Printable View
SoF: Hope you get a probber mobo (or get the one you have now to run som high FSB) and get you ram tuned to the max.
That's a really nice clock you have got there :slobber:
With some fiddling around with your timings you should be able to easily beat my 3000+ with a vapochill SE:
http://i10.photobucket.com/albums/a1...perPIOCIII.jpg
can you please do me a favour and post links instead of pictures?
this is making the topic too complex and you see more pictures instead of text ;)
trying to get 26s but my ram is already maxed...whatever I do in A64 tweaker...it will get unstable, slower (no matter which direction I change settings) or crash...I tried some of my BH5 settings but they won't work...I think it's a disadvantage to only have CL3 cl2.5 will crash...CL2 the same...1T already...
here is my last run:
http://www.casextreme.net/pixx/bench.../27_5s_a64.jpg
suicide shot udatet in the first post ;)
it will need a lot more volts for OCCT and Prime...both are very similar when looking at stability...If OCCT will crash within 15min, Prime won't run longer than 15min as well.
http://www.casextreme.net/pixx/bench.../3059_occt.jpg
SuperPi will even run with 1.55V ^^
This was yesterday, currently I'm not able to reproduce this, I have to go to 3044MHZ :(
http://www.casextreme.net/pixx/bench...rime1_3044.jpg
nevertheless I finally broke 10k with m 6600GT...a laugh for all with 6800/X800XL and upwards but I think ok for a 6600GT
http://www.casextreme.net/pixx/bench.../10k_3dm03.jpg
I hope to reach at least 3100MHZ stable with some burn in ;)
Regards
SoF
€dit: Installed MBM5 and now OCCT reads my values, showing -63°C CPU on load, but I think it's ~ -35 to -40°C real.
I had another 3059MHZ OCCT but forgot to update the screen :slapass:
It's now running again ;) Seems to be related to RAM timings as well, because I have to make several changes to get it stable compared to my SuperPi settings...I have a small sheet of paper here for SuperPi RAM settings, Prime RAM settings and OCCT RAM settings....I always have to switch it to get the program stable...I'm happy to know what I'm doing, else I would go crazy right now :cool:
hi sof :) nice oc :D
try 2 beat me an reach 26,766 sec in super pi :P
good luck dude
Xtasy
I'm on it :D
But I can't pass Pi @3220MHZ atm.
Think I have to buy the TCCD Agent offered me in c³ :D
cheers mate seems both of us finally reached 3.2GHZ :toast:
SoF although I posted this photo in "CPU burn-in guide" I post here again to show to other users something strange... :mad:
http://www.xtremesystems.org/forums/...chmentid=38250
For any FSB > 265 Mhz :stick: StressPrime LOL shows anything... :D
Conclusion:
3.2 GHz rulez man :woot: borrow 250 MHz to me I want 2.9GHz 32M stable.. :(
thanks for this one - 32m and OCCT is a difference of 150MHZ!
I think it is the crappy IHS!
The temp sensor on the board isn't reliable in accuracy but it shows intersting things:
I will have idle shown temp of -85 to -80°C.
Pi will get it down to -70 to -65°C.
OCCT will run even at higher speeds as long as this temps are shown...but it will go below -59/58°C and stop!
This is the case with higher voltages and high speeds.
I decided to rip it of this night, will mod it a little bit (will show pictures), apply new paste, and re-attach it. I don't like to run without because it is more difficulty to mount the prommi correctly to a processors without HS to get display temp = bios temp until 0° (gigabyte bios stops there) than with HS!
I'm excited if I'm correct because only Prime and OCCT will get the processor down to these temps, Pi will not for any reason and runs stable.
And if I can boot without any problems @3280MHZ (I can start to bench but it will crash after a while but NOT immediatly) it should be possible to get 3.1GHZ Prime. I hope it is just the f*cking HS...
...Regards
SoF
My 3200 venice 3209mhz spi run
http://img361.imageshack.us/img361/7...ncer7rx.th.jpg
Redline limit /dfi rev ab0 limit
With tccd I can't go beyond 305 1:1 :( (yes, my centons advanced can do more than that)
damn I tried to get the HS off but I can't cut anywhere. seems the razorblade I used was crap :(
stopped before damaging the processor. will retry when I have a new razorblade :/
anyway I was a little bit angry this morning...prime just stopped after reinstalling the cpu after 38min while I was sleeping (thought 3030MHZ@1.7V would be no problem) so I tightend the screws more and see - I had 5°C better load temps and OCCT@3045MHZ is possible again.
It really seems to be pressure related...I won't get the IHS off this week-end. Hopefully tuesday I will do so...I have to be carefully I would kill myself when destroying this wonderful chip because of a silly HS :slapass:
:slobber: this is a great difference OK we all know that OCCT is more sensitive than 32M but for my system is 30 MHz only...Quote:
Originally Posted by SoF
Anyway each system is different but I think you are right --> IHS...
Of course you compare MSFs with same settings and Vcore? :rolleyes:
SoF if you fold try a folding program many people use folding programs for stability testing and tell us your MSFs etc...
Take a look here I created a thread for you! :)
big thanks Sean!
I did a lot tests today and for 1.65V it is a difference of ~100MHZ between 32m and OCCT...3090MHZ will pass 32m @3100MHZ while temps staying under -64°C.
As sooon as I start OCCT or Prime temps are rushing down to -63/-62 during the first minute!
I think we should consider the HS when asking people about Prime/OCCT problems.
I think the thermal paste AMD uses craps out under subzero and under high load...I can't even burn-in now because I don't want to risk burning in cpu with hotspots on the die ^^
There is no problem with A64 and prime in general!!!
I had 6 processors now and even my coldbuged winchester were able to run prime on stock voltages under prommi. under air they got ~2650MHZ....it seems to me that Prime and OCCT causing hot spots where the AMD's 90nm processors are very sensitive to...just my opinion ;)
hi Sof
did you try 2 tweak a little bit. well i did
http://www.forumdeluxx.de/gallery/da.../20356pi11.JPG
Excellent way to go.
WTH? NICE!!!!! something I have to beat hrhr
I'm preparing something xtreme ;)
Just got some tools for tonight's HS modding action -.-
Becaused I crashed my windows another time I will use another windows for the next runs as well...if I can pass 327HTT with windows tools it will be a hot night :D
Greetz SoF
mhh i think i´ve 2 go 2 buy vaseline ^^. crappy hs go away ^^. now i know why my mach 2 gt is modded 4 cpus without hs, but hopefully i wont kill the cpu. it is the first cpu on which i take the hs away :S
Edit
Sof i reached 26.00 sec in super pi :D now ill take the 25,XX . if ich rech them ill post pix :)
my HS is OFF!
its modded and reattached...but after 2 hours on working carefully and concentrated getting this biggest f*uck of hardware "gimmick" ever off, I was too struggled and just polished it a little bit, applied silicon paste and reattached it...first run --> screws not tightend enough (I'm always facing this problem with my self-made socket kit) and not even 2890MHZ @1.45V....I finally got it, but got no Pi 1m @3150MHZ because temps raised fast down to -55°C and Pi stopped...tried a little bit, but I haven't looked close enough on pressure between HS and DIE i guess...will do it again tomorrow...if it still won't work I will TRY to mount it naked...but this cute little DIE seems so sensitiv ^^
It will never stand against my Prommi, not thinking about it will always mount not 100% planar it will fix when tighten up screws........
...hate to know I could get 3270MHZ Pi 1m relativly easy when I can improve heat-transfer between evap and die...but I always have to risk the best piece of hardware I ever had and I would have no money left this month for buying a new one....this will be a loooong week (it's again 2:30 in the night in germany for all the ones in other time zones and I have to work tomorrow but I can't give up and think about nothing else - lol)
greetz SoF
ahh here's your 3,2 that couldn't break 27" barrier i wonder why i had that time 2 weeks ago with TCCD @ 3058 same chip that gave me the 26"
but anyway nice clocks there :woot:
here's the link of my 1st run with water from the kitchen 2 weeks ago :D
http://www.xtremesystems.org/forums/...ad.php?t=75640
:woot:
u have the same chip/batch of mine :toast: so i guess i can push mine a bit more..but i need some serious cooler :slobber:Quote:
Originally Posted by Leverpospade
slept 5 hours, back to work...I'm that finished, totally.Quote:
Originally Posted by |SiLA|
Huh nice watercooling ^^
I installed Win64 which gave me a good Pi boost I thought....31s (non-mod) @2750MHZ but I couldn't even hit 27s on 3100 ^^...maybe I was too tired yesterday...this is beeing non-stop overclocking since almost one week...I have to get this finished or I will loose my job...I'm always late for work because of clocking until early in the morning :slapass:
2750 is 30 :) modded and not :) tccd bh5 a bit lower :)Quote:
Originally Posted by SoF
Alittle slow for 3300mhzQuote:
Originally Posted by Xtasy
This is mine with bh5
http://img402.imageshack.us/img402/2082/silencer3hk.jpg
And this one with tccd
http://img415.imageshack.us/img415/3756/tccd6hk.jpg
Yes for 3,3Gig 26.X sec is a bit slow... ur right.
but i dont think xtasy tweaked his windows for spi.
with a little bit of tuning 25 seconds in 1m are possible i think.
raiden
hi raiden :) thats correct. ive got 25,xx sec in super pi :) ill show new pix
hi raiden nice to cu here ;)
so I finally did it - prommi mounted without HS!!!
HUUUUUUUUUGE load improvements of 10-15°C under low load yeah...high load improvements as well, first time I ever passed OCCT @3070MHZ @1.65V today...
IDLE improved by only 3-4°C but this was exactly what I expected.
PICS?
http://www.casextreme.net/pixx/bench...ce/hs_off1.jpg
http://www.casextreme.net/pixx/bench..._off_tools.jpg
and now the load comparison (temps on occt aren't correct but the differences between the temps is allright!)
BEFORE WITH HS:
http://www.casextreme.net/pixx/bench..._3032_load.jpg
AFTER WITHOUT HS:
http://www.casextreme.net/pixx/bench...occt_HSOFF.jpg
Think it was worth it....load temps over -60°C (OCCT) before, until now (3150MHZ @1.7V 32m try, facing problems with 1T but for 1m it can hold it :D ) I never saw it again....think I finally will make a prime run again tonight...time for burn in I think.
temps cant be correct ;)
any inprovements in oc?