performs better if you use slightly more than recommended 'minimal' amount.
You will have an easier time passing 4500Mhz LinX if you turn Hyper Threading off in BIOS, its very hard with it on. Also, you need to set "maxmem" in LinX to ensure full stability.:up:Quote:
Originally Posted by thesaucier
Good to know, I'll be playing around when I get back home and have my full setup back together.Quote:
Originally Posted by spoolindsm127
All or nothing! lol.Quote:
Originally Posted by deachus
How does 4.2 HT on compare to 4.5 HT off?
I run 4400Mhz HT on and have run 4500Mhz HT off both LinX stable and see big points differences in Vantage with 4.4 HT on being faster and scoring more points in Vantage then 4.5 HT off.Quote:
Originally Posted by fl0w3n
Haven't taken the time to do any other tests. I run the 4.4 HT On 24/7 under water 35% fan speed for bluray video encoding no problem.
of course 259 bclk was on slow mode, I hit 4100MHz qpi link with slow mode on that bclk, without it will be around 5000+MHz qpi link and that is impossible.Quote:
Originally Posted by spoolindsm127
I cant go over 234 on air cooled CPU and motherboard because it starts throttling over 4200MHz qpi link speed. I think with better cooling on the CPU and board it will go higher easy.
Slow mode is ok for 2D benching and I get similar scores with slow mode and without. But for 3D slow mode is useless.
I did some testing of 4.4 HT On vs 4.5 HT Off in wPrime 1024M and Vantage CPU test to confirm my findings which I saw when running Vantage with HT On vs HT Off. Here are the results which show a big difference favoring the lower clock speed with HT ON.
4523Mhz HT OFF wPrime 1024M
http://i580.photobucket.com/albums/s...Prime1024M.jpg
4438Mhz HT ON wPrime 1024M
http://i580.photobucket.com/albums/s...Prime1024M.jpg
4523Mhz HT OFF Vantage
http://i580.photobucket.com/albums/s...FF_Vantage.jpg
4438Mhz HT ON Vantage
http://i580.photobucket.com/albums/s...ON_Vantage.jpg