A bad CPU simply wouldn't boot....it wouldn't cause OS errors....look at other components to blame.
Printable View
A bad CPU simply wouldn't boot....it wouldn't cause OS errors....look at other components to blame.
DFI LANParty nF4SLI-DR...... Hmm, my addition to this thread is I was able to OC OK, but now..... with my new RAM (RAM : OCZ 2GB PC3200 2-3-2-5 Platinum Edition RAM (1GB x2)) when I try to OC even 200 MHz with voltage up to 1.5 volts, it posts and then says that the RAID is good and all, then there is the big list of IRQ reservations or whatnot, and when it SHOULD say Vertifying DMI Pool..... Update Success, it does nothing, and reboots...
Any ideas to help me out??
Yan
Quote:
Originally Posted by Vapor
how can i blame on other components in my system when they ran fine with a 3200 venice for months. mind you i'm no joe six-pack and i read many threads about how to clear cmos, update bios, the proper procedure when installing a new cpu(with the exception of reinstalling windows), etc.. anyhow i'm rmaing this cpu because i want quality for the money i paid for it.
Yep I had this issue sparatically when testing my new DFI and OC testing, reset cmos and load optimized defaults and start again, might need some more voltage to RAM or CPU, just OC one at at time to resolve issue, leave RAM at low OC settings and get max CPU and then max RAM or vica versa.Quote:
Originally Posted by YanBooth
I had that issue and thought I had a bad mainboard didn't know what to compare to, but after changing multi and getting everything good like voltage for CPU etc I have not ever again had that exact issue you describe.
Maybe one of them stopped working properly or a non-WHQL driver you installed recently is mucking things up...the CPU isn't causing the problems.Quote:
Originally Posted by dng29
I used to have exactly the same problem, try re-flashing your bios with all the commands eg. awdflash BIOSFILE.BIN /py /sn /cd /cp /cc /LD /R /fQuote:
Originally Posted by YanBooth
only do this at default with maybe a little extra Vcore, Vdimm, chipset and HTT v
This fixed my problems after two days not even being able to load windows with a brand new board and CPU that would not even boot @ default :rolleyes:
Hiho,
try to set Max Async Latency and Read Preamble to Auto, oder use higher values.
For example if i use Max Async Latency 7 Read Preamble 4 with orange slots it works fine up to 260 MHz, with yellow slots (more stable for me) I've to use 8/5ns with 260. With frequencies above I've to use higher values for this setting, or set it to auto.
If I set it to low I've the same effect you describe.
I hope this helps.
Is not the CPU but you'll RMA anyway, you do realize overclocking voids warranty, not trying to insult you but just stating a fact. Unless I have a CPU that goes DOA on me I don't RMA CPU.Quote:
Originally Posted by dng29
I had BSOD and other issues while I was initially setting up my 4400+ reinstalled XP Pro SP2 after all my testing of ram / cpu etc and has been perfect since.
alright, i guess i'll have to reinstall xp. that's a pain in the arse though.
Really? Not to insult but how long you been fooling with PCs? Used to take me 45 min to 1 hr to install an OS in the past. I was HAPPY to install XP on my X2 took 8.5 minutes (NOTE this Excludes format, file copy during setup, reboot and entering keycode) basically 8.5 min where you are done inputing and it is completing setup (hated when people try to tell me 10 min etc for entire XP install)Quote:
Originally Posted by dng29
Anyway then Nvidia system drivers and NIC driver, with 6mbps download updates are 5 min to download and about 5-10 to install, video and sound driver, you're good to go.
Alienware's main selling point on their insanely priced $4000-$5000 PCs if you read their site is how much troubleshooting with hardware compatibilities and performance checks and possible bad parts as a "selling point" to buy their system versus buying theirs.
When I configured an SLI setup it was a bit over $5000 totally insane but for some people worth it, had a client wait 4 weeks from Falcon NW for SLI 6800U and FX-57 paid $5500. :banana:
A little update on my progress, but I did a 3D Mark 01 SE run and my X2 4400+ @ 2827 actually beat my FX-55 @ 3096! I mean not by much, but it still beat it. Here are the 2 links:
X2 4400+ @ 2820
FX-55 BN @ 3096
Same video drivers, same bios, sam RAM timings...
:woot:
That is a sick system mate.
Something tells me not all the settings were the same, SVTSnake. The 4400's run only wins in overall and Nature (by a large margin).
Quote:
Originally Posted by Vapor
Everything is the same! The only RAM timings that were changed was on the FX I was able to run at 8 and 1 for the DRAM drive and data strengths and on the 4400+ I have to leave both at AUTO. Everything else is the same!
It won 7 out of the 17 tests....
Then how did Nature jump 100FPS? This is puzzling me...LOD me thinks.
Nature is a GPU bound test...no CPU setting could do that....no CPU could do that.
Well considering I unplugged the FX and plugged in the X2, with only changing the 2 items in RAM settings that I said above, everything is the same. Maybe the increase in scores comes from the 2nd CPU doing all the background processes. I don't need to cheat on bench's, I could honestly care less. I thought I would just post a quicky showing that these X2's can compete with the FX's.
I really don't know what to say then...wow. All the CPU bound tests are higher on the FX, but Nature is higher on the X2....this needs more investigation, I'll see what I can figure out with my X2.
Lucky you that you got almost nothing to worry about on your pc... in my case, reinstalling means good whole day with all the applications, games, transfer of settings and bunch other configuration.... (mind you most of user files do sit on separate HDD but still takes full day).Quote:
Originally Posted by pentium777
Well here are my two runs, one with my X2 4400 and the other with the SD 3700+ in my sig.
The SD was at 3100mhz and the X2 at 3050mhz, guess which one was faster ? lol
The link works for the SD run in my sig, I have the free version so only can publish one
same memory, same multi, same settings, same drivers, same OC on video cards, only difference was 5mhz more OC on ram for the SD
Nature's the same...the other stuff changed more. Good work Major!!
As for why SVT got a massive Nature boost...IDK, maybe some bugged run?
As does me, but while it is a pain in the butt I have gotten used to it, I have a server and images drive for reloading software, CD tower for organizing and a 200GB HDD that makes daily differential backups and a weekly full NTbackupQuote:
Originally Posted by jinu117
Nice. Reloading software... wish SQL Server, Visual studio, Visual studio.net, etc as well as few candid utilities I like to use are like that. Best I can do is on my data drive, shove them in orderly fashion so that I just go through directory of installation in order. Maybe dual core will help me on this one eh? :PQuote:
Originally Posted by pentium777
Excellent X2 owners rejoice! One set of scores might be a fluke but 2 of them is quite likely :fact:Quote:
Originally Posted by Major
Major's is definitive. SVT's is a fluke...his nature score is the only thing bringing the X2 up and over.
From what I see, Major's test has all 7 tests with X2 winning, the left side 3051mhz is the X2 and every score is higher including nature.Quote:
Originally Posted by Vapor
Oh nevermind you mean "his nature score" NOT meaning Major but that other guy gotcha
Yes. I know....look at SVT's now.
Maybe it was a bugged run, but it still won in 7 of the 17 tests...
I was having power issues at first with the FX and dual 7800GTX's before I swapped out the OCZ600w to teh PC P&C 850w. I don't remember when I ran the FX bench if it was before or after the power supply swap.
I have 2 other runs on the FX with both other runs about the same in nature scores. 1 was with the GTX's and 1 was with my 6800GT's highly overclocked..
I'll run a few more tongiht and see if there is any difference...
anyone got a 4400 x2 with a bad second core? btw, reinstall winxp and everything is ok now, sheesh. anyhow the second core on my 4400 x2 is bad :(.
How are you determining this "Second core is bad" there is extensive information about X2 floating around. Basically you WILL find that while you will fail on a particular core during dual prime you can prime all day all night on that ONE core, same OC speed and it won't fail :fact:Quote:
Originally Posted by dng29
Reason for this is that dual core priming generates quite a large amount of heat/stress, generally for my purposes if I can dual prime for say 8 hrs without a problem then get an error on one core after say 9 hrs for example to me that's good enough stability. Think about when in the near future you will have an SMP app push both those cores 100% for 8+ hrs, probably not going to be very soon.
Do definately test both cores separately for a long time to make sure though
i did test both cores seperately and at the same time. cpu0 or core 1 can run prime95 for hours. cpu1 or core 2 failed prime95 after the 1st iteration. this is on default speed btw. i'm not bashing amd or anything here, but this dual core seems defective to me.
Ok well that certainly is the first I've have heard of that yet. Sorry to hear that :( This reason is why I ran my X2 (first AMD I've owned since the pre-1ghz days) at default for a couple days of testing to make sure I had nothing defective @ $632 it marks the second most expensive CPU I've ever purchased (I think I paid $700 for a P2-300 way back in the day) I wanted to ensure my chip wasn't defective and if it was there would be no way they could throw the "you overclocked it out of warranty" thing on me.Quote:
Originally Posted by dng29
I am still unsure how they know but I am confident that there has to be a way for them to be able to examine the CPU for overclocking since it is so well known now. Whether it is worth their time to check RMA'd CPUs or not I don't know. Sorry to hear about your problems really sucks
Hello companions, I don´t know if to buy a 3800+ x2 with following step:
ADA3800DAA5CD
CCBWE 0525XPMW
1245498G50098
or sandiego 3700+ of the week 0517UPBW . What would do you?
The processor would go exclusively destined to play with one 7800 GTX.
I wait for your answers with attention.
A great greeting to all and thanks for everything.
I seriously doubt there is anyway for them to tell if something has been overclocked or not, if there was a register or something it would immediaetly reset when it was powered off... and i really doubt they are going to put some kind of rom or something in there to hold the frequency the chip was last running at... and otherwise they'd have to take the entire chip apart and examine it to see what happened, which is also extremely unlikely.
what do you plan to do with it? Do you mean play games?Quote:
Originally Posted by yoni
What is your upgrade tolerance? i.e. How long would you have to use it before upgrade to not feel like you wasted this chip?
well i'm the unlucky one. this is the 1st most expensive cpu i've ever purchased and i don't want a defective product. afterall, amd means quality right. i don't care if they check the cpu once i return it back to them. i didn't do anything or inflict any damage to it. it still works, but not stable dual core. i don't want to pay full price for dual core and only get 1 core working.
I agree, I hope you bought the retail version of the chip which is boxed and includes the HSF, that way you can do the RMA online with AMD and get a replacement directly, OEM is not something I'd ever do after reading the pain to get a replacement not to mention only like a 90day warrantyQuote:
Originally Posted by dng29
Well I was deciding a while back on 4400+ or FX-55 and I took the gamble of hoping to have a fairly large OC from 2.2 to 2.6 under air and it paid off in spades. I have all the FPU gaming performance I need with another core standing by. Let me tell you it makes a huge difference in XP to me, used to be on startup of XP I'd hear the standard windows theme song then 5-10 seconds my audigy sound music now there is approx 1 sec delay and sometimes one is almost overlapping the other. Everything in XP much snappier and fluid, and I have the gaming CPU I wanted with my XFX 7800 GTX Extreme 500/1400!Quote:
Originally Posted by yoni
Finally I get a good chip. My 4400+ is rock stable at 2800MHz@1,488v, IHS lives on as a keyring....
From your sig I am assuming watercooling? hmm maybe some of you with air cooling or water cooling can answer this for me so I can compareQuote:
Originally Posted by type_a_positive
What temp do you run at dual core priming on X2 4400+ >=2.7ghz 1.6v or higher. I am curious since I have the XP-120 and extremely good airflow I'm still around 55-58'C depending on ambient temp and how cold AC in room is set. 1 core prime I am around 47'C and IDLE is around 37-39'C
Specs 246x11 2.7ghz 1.586v X2 4400+
Just want to make sure I don't need to pull the HSF / chip and redo it, in comparison same case and airflow my wife's PC (P4-3.2EE runs 41'C priming) and it's at 1.6v overclocked from 3.2-3.46, using the XP-90C, that same chip with my old thermalright pre-xp-90c did 3.35 stable prime
Does anyone think the XP-90C is better than the XP-120 I've read that the XP-120 was best due to large fan etc and massive amount of fins on it. Is it worth the time to pull HSF off both and try the XP-90C on my X2 to see if I can lower temps?
you must be brave to take the ihs off :). for something this expensive, i'd treat it as a piece of gold :).
Right now with 1.6v and watercooling, dual priming @ 2.82 its running about 48 degrees with IHS on. I'll probably take off the IHS tonight, and will probably be about 39 degrees while stressing.
At idle its about 36 degress...
Damn well I hope someone with good air cooling can confirm today what they are running for temp, because that's quite a bit difference of 10'C but it is watercooling though. I just would rather avoid the hassle of removing the XP-120 and XP-90C if it really is a nominal temp for good air cooling.Quote:
Originally Posted by SVTSnake
I'm having problems OCing my 4400 X2. Even if I put the voltage up to 1.61v I cant get stable at 11x237. CPU burnin with errors on works no problem, but prime94 and superpi timeout after a minute. I have the DFI N4 Ultra-D board. I have the big tornado for a cooler. What else could be causing it to time out? Is my 2nd core that bad?
1. Sell your mobo cpu and ram on ebay and buy a Pentium D.Quote:
Originally Posted by spiroh
OR
2. Supply more info on your system so someone can actually help you.
Two things I HIGHLY doubt Dr. Ruiz would ever say...buy a Pentium D or give us more info so we can help the OC?!Quote:
Originally Posted by Hector Ruiz
Sorry...your alias is just, ummmm, not appropriate unless your are in fact THE Hector Ruiz.
DFI NF4 Ultra-DQuote:
Originally Posted by Vapor
AMD 4400+ X2
2x512 OCZ Value [OCZ4001024WV3DC-K]
eVGA 7800 GTX
2x Hitachi Desktar 250GXP RAID 0
BIOS 623-3
SB Audigy 2 ZS
Seasonic S12 600
There's my info. Let me now if you need anything else. I dropped the voltage back down to normal, and went back to 2.5 where it is stable. Prime95 works there.
Any sort of memory divider? I'm not sure your memory can handle 237, tbh.
I wasn't using a divider. Would that affect Prime95 though?Quote:
Originally Posted by Vapor
Stepping of your CPU would help greatly. Was it retail or OEM also.Quote:
Originally Posted by spiroh
Sorry, I dont have my stepping it's a pain to take off my heatsink so I can get that tomorrow. It was OEM purchased from Newegg 2 weeks ago.Quote:
Originally Posted by pentium777
Well I dropped my memory divider to 9 and so far it has been running longer than before. Im thinking I might need to buy some OCZ Platinum.
Hi all,
been playing with my new rig about a week now, replaced my shuttle xpc with a dfi sli-dr(latest official bios), X2 4800+, good old 2x512 EB 3700, and a xfx 7800gtx. PSU is Tagan TG580-U22(580W).
I've replaced stock hsf with a zalman reserator I had from previous setup which resulted in a couple degrees lower temps @ load.
Starting from stock voltage and using S&M(latest version supporting dual processors/cores) 2xprime95 and 2xsuperpi 32M for burning purposes and stability testing, managed a stable frequency of 2580MHz.
Upping the Vcore to 1.43(1.4 reported by smart guardian) resulted in a stable frequency of 2712MHz(12x226).
Load temps using S&M(it heats the processor much more than 2x primes, at least in my case) are 54 degrees max for cpu and 55 for pwm(smartguardian readings).
So far so good, I've decided not to try with higher Vcore with this cooling.
Next step was to lower the multi to 11 and try matching the same cpu clock by clocking the memory.
After hours of memtesting and bios tweaking, found the limit of the EBs @ 240Mhz(while maintaining a good unbuffered BW) with 2.9V(from bios). Since I dont use any active cooling on them yet, just coolermaster ramsinks, passing 9 hours of test5 and test8 was a good point to check how this config would do in windows benchmarks.
Here comes the funny part.
Although one instance of superpi 32M would complete without errors, 2 instances would crash the one running on the second core. Running an instance at each core alone, will complete fine.
Same thing happened with primes custom test(I chose custom to test the memory and set 400MB for each instance to use, since blend test uses 768MB, no point running two instances of blend test with 1GB memory), testing each core alone will run fine for hours, but two instances at the same time will crash the one running on the 2nd core.
In superpi case I've set the affinity to each core from task manager, for prime through the advance menu.
Since the crashes on the 2nd core while both cores are tested, occure only when overclocking the memory and not at the 12x226 config, may indicate unstable mem clocks.
I'm not 100% sure though if that's the case, since I dont understand why a single instance of superpi completes without errors and two of them crash. When it comes to memory part, the only difference when running a 2nd instance is that it uses another 268MB of memory.
The same thing is with prime. Testing 768MB with one instance at each core, runs for hours. Assigning 400MB for each instance to use simultaneously, would crush the one running on the 2nd core.
Maybe the answer is that the mem controller is stressed much more while two cores are at load and both of them run a task that involves memory, but I'm a nooby when it comes to A64 setup.
Thanx :)
Thanks man, still running at 2.6ghz and I keep lowering the voltage with no problems. Looks like it was the ram all along.Quote:
Originally Posted by Vapor
Well, I have the 2GB kit of OCZ Platinum, and well, ever since I popped it in, I can't OC at all :(Quote:
Originally Posted by spiroh
Yan
Bad ram? Did you set it to 2T?Quote:
Originally Posted by YanBooth
No, I havn't set it to 2T. The RAM is great, improves BF2 frame rates a lot, but after I put it in, I can't over clock past stock speeds.
I am so pissed.
Yan
That memory should OC easily with your setup, mine runs 240 at 1t cas2.5 other timings at default, I have a very similar setup, running 623-2 myself.Quote:
Originally Posted by YanBooth
Maybe I should try switching the slots over? I am using the orange, and 623-3 at the moment.
Yan
It is after all PC3200, lower latency 1GB sticks "seem" to be tougher to raise the clocks on. I'd really suggest PC4000 sticks, Crucial Ballistix seemingly the best currently.Quote:
Originally Posted by YanBooth
I moved from a GB of G.Skill's best, but only 1 GB wasn't enough.
I tried resetting the BIOS and stuff, upping the LDT voltage (someone said to try that) and it won't get past that part, I am going to try the yellow slots right now.
Yan
I've tried both slots myself and really didn't see any difference but these boards seem to have a personality of their own sometimes, I am currently running the orange slots though.
I've had two sets of the OCZ 2gb plat kits, the second set I got not long ago even did 235 at CAS2.5 other timings at default so the memory should have a decent amount of headroom left.
Have you tried bumping up your ldt voltage or any of the voltages yet?
I have tried getting dangerous with the voltages... All the way up with the LDT voltage, up to 3.0 volts with the RAM, 1.5 volts for the CPU. What I have found out it, it WILL overclock just a little, I raised the FSB to 205 from the default 200, with the multiplyer on default (11x) and it worked, but since it won't get to the "DMI Pool" thing on 2.4GHz which it did fine with my G.Skill. This must be the RAM, right?? What do you think would be a good plan of action, if this cannot be resolved? Just send this in, and pay another $50 in restocking fees etc. or get a replacement, could it just be that my RAM is retarded somehow?
Thanks!
Yan
Hmmm, I wished I knew of some quick trick to try but the two sets of that ram I have worked without any issues when OC'ing.
I do know that neither set of the plat 2gb kits I have responded to voltage higher than 2.8vdimm so I wouldn't worry about getting to crazy there to achieve good speeds.
Strange, it very well could be the memory, if you just have them exchange it there shouldn't be a restocking fee unless you return them for a different item or credit.
I've never had any problems with mine by just slapping them in, no special tricks or voodoo rituals so there may very well be an issue with the ram.
Have you tried OC'ing using dividers for the ram just to be sure the board and cpu are OC'ing fine?
What do you mean by that? I am still pretty new the the whole overclocking world.Quote:
Originally Posted by highoctane
Yan
Running your memory at a lower speed than the htt, if you go into your memory options you will see an option for it, probably the first of the memory options.
Hey guys,
Happy to report some good news, I was at 246x11 and now I'm happily at 250x11 (2.75) on my Retail 4400+ :toast: The change occurred from switching XP-120 to XP-90C (should have done a bit better research)
I swapped because my wife's P4-3.2EE @ 3.35 got the XP-90C when I got my XP-120 I noticed her temps went from 114-116'F down to 100-104'F at full load. Thinking my temps were too high on my X2 I switched out and voila 5'C temp drop at the minimum and when I have my room AC on it's even larger up to a 10'C drop if cold enough.
Had issues with video card cooling went from 50'C at idle to 57'C when CPU was at full load this was due to the heat blowing down on HSF (which for my testing was the best cooling method for the CPU. Obviously this hindered my GPU overclock so I now have better positioned fans, my side panel of my case is off currently and I have a 120mm 104CFM fan blowing at the side of the XFX 7800 GTX which allowed me to go to 500/1400 speeds with no artifacts. I was able to bench at 520/1420! highest 3dmarks05 has been 9239
For those wondering single core or dual core and thinking gaming is affected..... The proof is right here in these two urls :fact:
First URL is mine running 250x11 (2.75ghz 4400+)
http://service.futuremark.com/compare?3dm05=1116394
Second URL is a FX-55 SD running 280x10 (2800) Same video almost same clocks on video also
http://service.futuremark.com/compare?3dm05=1046872
Any doubts left out there still about whether to buy single core or dual core, I think it was only gamers that had the dilemma, shouldn't anymore, the only real issue is if you can't OC your 4400+ to 2.6 but at 2532 my X2 was comparing just fine to any 3dMarks05 scores out there :woot:
Note: Don't want to violate any posting regs here but want to help people out, I just found ZZF (Zip Zoom Fly) has the 4400+ Retail for $557 free 2nd day shipping! Lowest price I've seen yet, paid $632 for mine then in one week it went to over $700 at Newegg. With this price available you shouldn't even think about it, but of course that is only if you are shooting for a higher end CPU and you were considering an $800 FX-55, since this is XtremeSystems I assume most people want the best they can get within $$$ reason.
Finally got my 4400+ stable at 240x11 1/1multiplier. Man, I would love to hit 2.7ghz. My temps right now are hitting 48c max. Is that pretty good for air? voltage at 1.43v.
Can somebody explain how I can run 2x SuperPi?
Copy the entire directory. Then run each instance - optionally assign affinity if you want.Quote:
Originally Posted by the farmer
Really low voltage there, I couldn't even come close to that voltage, your 48c temp if it is with dual cores priming then that's an excellent temp. For me to hit 2.7G I had to raise to at least 1.5 to boot up then 1.55 for priming and to get more stability I did 1.586v, since I was around 55c I didn't want to pushQuote:
Originally Posted by spiroh
Now that I've done some better cooling options, still air, I am getting 2.75 250x11 but had to raise voltage to 1.55v*104% = 1.61v temps are right now as I am typing and priming at 53c but room AC is blasting and side of case is open so ambient temp is optimal. One core failed at 2.75G in 30 min, running only 1 core testing provided no failures.
If I can get dual core priming for several hours before 1 core fails I will probably keep this setting as I don't want to push my vcore any higher and I won't be taxing 2 cores at 100% for hours anytime soon anyway. When I prime just 1 core the temp drop is dramatic usually 5-7c which when testing either core is 100% stable.
Thx, this works :) .Quote:
Originally Posted by NickK
Well it looks like my ram wont go any further than 240. Truthfully I was expecting just a bit more. How can I get my speed to 2.7 yet have my memory fsb close to 240?
You can use a divider in the bios, basically you'll do a higher HTT and then use the divider to lower the ram down to whatever you can get close to your max, wish I could tell you what divider settings to use but I haven't done so and this is my first AMD rig all others have been Intel and I never used dividers because the last few CPUs weren't great overclockers.Quote:
Originally Posted by spiroh
If you do a search on ram dividers you'll most likely find a wealth of information out there and other people on this forum can help too with that
Oh and what ram is that you are using? Maybe you could loosen the timings a bit? Are they 1GB modules? For me I'm running 2x1GB ballistix tracers and they paid off bigtime was able to bench them at 289 but stable was around 282 unfortunately my chip doesn't seem to like the 10x multi so I backed down HTT and raised to it's normal 11x and got excellent results
I'm using OCZ VX value. I'm seriously considering going with different ram.Quote:
Originally Posted by pentium777
Isn't VX supposed to be the really high end stuff? that's just 2x512 right? Thought that stuff was TCCD or TCC5?Quote:
Originally Posted by spiroh
Oh wait, you said "value" hmmm what are they designed for? just PC3200 and tight timings, if that is the case usually tight timings mean not very good overclockers, ICs designed for looser timings and higher clocks seem to do much much better with exception of things like old BH-5 and TCCD etc
Hey pentium777 how much louder is your rig now that you are using the XP90 that uses a smaller fan that 'should' make more noise :)
I have found my XP120 great because of the near silent running
Oh and BTW the VX stuff is similar to the G.Skill GH I think which is new BH-5 :)
Well not too bad because I am using a 92mm 72CFM fan from sidewinder linkQuote:
Originally Posted by alpha0ne
http://www.sidewindercomputers.com/de92ff1.html
^^ Admins if this is against policy please remove it, it is purely for informational purposes definately not advertising.
Only 42.5db which might be too loud for some people I don't know. I do know that the crappy ThermalTake fans that came with my Tsunami Dream while very very silent pushed practically nothing for air. Cost me some $ to replace the fans but it is really a Tsunami in there and my clocks for video and cpu are fantasic for air cooling.
As far as overall sound, I have a P3-800 as a server in here full tower 7 HDDs, my rig in sig, my wife's P4-3.46EE which her case matches mine same cooling since she has a 6800 Ultra, Astaro Firewall Mini ATX case 2.4Celeron couple RPM adjusting fans in there, but the biggest noise of all is the Windchaser 9000 BTU Portable AC unit this overrides all other noisemakers in the room so when it is on, which is most of the time I'm in here I can't hear any computer fans. All of this equipment in a 12x10x7 L/W/H computer room
I've spent most of the morning dual priming at 2.75G 1.61v and temps are around 52-54c CPU, 54-57c PWM IC, 38c Chipset, also I am in FTL, FL it is 93'F outside! and humidity probably close to 100%, when I turn off room AC and turn up house AC to 83'F CPU is 60-62c, PWM IC 63c+ etc this is with dual core priming though, will be finding out what it is at idle Monday when I go to work.
Unfortunately with a 210G saltwater aquarium fully decked out with a chiller (cools water), 900Watts of lighting, Pool Pump running 5 hours a day, HDTivo and all the computer equipment with the AC my electric bill skyrocketed to a $287 lifetime high for me. In comparison last year at my condo before buying the house with wife etc I had a $100-$120 electric bill during the summer :(
This MAY be because I don't have the AMD CPU Driver installed, I am not sure but here is what I found with my troubleshooting.
I run XP SP2 with 2 workstations and a Windows Server 2003 Standard server with Exchange 2003 installed.
I noticed Event Viewer messages with the following information:
Event Type: Error
Event Source: Userenv
Event Category: None
Event ID: 1054
Date: 8/13/2005
Time: 1:21:08 PM
User: NT AUTHORITY\SYSTEM
Computer: PENTIUM
Description:
Windows cannot obtain the domain controller name for your computer network. (An unexpected network error occurred. ). Group Policy processing aborted.
Naturally you will only see this if joined to a domain controller. The fix? well I tried many many fixes but finally at the end of my troubleshooting and using gpupdate /force over and over to test, I found a racing condition existing with the dual cores, used my patent pending explorer.exe set to one core fix. Voila when the gpupdate.exe runs (runs so fast I had to turn on dual prime to slow it down to confirm only one core was selected) it does not error and shows:
Event Type: Information
Event Source: SceCli
Event Category: None
Event ID: 1704
Date: 8/13/2005
Time: 1:21:08 PM
User: N/A
Computer: PENTIUM
Description:
Security policy in the Group policy objects has been applied successfully.
Now just for more info, this issue only occurred occasionally definately NOT everytime it is very random and the reason I investigate so much is because some of the solutions pointed to NIC communication issues, naturally with online gaming and my home network which I use for work also I did not want networking issues, also I'm using the Marvel LAN Adapter because I read somewhere it used less CPU.
I ran the group policy update 20 times in a row and not a single error in application viewer, definately a racing condition which was one idea someone had on eventid.net but he didn't provide much info on it really. Of course I really don't care about the group policy erroring occasionally I don't have a need to have group policys on my home network and the fact it errors once in a while doesn't mean much to me now that I know the reason.
Thought I'd put this message up because it kind of goes with the other racing conditions mentioned here for gaming and how setting one core to the application / game corrects the issue. Hope this helps people out and saves someone time.
They are supposed to be BH-5. I'm thinking of buying some TCCD ones.Quote:
Originally Posted by pentium777
Talk to OCZ and see if they can help, been very helpful to me over teh years.Quote:
Originally Posted by YanBooth
Confirm with prying heatspreaders a bit and look real close with a flashlight or remove the spreaders (probably more difficult) I never removed mine from my Corsair PC3500 BH-5 but I pried them enough to verify BH-5Quote:
Originally Posted by spiroh
I think you will find with a little more research (I dont have the link handy ATM) that the Nvidia native controller uses fewer CPU cycles :)Quote:
Originally Posted by pentium777
I am wondering if someone here can help a bit ...
I am trying to oc my x2 4400+ on an Asus A8N Premium with 2x512 G.Skill La memory and i'm having a couple of questions and issues.
1) Asus probe 2.23.026 indicates Vcore of 1.344 when i choose any Vcore settings between 1.3 and 1.4625V in the bios. Then when i choose 1.475V in the bios 'asus probe' indicates 1.456V.
Does this happen to anyone else with an X2 cpu and an A8N SLI Premium motherboard?
2) Bearing in mind the above is happening, i am forced to use a Vcore of 1.475V in the bios, reported to be 1.456V by both Asus probe and Speedfan. When running at 10x250 =2.5GHz with 2 Prime95 (1 on each core) i get temperatures of 56 Celsius when the motherboard/chipset temperature is reported 40 Celsius. That Vcore and the temperature should be ok (i.e. not too high) or? What voltages do you use with x2 cpu's and what temperatures do you get? what is considered safe?
3) I have not tried anything above 10x250 = 2.5GHz although bearing in mind that my x2 4400 does 2.4GHz with Vcore 1.344V (reported by asus probe and Speedfan) i guess that if i leave Vcore to 1.465V (as mentioned in 2 above) i could try 255 or 260. But then would the stock AMD heatsink/fan still be ok, or should i use something like an Thermaltake XP120?
IHS still on. Nothing extreme on the ram , but it's generic samsung chips that don't OC for squat anyway ...
http://199.202.73.113/mb/0365578/X2/2.8_X2.JPG
I'm getting a new Swiftech Storm block tomorrow so let's see if this can go higher .
Well, I fixed my problem! I took my DRAM:FSB from 200 (1:01) to 166 (3:05).
I wish there was a way to fix that without having to slow my RAM way down.
Yan
1. Mainboard software to monitor is generally off compared to bios readings... IE: I am running 1.61v and software says 1.56vQuote:
Originally Posted by skipio
2. First your vcore is rather low for higher overclocks, you'll want to try 1.5, 1.525, 1.55 for +2.5G and temp WILL go up, safe is up to 60'C lower always being better, 65'C is AMD's posted MAX
3. You might want to try 11x multi as I could not do 2.7G with 10x270 but now I'm at 250x11 (2.75G) where the 10x multi would cause strange issues, some people report 10x multi working better for them then 11x also so you'll have to experiment. You will definately NOT want to use stock HSF for high overclocks >2.5G, I recommend XP-90C, I had a XP-120 and it wasn't as efficient cooling as the XP-90C (tested both) plus the XP-90C works in most every MB unlike the XP-120
Hey guys, I was in this thread for about 5 mins before my DFI mobo blew..
I am back now with an ABIT AN8 SLI and ready for action.
A couple quick questions, I am running air cooling for the time being (liquid cooling will be here in 2 weeks)
What voltages are considered safe?
My mobo is reporting temps a little high (classic ABIT) but what is high enough to damage?
So far the highest I can seem to run at dual core stable is 11x240 fsb @ 1.65 v
temps are extremely hot, reporting 70+ degrees so I cant stay here very long.
It will go into windows and play games etc @ 11x250 but not even single core prime stable.
I think I am hitting a heat wall here. Any suggestions?
My goal is 11x250 stable as I am running VX mem so this is the sweetspot.
Could someone please explain how these new mem multipliers work on the E revision cpus? I now have 433 466 and 500 options...how does this affect mem? I tried 433 multi and went into windows...cpuz reported mem speed at liek 120 mhz...is that right?
Oh yeah, I dont ever think I could own a single core system again, performance, response, and overall usablity of the system is through the roof!
65'C is AMD's publicly stated on their website MAX for this chip, not even single core stable? You got serious issues with stability then my friend, going to have to do something, I'm on air and around 55-58'C and that's dual core 1.6v @ 2.75G priming, while I am only about 5-10 hours dual core stable before one of them gets a rounding error this is only due to a bit of heat, if I keep my room AC around 72'F ambient I can dual core for longer but I can single core prime unlimited on either core so for me it is 100% stable until someday I'm running dual cpu 100% for many hours then I might have to back to 2.7GQuote:
Originally Posted by dsumanik
Quote:
Originally Posted by pentium777
Thank you for your reply!
1) Ok, mainboard software and bios don't show the same Vcore in general. But thats not the problem. The problem is that all Vcore settings betweeen 1.3 and 1.4625 (in the bios) are reported as 1.344 in asus probe and speedfan. One would expect a progressive relative increase on both the bios and the monitoring software.
2) According to monitoring software i get 2.4GHz@1.344V and 2.5GHz@1.456V. I think that if the issue above did not exist it is very likeley i could do 2.5GHz at Vcore below 1.45V. So i guess i'll leave the voltage as is and simply find out what it can do max at this voltage as long as it remains below 60 degrees.
3) Did you try both the XP90 and the XP120 with a 92mm fan or with 92 and 120 fans respectively? The reason i mentioned the XP120 is that i could borrow my old XP120 which is on my old system which is now my brothers system and avoid buying a new heatsink.
I suppose you have the A8N SLI premium? Which bios do you use? It seems that untill 1005 and 1006 it seems that the DDR433, DDR466 and DDR500 are not independent from the cpu fsb. To test set in the bios the cpu at 240x10 then go and set the memory to DDR466 or DDR500. Then go back to the cpu settings, are they now greyed out? Save the bios and boot, go in windows and check 'my computer', 'properties', 'general' and it will probably not report 2.4GHz. I have this issue, i have seen others report it on forums and i have seen it also in a review. It seems that Asus may have not yet fully implemented the new DDR multipliers in bios 1005 and 1006. This may be /have been corrected with bios 1007 (still beta) which i haven't tried.Quote:
Originally Posted by dsumanik
Then again it may only be happening with some mobos only, i'm not sure.
Sorry, i am using the the ABIT AN8 fatal1ty SLI not ASUS A8N Premium.
Bios 16
Ooops, somehow mixed up abid and asus. Sorry.Quote:
Originally Posted by dsumanik
Can't comment about your board unfortunately as this is the first AMD and nforce board I've had.Quote:
Originally Posted by skipio
Voltage seems to be ok up to 1.6 some people even run higher up to 1.7 which I would consider dangerous.
I ran a panaflo real nice 120mm fan on that XP-120, the XP-120 certainly is decent but not as good as the XP-90C from my testing but of course that depends on many factors like ambient temp, case airflow etc, if you got the XP-120 free and it will work might as well use that.
Mine X2 4400+ at 3.2GHz......... :D
http://valid.x86-secret.com/show_oc?id=33428
http://www.pctechnology.gr/vbull/vb/...tid=7680&stc=1
Quote:
Originally Posted by pentium777
Thanks for all the advice pentium777!
When i have time to unmount the XP120 from brothers sytem and move it on mine, i will give it a go and see how much better from the stock AMD heatsink it is. Could you please tell me how much did you gain in degrees celsius with the XP90?
As for the voltage, i think for the time beeing i'll stick on the conservative side and try to avoid Vcore voltages over 1.5V and i'll see where that gets me with the XP120.
Thanks.
You have any screenshots of dual core stable for X hours on Prime large and small FFT?Quote:
Originally Posted by ThorLite
Hard to give an exact figure on the gain to the XP-90C it really depends on ambient temp etc, I noticed though when my room temp was a bit higher and I was dual core priming I got about 3-4'C lower temps sometimes up to 5'C if room was colder with my AC on.Quote:
Originally Posted by skipio
Let us know how you do at 1.5v and what temps, also you will notice a huge difference from stock HSF to XP-XXX especially under load.
The XP90c may be better cooling but for quiet cooling IMHO the XP120 is hard to beat :)
And the verdict for #1 air cooler is: Big Typhoon (especially for quiet cooling).
My X2 3800+ sucks....under a single stage phase change it can only dual-prime stable at 2.80ghz @1.64v with IHS OFF!! I'd remount the evaporator so many times that I can now do it with an eye closed while I'm eating hamburger...but the darn thing still won't crack 2.80G. Bought a 4200+ couple days ago and I'm having a better luck this time, I won't bother with any SuperPI 1M at super high clock speed or any suicide screenshot, I'm still dual-priming with my TCCD ram and so far here is what it was running at with Redline XP4000:
http://www.xtremesystems.org/forums/...chmentid=35867
http://www.xtremesystems.org/forums/...chmentid=35868
http://www.xtremesystems.org/forums/...chmentid=35870
http://www.xtremesystems.org/forums/...chmentid=35869