Printable View
hey just for know how u can hide the voltage on cpuz...never saw before...then in windows vcore is higher than bios??that s strange..i did the same just for bench but was 1.587 on bios and 1.56 on windowsQuote:
Originally Posted by alcatraroz
full load tudo bem ok...
http://www.pctunerup.com/up/results/..._cpuzsub10.jpg[/QUOTE]
what try do you say man? ...wtf?
the Q746A500 are a great stepping see it and only by air
http://img408.imageshack.us/img408/6...minutoszv5.jpg
Whats the point of all that with a Geforce 4 MX 64Mb PCI, No offense, Just curious
That foxconn Mars board looks Sweeeeeeet!!!!
I've got my E8400 info right here, in my exitement with installing I forgot to take pics of the motherboard so I'll just have to nick some off google :p
http://i118.photobucket.com/albums/o...v/DSC00187.jpg
http://i118.photobucket.com/albums/o...vvv/MSIP35.jpg
http://i118.photobucket.com/albums/o...v/DSC00128.jpg
http://i118.photobucket.com/albums/o...v/DSC00132.jpg
Some pics of the HSF I use:
http://i118.photobucket.com/albums/o...v/DSC00096.jpg
http://i118.photobucket.com/albums/o...v/DSC00097.jpg
My Super PI 1m result:
http://i118.photobucket.com/albums/o102/bhavvv/Pi3.jpg
And my 100% stable clock:
http://i118.photobucket.com/albums/o.../maxstable.jpg
All my overclock results with voltages:
3600 Mhz @ stock Volts
4050 Mhz @ 1.3v
4275 Mhz @ 1.375v
Then my Ram fails :(
I have 4x1 Gb of G. Skill HK 800 Mhz @ 4-4-3-5. With one pair installed its good for up to 1000 Mhz at these timings, with 4 modules it can only get to 950 Mhz stable and doesnt respond well to looser timings or extra voltage.
the point,exactly is a "little" power supply,so,with the little geforce 4, the oc go easily without troubles with power supply. :up:Quote:
Originally Posted by Furiøùs
Ah, makes sense. Im hitting that wall now, my BFG 650 has hit the wall at 3.83 because if i put anymore on the NB or Vcore / VTT anything it will crash Prime / OCCT within 2 seconds. Meh, gonna get a PCPC 1k Watt powersupply, been using this BFG since 05 i believe heh :cool:
You need a P35 like the Foxcon Mars, Abit IP35 Pro, or MSI P35 Platinum for the best overclocking results on Intel CPU's.
The new X48 is also looking great, but it costs 2-4x as much!
My MSI P35 Neo2 was only £66 and is a great overclocker :up:
But I wish I had spent extra for a Foxconn Mars now, that board is Nucking Futs awesome!
My E8200 primed for an hr on 1.29 Vcore
http://i95.photobucket.com/albums/l1.../4Ghzprime.jpg
Primed for 2hrs.
http://img.photobucket.com/albums/v7...4ghzclub-3.jpg
wow not bad. I need a new powersupply i believe, I get freezes in the BIOS if i set the FSB to 1780. Meh.. 2 year old bugger it's gone long enough
bhavv: Why not download the updated, XS version of SuperPI. We need some more SuperPI 1M 4.5 GHz runs for comparison. :D
Meh, try a PC Power and Cooling 750 Silencer, Meh:DQuote:
Originally Posted by Furiøùs
Quote:
Originally Posted by hersounds
what try do you say man? ...wtf?
the Q746A500 are a great stepping see it and only by air
http://img408.imageshack.us/img408/6...minutoszv5.jpg[/QUOTE]
that batch is decent stepping like mine (q746a589)but its not the best i ve seen here some batchs much better and i am on air too:
http://www.pctunerup.com/up/results/...082252_485.JPG
Your going to have to show me where you get that cooler from... I want one too!Quote:
Originally Posted by Furiøùs
Thank you. I was wondering how to get decimal points :pQuote:
Originally Posted by unclewebb
I'll also try remove a pair of ram for testing above 500 fsb.
Well, that setting isn't stable although it gets to Windows @ 4.6+ Ghz. The next thing you click might have your screen stalled and hang right there :)Quote:
Originally Posted by Furiøùs
So all these Super PI screenies and high overclock without stability tests are more like cosmetics and fun rather than everyday pratical usage.
Hughy, if you live in the US i get all of my stuff from either www.newegg.com
Or www.FrozenCPU.com Frozen CPU is one of the best reseller for Fans heatsinks you name it :up:
From another thread i read that you need a 60mm fan, on the eVGA forums some guy did it, the SPP is exactly 60x60 so you can fit a fan on there. I heard its cooler and less noisy
Actually... my board came with a fan for NB. It's plenty big... and plenty noisy too!
All of ours did, but some are saying its not very effective. The performance:noise ratio is terrible, idk but im gonna give it a shot anyway, what harm could it do :)
Trying to find 24/7 again with correct volts, after eliminating reboot problem. 4.2 at 1.376 so far. Still dont know if better to use loadline or not...loadline does at least decrease vdroop (in addition to decreasing vdrop/sensor calibration) below what a stepped .016v sensor can detect.
I notice you're using custom and not small FFT/blend. Your FFT size is only 28K? lol
Hmm, it's too easy to pass with that :)
loadline reduce vdrop...that s for sureQuote:
Originally Posted by rge
Quote:
Originally Posted by aoch88
Orthos small ffts using same 8K over and over is not as efficient at finding unstable CPU settings as the updated prime 95 small ffts set for 8-64K. Orthos simply has not been updated for newer cpus in a long time, which is the only reason it just uses 8k only.
By using orthos custom 8-64K to test cpu, I am simply simulating as best as possible the updated prime 95, though granted with older 24.14 version.
If you think that my custom run is to easy, then in effect you are saying it is easier to pass prime 95 small ffts at 8-64k, than to pass older 8K only runs of orthos small ffts, which I have found to be exact opposite.
Finally got around to finding a prime stable configuration (9+ hours blend). It goes much higher if I'm just messing around. Runs hot as sin though - even without the IHS. Hits the low 80's on good water setup. :)
http://img91.imageshack.us/img91/865...imestabtw2.jpg
By the way, I'm loving this RAM - $80 for 2x2GB and it clocks like a beast! Got some Team PC2 9600 on the way (100 bucks after rebate at newegg) - hopefully I can go 1:1 at 7.5x573. Don't know what the processor has to say about that though!
I will specify 24K like you said to see if it's a lot harder but I don't think it will though.
custom, in place, small ffts, 8-64KQuote:
Originally Posted by aoch88
If you have never used orthos before, you need to know that just because it happens to be testing 28k at that point in time, it does not mean it is just set there only. Prime 95 is same way. How old are you?
Edit: you can see its increasing just by looking at 24K passed message just above.
I'm old enough to be your father :) Well, I'll give it a shot like you said.
let me know which one you get and how it works... also what are you using to monitor temperatures? I use speedfan but I can only see cpu temps in it... I dont think it reads the others on my rig... Anybody know what's up with that?Quote:
Originally Posted by Furiøùs
Are you using speedfan 4.34 beta 38 (newest) and hit configure, temps, and see if core 0 and core 1 there and checked?Quote:
Originally Posted by Hughyhunter
Another q746 :D Looks like 46 week is a nice low-voltage one.
e8400 q746a556
P5b deluxe (no mods)
vcore bios: 1,3375; cpu-z shows 1,316 idle and 1,280 under load
One question about vtt- is it better to set it to 1.20 or rather leave it to auto ?
Bios says that auto is a "safe mode"- should I believe him ? ;)
Now running on auto...
Actually I think I just got the miscommunication, sorry bout the snip:p: . I posted 8-64K and I guess you thought I meant testing anything inside that range...like the current screen shot of 28K, when I meant setting the test run at min 8K and max 64K, so it tests 8k, 10k, 12k, 14k, etc like prime 95 small ffts does.Quote:
Originally Posted by aoch88
I would say, HeHe, to keep the VTT on Auto... 1.20v is probably ok also, but the recommended voltage is 1.10v and the maxiumum is 1.115v, I believe. I keep my VTT at 1.10v and my system is stable at 450 x 9 @ 1.365v in bios; 1.25v NB.
I guess thats not bad, but its not the best. Ones that go up to 4050 Mhz on just 1.3v are better :yepp:Quote:
Originally Posted by Hehe
Guess who got one of them XD XD XD
<-------
I just need to test with only one pair of ram, I will try and do it this week.
Or is that just a problem with your motherboard having to set the voltage that high and only getting 1.3v sent to the chip? That seems a bit odd.
Mine wont stay stable at 3.9. I think i just need a new bios and i'll be good then
has anyone figured out why most e8400 temps shows up high?
I think my 745 is a decent chip also.I am just trying to decide if I want to go over 1.3 (1.34) 24/7
http://webpages.charter.net/chew_toy/DFI8400/O461.PNG
or under 1.3 (1.28)
http://webpages.charter.net/chew_toy...00/orthos9.PNG
Or if 1.4+ is safe for 24/7 which I would guess is needed for 4.2+. I can do 4.2 1hr of occt at 1.34 but will get a random crash. I haven't had my voltage any higher than 1.344 yet and dont plan to for a couple months to see how the chips hold up with the higher voltages.
I am looking at some data which shows pretty conclusively that core temp is not a meaningful number. Give me a little while and I will post this information so that others can draw their own conclusions - I am a little busy right now.Quote:
Originally Posted by rusty
i seriously honestly think you'll be ok up to 1.4v. i think some people might have just gotten duds or put more voltage into it than they say they did. a few people here are running 1.5v with no problems since they first came out.Quote:
Originally Posted by Ancient_1
great chip though, 745 seems to be a crazy stepping.
I will check when I get home... another question someone needs to help me withQuote:
Originally Posted by rge
Here is the thing about super pi... I was looking through people's posts of overclocks and checking the validater on xtremesystems.org and every one I checked it said invalid checksum... so either super pi is unreliable, everyone's systems are unstable, or I'm just reading it wrong or entering the info wrong every time... I dont know!
I got my 4GHz, Woot! 1780QDR.. hehe im happy now, got the big 4.0 :)
Scan tech support over on hexus did a test on a first batch E8400 using an intel motherboard and stock cooler and speeds. Idle temps were in the 30's, max temp after 2 hours was in the low 60's (stock cooler dont forget). Both Intel and scan confirmed that the CPU is working normally within its specification, and higher temps on other manufacturers boards are due to an immature bios.Quote:
Originally Posted by rusty
that's what I thought it's a damn bios issue just need more confirmation on that, thank youQuote:
Originally Posted by bhavv
How timely. I don't think that is the case. Check out this thread for my reasons: http://www.xtremesystems.org/forums/...70#post2793670Quote:
Originally Posted by rusty
so what should we use then to get the correct readings ?Quote:
Originally Posted by ixtapalapaquetl
Alright... need a look over.
Performed Small FFTs, In-place large FFTs, Blend... all 15 minutes no errors.
CPU CORE: 1.365 (I think... cant remember) I checked and it was at 1.4 so I set it to 1.375 and it's at load right now at 1.312v
CPU FSB 1.4 This is 1.3
SPP 1.4
MCP 1.6
SPP<>MCP 1.3 This is in Auto
How does everything look... I wonder if those temps are accurate... what is everyone else getting for temps at that overclock?
Should I try running my memory at 1T command rate? I set it to 2T for the overclock. I am running 2X2 at 4-4-4-15
grrrr got new e8400
1 core is 10° hotter then the other in idle....at load they pretty match up....wrongly appllied TIM?
Na, this seems to be a common thing with these CPU's, one Core being hotter than the other.Quote:
Originally Posted by TEDY
yeah cause on load they're matched +-1°c :)
e8400 pownz :)
coretemp shows 5°c more then speedfan, which one to trust ? :P
http://shrani.si/t/2J/SO/1qf42Zpo/e8400-3600.jpg
lol on idle mine are the same and on load Core 1 is around 61-63C and core 2 is around 52-54 :cool:
None of them are right. Just get Everest. Reads mine fine :up:
coretemp and everest give me the exact same temps
Here is my E8400 tray, Q746A482. I am relatively happy with it, it clocks decently but unfortunately cannot do 4GHz with anything below 1.414V real. It does 4.3GHz at the same voltage though.... Haven't and wont try any higher since the CPU sits in a crammed HTPC, which I would rather keep silent than uber-fast.
http://farm3.static.flickr.com/2146/...6daf0223_o.jpg
guys use preview not full pics....has anyone have abit ip35-e and e8400...mobo undervolts like crazy....
bios 1.305v idle 1,152v load 1,16v cpu-z
bios 1.385v idle 1,264v load 1,232-1,240v cpu-z
grr, iritatting.....
Wow thats really bad o.o
yeah i've heard of the abit p35 series really not keeping up with these chips.Quote:
Originally Posted by TEDY
which vcore is more reliable?
I bet for ramQuote:
Originally Posted by emoners
@Mykou: thanks! hmmm if ram is the case, then probably the vcore would need lesser volts if i pump up the vdimm or loosen the timings
Just received my 780i upgrade from EVGA ($89=ftw)
Here are my results:
https://publish.comcast.net/rpath/Wt...qR2QTeNDfLTqY/
That's a result of using a bios version that doesn't fully support the E8X00 series yet. So far the only one that we've got that works correctly is 16B09, that should make your selected voltages match up much better and get your temps inline if they're being reported funny by the bios / hardware monitorQuote:
Originally Posted by TEDY
Q745A567
This is what I run mine at 24/7.
http://i8.photobucket.com/albums/a11...6600_E8400.jpg
http://i8.photobucket.com/albums/a11...ixen/specs.jpg
Damn nice chip. I have a feeling you could probably do 4.2 and still be just under or over 1.3vcore (Windows).
Wow ! Nice Chip Vixx
Vixx, we have the exact same batch. They're great!
I think my 8400 just degraded, I did prime95 stable for 9hrs 2 weeks ago at 4.15GHz , V1.38
Now I need V1.39 to be stable , or maybe more
WTH with 8400 ??
if u have another board try on that one and see what happends, as my maximus began to act funny after right after I popped in a wolfdale, but the board is sold now. U sure your ram is stable in the first place?
Please try to run prime again.Quote:
Originally Posted by icecpu
You cannot call one failure not stable. You must try at least a few times..likewise the inverse is true, did you run prime for 9hrs just once? It's more than possible that it was a "false positive" for stability. Prime is not the end-all, be-all.
I had the same on my Abit IP35 - I think BIOS it is not ok yet, nad you have to wait for the new one :(Quote:
Originally Posted by TEDY
boo! where's the official e8200 overclocking thread : (
Currently running stable @ fsb 475 vcore 1.304 real, what do you think?
i'm using 14 so best way is to update to beta 16,right ?Quote:
Originally Posted by HaTaX
Unfortunately the IP35-E is not getting much attention, there is no 16 for the E version of this series. :(Quote:
Originally Posted by TEDY
There is a beta: IP35-E - V1.6 Beta 01 E8400/E8500 Support Improved at http://forum.abit-usa.com/showthread.php?t=122393Quote:
Originally Posted by ArCElM
yes it is :)
IP35-E - V1.6 Beta 01 E8400/E8500 Support Improved
http://www.lvcoyote.com/M630A_16.B01.zip
well i should upgrade bios hehe....
I find that for this CPU, I only need 1.31v for 4.0 (8x500) but to try for anything higher requires big voltage increases.
I know everyone's sick of hearing the d word, but the Q745A I have definitely needs more juice than it used to. It still does 8.5x500 (4.25ghz) 2 hours Prime stable, but it needs exactly .05v more vcore than it did before I took it over 1.4v.
This is the second e8400 of mine that seems to have degraded, so to prove this to myself at least, I started tracking this one right out of the box. I got it prime stable (around 1 - 2hrs) at about ten different clocks, using as little voltage as possible. I settled for 8.5x500 with 1.275vcore, which was priority 9 Orthos blend stable for 2+hrs.
I kept this log of settings around and after about a week of real world use testing, I decided to run some benchmarks. I did maybe 30 - 40 superPi and 3dmark runs, bringing the voltage up to a max of 1.42vcore. I never ran prime or any other torture test over 1.3125vcore.
After benching I wanted to see if the chip had taken a hit, so I went back to my log and to my saved profiles in cmos reloaded and ran prime again at the settings that had been stable. They all instantly failed. They began to pass as I notched up the vcore for each clock. It took a total of exactly .05v more at each clock to pass 2 hours than it had before benchmarking. It needed the same .05v at each clock.
So just to be clear this is not proof to anyone else, just what I've found using my simple note taking method.
The chip still does 8.5x500 prime stable with only 1.325vcore, so it's all good.
Another note - temp readings are fine and they scale up and down in sync before and after the possible degradation.
Quote:
Originally Posted by cadaveca
I ran OCCT and prime crashed within 2 minutes , so I just pumped v1.406 in bios (v1.39 in cpuz and everest) now it's stable, I will run it again in 2 weeks to make sure it's not degraded.
Well... You really cant expect a chip to last very long running it under full load for hours and hours on end. If you did that to old Atholon chips you almost guarantee that they arent going to last very long.Quote:
Originally Posted by icecpu
I only do stability for a solid 15-30mins with each test. Then it's stable in my eyes... if prime95 finds instability at the 10 hour it's probably because the chips so hot and soft and worn... but really who cares if it's not stable at the 10th hour... who honestly puts their chips at full load (ie gaming, movie making) for 10 solid hours...? even then it's not going to be full 100% load the whole time.
Prime95 should not be run for 24 hour stability checks... the chips were not meant to do that when being overclocked. I could see that at stock volts but not with more juice. :shrug:
Picked up a Xeon e3110 from Buy.com before they sold out. Looks like all of them were Q746A534 according to Reefa_Madness. Its slightly better than my e8400 Q746A510 in that it takes less voltage to get to the same place. Both run hot but the Xeon is slighty hotter. Swapped out my Kingwin/Xigmatek HPDT cooler for a GeminII with bigger CFM fan.
Pic in preview in deference to TEDY and any members who are low bandwidth or pay metered rate.
http://img261.imageshack.us/img261/1...eairsh4.th.jpg
3DMark06 fun with a vanilla 8800GT
http://img261.imageshack.us/img261/9...iniicv5.th.jpg
@mrcape. If you are using cmos reloaded to load your clocks, did you check to make shure that your "Clock setting Fine Delay" settings are the same as they were when you ran 9hrs stable. I have noticed that when i use cmos reloaded, the fine delay settings change from one load to another. If they are different, i also get Orthos failing very fast. But if i load from "cmos reloaded", shut the mobo down after the bios is saved, discharge the mobo then boot up again, the fine delay settings are then the right ones and it sails through Orthos with no problem. I would be prepared to bet that a large number of the "cpu degregation's" we are seeing, are actually due to the wrong "fine delay" settings.
i don't care i have flatrate 10/10 mbit/s but it's anonying to see those big pics :P
Hmm that's interesting and I'd love to explore that possibility more. You're my hero if you figure this out and I can apply a fix so that I can get it back to lower v. Not sure about it in my case as I had fine delay always on auto before and after the symptoms. Some of the settings I retested with came from cmos reloaded and some were manually entered.Quote:
Originally Posted by kitfit1
I did however try about 40 - 50 tests before and after clearing cmos, re-flashing, a fresh windows install and fresh ram.
Also I never ran 9hrs, max of 2hrs for me.
A little off topic, but man am I sooooo glad I bought my E8400 at the start of feb!
I paid £129 for it and was lucky enough to order a day before stock came in and got it two days later!
Look at the prices now:
http://www.scan.co.uk/Products/Produ...oductID=736334
http://www.overclockers.co.uk/showpr...odid=CP-192-IN
And I got a good core, 4050 mhz on 1.3v, 4275 mhz on 1.35v, but my ram isnt 24/7 stable at 475 Mhz fsb.
I'm going to test 4100 mhz @ 1.3v now.
Because i have fine delay set to more aggressive, when i first ran my 4.6ghz clock i noted down the fine delay values. So of course i know that those work for me, with all the other settings at that clock. If you did'nt note down your values, even on auto, then finding them again could be difficult. One way, would be to set up the rest bios for your clock and then shut down as i said in my last post. Then run Orthos and see if it runs clear.Quote:
Originally Posted by mrcape
I've tried shutting down with the settings applied, fine delay at auto always. The only thing I could do now I guess would be to tweak fine delay and see if I can go lower v orthos stable.
I have run about 280 hours of prime + orthos + occt tests, in fact mine has run stress testing practically non-stop since I got it...first trying to sort out a reboot problem, which ended up being a software issue, and then trying loadline versus not, etc.
80% of that 280 hours (documented, actually probably much higher) is between 1.3625 volts (cpuz) up to 1.39 volts (cpuz). And still boot voltage at 4, and voltage for 4, 4.05, 4.1, 4.2 for 12hrs+ stability is exactly same. I have zero degradation after 1 month of near continuous stress testing at vcore 1.36 to 1.39.
I think stress testing simulates greater use. By way of example and not implying accuracy, 1 day of stress testing might equal 1 week of normal use.
If I saw degradation using two chips running stress testing with high FSB and ? high corresponding VTT, I might conclude that these chips dont like running with high FSB and/or high VTT, and that stress testing brings this fact to light quicker than 24/7 use. But I would not conclude, especially based on my chips (granted just one, but so is his just 2), that stress testing in itself is damaging, unless your settings are a problem to start out with, and stress testing just brings this to light quicker.
Are people linking memory to FSB? I have mine set to SLI enabled and to unlinked. So that I can run 800mhz on the memory. It says it's running 9:8 ratio in cpuz. What do you guys think? Also I am using about 1.375v cpu for 1800fsb. Is this kind of high compared to what you all are getting?
Hi not sure if you were directing the comment about stress testing at me or not. I have no issues with stress testing, use it all the time.Quote:
Originally Posted by rge
Question about your testing -
Did you get a baseline to begin with to see how low V the chip could run at each clock before taking it over 1.3v at all?
I'm running non-SLI, and unlinked memory. Set to 1000MHz, with 5,5,5,15 2TQuote:
Originally Posted by Hughyhunter
flashed to latest beta16Quote:
Originally Posted by TEDY
bios 1.305v idle 1,272v load 1,248v cpu-z
showing much realistic now.....altho on stock cooler needs too much vcore? or am i doing something wrong? i see people here below 1.3v for 450x9 me 1.35-1.4v bios :(
lol, had to reply this one.Quote:
Originally Posted by bhavv
You got a good price although if you were a good shopper like me. You would have found it for £127 like I did, not forgetting they also have given me £10 voucher. So at the end if I buy it. I will be paying only £117. That is the best price in the whole UK. Well now I just need to decide if I will buy it.
So The E8400 for £117 or E8200 for £107 with a voucher discount.
Does somebody know if the reading temps affects also the E8200 or it is affecting only the E8400?
Metroid.
Had this Q745A703 sitting in my P35 DS3P for a few weeks. Have not had the time to bench but I knew from reading here and other forums that the stepping was a beauty.
Nice to see my one followed suit.
Water gear is on the way back so will push it harder over the next few days/weeks.
Hot little bugger though!
http://img150.imageshack.us/img150/5026/4gigtp1.th.jpg
Where from?Quote:
Originally Posted by Metroid
Several places still have them priced at £122~£130.
Just sold my second one on ebay so i might pick up an E8500
Yea I should have like bought as many E8400's as I could at under £130 then put them all on ebay for £175 or something crazy!!!
But I love my E8400, I got a good one ^^
1.30v, 455 Fsb x9, 4095 Mhz, 910 Mhz ram @ 5-4-4-12 (Cas 4 is unstable over 900 Mhz), and 62 degrees.
Just increasing the fsb by 1 mhz to 456 for 4104 Mhz @ 1.30v, and the CPU crashes in under 2 minutes XD
I found my 1.3v maximum :)
http://i118.photobucket.com/albums/o.../maxclock1.jpg
:clap: :up: :up: :up: :clap:
My ram is terribad over 900 mhz at higher FSB's. I can get my CPU stable at 475 x 9 for 4275 Mhz with 1.35v, but the ram BSOD's within an hour or two and fails PI 32m and Orthos, regardless of voltage and timings. But at 333 - 350 Mhz FSB it is stable up to 1066 Mhz.
I'm going to make my sig now for my max stable OC :p
That's great, the Q745 rox. :up:Quote:
Originally Posted by C0ncrete
Sig POWER!!!!
:rocker:
Thanks.Quote:
Originally Posted by mrcape
Have a feeling it can go 4gig with a little lower voltage.
Will be 37deg here today so no stressing this baby until it cools down later tonight :)
Damn, 37 is hot! I bet it can do prime at 4ghz with 1.2v-1.22v
Not you, but whoever was suggesting that your chips might be degraded from stress testing in itself.Quote:
Originally Posted by mrcape
After booting at default, I checked booting 4ghz at 1.25v and orthos crashes in 1-3 seconds, then many runs at 4ghz, orthos stops in 8-40 mins at 1.275v, and 2 long stable runs at 1.29. Then 2 long runs 4.05 stable at 1.3375, and orthos stops 10-20 mins at 1.325v. None of those points has changed. At every single voltage I list as stable, I have an orthos crash around 20 to 40 mins 2 notches below, and orthos crash around few to 10 hours at 1 notch below. If I degrade 1 notch I will be aware of it, but so far, not one notch at any stable setting even up to 4.1 and 4.2 ghz.
Now, at one notch below stable especially, I have seen massive variabilty, especially when running two identical orthos runs in succession. For example, (all runs at same settings in bios and in orthos) I run orthos at one notch below stable and it ran 8 hours first time. I immediately run it again, and it crashes in 17 minutes. So it degraded? The I run it again and it crashes in 3 hours. Then it idles for few hours while I am at work, and then run it again and it stops at 10 hours and 50 minutes, so it ?upgraded.
AT one notch less then stable, and sometimes two notches, I have seen massive variability, especially on quick reruns with no rest in between, but no degradation.
RGE, that's good to hear that you haven't seen any major change. I really hope I can sort out a good reason for this other than damage but so far not yet.
I need an even seven extra notches consistently across all of the clocks I had recorded.
I wouldn't expect to keep that stable for long, actually.Quote:
Originally Posted by mrcape
@bhavv: Very nice for 1.3V--quite nice in general, really :up: Looks as though you got a good chip. BTW, as you say cas @ ~4 gets unstable above 910 Mhz... what about cas @ ~5 and hitting around 1000 or 1066?
Thats what I have seen as well, strange variability that could be causing some type of thermal breakdown yet at least with my chip has no major long term damage as I can still run my 24/7 speed of 4500 at 1.45v going on a month now and bench anything like the day I fired it up. Strange indeed..:)
New gens should fix the fluctuation up, and we'll be seeing less chip individuality in all likelihood. Hopefully less issues in general, but who can say.Quote:
Originally Posted by flesheatinvirus