My 4400 can't get past 2.65 or so...why?

[LoneWolf121188]

I've got an XP-120 and AS5 and this thing (well, the second core specifically - the 1st is great) still isn't Prime or SuperPi stable above 2.65 :confused: . I'm seeing people get over 2.8 on the stock cooling, and here I am...I've been burning in over the past few days, but my temps are fairly high, so I don't really want to raise the voltage much: high 40s-low 50s, depending on which mobo sensor (if either :rolleyes: ) is right. A8N-SLI dlx for the mobo (BIOS is 1012.001), Patriot PC3200 RAM...Any help? 2.7 is my original goal, so if I can even hit that, I'll be happy. I've got a thread over at OC Forums with the same user name ( link (http://www.ocforums.com/showthread.php?t=400811) ).

Again, any help would be great.

[turtle]

The people here that are getting 2.7-2.9 most often have the IHS removed. It seems the "one core overclocking crappy" is a trend because the ihs is making bad contact on one core in almost all, if not all x2's. Removing it should 'cause load temps to drop by ~10c and should give you a healthy boost in clocks. If you're getting 2.65 now, you may get as high as 2.8 or 2.9 with the IHS removed. That seems to be the avg. anyway. Remember though, if you remove it, put felt bumpers or something on the chip to prevent from cracking the core (it'll end up looking a lot like an xp cpu). Also make sure you have good contact, as many HSF's are built prespaced for the chip with IHS on it, so a washer may be needed to insure good contact.

[spurious]

the 4400 and 4800 X2 have over 200 million transisters. it is a hi-tech achievement that is hard to comprehend.

why is AMD is letting a simple mechanical process of installing some thermal compound and adding a heat spreader be done so poorly. even running at stock speeds the failure rate will be above normal. AMD does not need high failure rates or reports that the chips run ultra hot.

i would like to know what percentage of X2 chips have a botched heat spreader. i have also seen reports of 3700+ and 4000+ chips running hot at stock speeds. do they have the same problem? how about venice?

i want to get a 4400 X2, what are my odds i get one that has the heat spreader on correctly? You would think a $600 CPU should be assembled correctly.

[turtle]

Who knows? It seems to be almost a fact though it's "botched" in most cases on the X2 as all the before/after temps people have reported that i've have shown a MASSIVE decrease in load temps after the IHS is off and the core is AS5'ed well. This lower load temp surely is the cause of higher stable clocks, and the IHS contact/stock TIM surely is to blame for it for the higher stock temps/lower stock overclock. I'm sure venice/sd may benefit from the removed ihs and as5's core (as opposed to stock tim) but I think the problem is mainly bad contact on one of the cores in the x2. As the others only have one core, it's harder to judge without a group of people doing fullblown testing on those cpus. Odds are they will do better though. if it's being held back as much as the x2's seem to be, who knows, but it's half as likely, as it seems before IHS is removed almost everyone has one core that does "better" on thier X2's. They seem to be more even after the operation. This is harder to judge on single-cores, although I guess there is a chance the single cores could be botched by one spot on that core not making contact and causing it to heat up faster and therefore it to top out lower than it could. That being said, 40c or so isn't ultra super hot (venice/sd/x2 are all complimented for running cool) and is far from overheating. It just seems they will run at a better clock when in the low 30's load, which is possible with normal cooling with the ihs removed.

Maybe it's a conspiracy to make this core last longer as they can release not-crappily ihs/tim'ed cpus in the future that stock clock at 2.8ghz that run even cooler than the current general cpus once the FX-57 has lost it's charm. ;)

JK.

If a group of people wanted to test SD/V's before and after removing the ihs for temps and overclocks, I know i'd surely appreciate it. Perhaps these guys are on to something...although you would imagine if it made a huge differance it would have been widely known about by now. None-the-less, someone start a thread and let's see some results. :D

[spurious]

better thermal interface material would cost 2 or 3 cents per CPU at most?

assembly is done in costa rica or malaysia? maybe they should do the assembly in dresden. after all the effort it takes to make a functioning CPU, i think they would put in a better effort to assemble it.

[jinu117]

It isn't like more than 1% of population care if IHS is well applied or not as long as it is functiong... Now if it was assembled in drisden vs costa rica or malaysia, don't you think it will cost a lot more then few cents to do IHS per chip? Cater to less than 1% of population and lose money? Somehow... doubtful.

[spurious]

if you miss covering all the cpu die on a majority of CPU's you build, its not just overclocked systems that will fail. reliability even in stock systems will be poor.

after looking at the pictures of heat spreaders removed, it appears the thermal interface material is being applied by a 3 year old. did they not adjust the amount of material based on the size of the die? is it applied by machine or do people do it?

am i the only one that is concerned about putting the equevalent of a ferrari engine into a ford pinto?

[Arkangyl]

sorry to be a tad off topic here but;

my 3000+ (winny) seems to be running hot as well, any chance that it could also be a bad IHS (first I've heard of IHS issues).

[jinu117]

if you miss covering all the cpu die on a majority of CPU's you build, its not just overclocked systems that will fail. reliability even in stock systems will be poor.

after looking at the pictures of heat spreaders removed, it appears the thermal interface material is being applied by a 3 year old. did they not adjust the amount of material based on the size of the die? is it applied by machine or do people do it?

am i the only one that is concerned about putting the equevalent of a ferrari engine into a ford pinto?

Give me a mobo with what WE consider as good application of NB heatsink. Give me one good memory with what WE consider as good application of thermal pad and RAM sink. Now, tell me how many people return for RMA from general build system (those who doesn't overclock) for it not having stability -_- (and I do mean H/W problem not their user errors)
Not that I wouldn't love to see ceramique on every chipset cooler, maybe better designed heatspreader or none in high performing memory, and kick ass spreader construction and application that can handle various range of temperautre across temperautre. But reaility is... we are not norm.

Back to topic: Grab even higher flow fan, lower the temp a little more. See if it helps. If it does, it might be worth taking HS off as you know what you will be getting out of doing so most likely... I have taken mine out finally on my single phase and gained another 100mhz dual prime 95.... finally doing it 3ghz... :P

[spurious]

if thermal compound and heat sink / heat spreader application is so bad, a manufacturer that says there are there for the overclockers should sell you the memory / cpu / motherboard without the heatsinks applied and let you do it yourself.

is it that difficult to get right. can someone point me to an article that explains WHY they do such a poor job?

[highoctane]

Search this forum, its plainly clear there are "definitely" issues with IHS contact on the X2 cpus, I could get into the details here buts it's already been discussed in other threads.

Why should it bother AMD, as long as it runs at its rated speed thats all they require.

[Absolute_0]

You guys really make me want to remove my IHS. I know it's got bad contact. With 1.6 volts it loads at 52C on XP90 and with 1.66 volts on my custom water it loads at 44C. Ambient temps are not particularly high, ~69F.

I'm curious, if the AXPs did not have the IHS, why bother putting them on the AMD64 at all...

[Arkangyl]

IHS's save cores. With the AXPs many cores were crushed because of the lack of IHS. Fewer crushed cores = fewer RMA's = lower prices

[pcdoc1]

Search this forum, its plainly clear there are "definitely" issues with IHS contact on the X2 cpus, I could get into the details here buts it's already been discussed in other threads.

Why should it bother AMD, as long as it runs at its rated speed thats all they require.There's way too much being made of this........ Because of all the post here and other forums I removed the IHS on my 4800 to test...... Upon inspection it was very clear that the manufactureing process was not faulty and there was extremely good contact between the IHS and the core...... The net result was a 2 degree temp reduction at both idle and load. Obviously it has no impact on your OC.... I guess you could argue a lack of consistency in the process but I don't think so...... I'd suggest being extremely cautious before putting your $1000.00 cpu under the knife......

Regards-

[jinu117]

You guys really make me want to remove my IHS. I know it's got bad contact. With 1.6 volts it loads at 52C on XP90 and with 1.66 volts on my custom water it loads at 44C. Ambient temps are not particularly high, ~69F.

I'm curious, if the AXPs did not have the IHS, why bother putting them on the AMD64 at all...

To be honest, your temp looks xtremely good if it is dual core you are talking about. Don't even bother :) I think I was @ 1.6v 52c with H2O... You lucky man :P

[spurious]

hopefully everyone gets their stock on 8/5, because i will still buy a 4400 X2.

just hope i get lucky and it runs at decent temps at stock speeds.

[highoctane]

There's way too much being made of this........ Because of all the post here and other forums I removed the IHS on my 4800 to test...... Upon inspection it was very clear that the manufactureing process was not faulty and there was extremely good contact between the IHS and the core...... The net result was a 2 degree temp reduction at both idle and load. Obviously it has no impact on your OC.... I guess you could argue a lack of consistency in the process but I don't think so...... I'd suggest being extremely cautious before putting your $1000.00 cpu under the knife......

Regards-

As with anything ymmv and theres a reason a 4800 is going to be binned as a 4800.

My 4400 on the other hand would absolutely refuse to run stable over 2.5ghz and I couldn't get load temps below 50C going from an xp120, to a swiftech h20 120 wc setup, & finally a custom wc setup with 1/2" plumbing, in the end the max load temps would still exceed 50C between all three different cooling setups.

In the end removing my IHS netted an instantly stable 2.7ghz speed and a 15c drop in load temps.

Just because your core didn't see much a difference there have been more than a few of us that have been able to gain substaintialy from what would otherwise be considered a bum core.

Anyways the results speak for themselves, if you have one core that is quite a bit weaker than the other odds are its going to be from your IHS.

Trust me the last thing on my mind was the chance of ruining a $600 cpu but after going through over $500 worth of cooling equipment and still not able to achieve noticably lower load temps I removed the IHS and it was like magic.

[Absolute_0]

To be honest, your temp looks xtremely good if it is dual core you are talking about. Don't even bother :) I think I was @ 1.6v 52c with H2O... You lucky man :P

my 3000+ venice actually

[joe2004]

#1. Heat
#2. Heat
#3. Heat

XP-120 cannot cool anything effectively above 130-140W.

[smids]

Not all chips are made equal. Most people will hit 2.6ghz with relative ease (1.45v-1.5v) but beyond that it really is pot luck. My duff core is really holding me back. The good core, Core 0 will run @ 2.6ghz prime stable on 1.39v whereas the bad core will only do the same @ 1.48v. Bit of a bummer really. I haven't got the grapefruits to remove my IHS. £380 chip going down the pan as I sneeze with a razor blade in my hand doesn't appeal to me. My watercooling does a decent enough job anyway.

[Zer0X]

I've got an XP-120 and AS5 and this thing (well, the second core specifically - the 1st is great) still isn't Prime or SuperPi stable above 2.65 :confused: . I'm seeing people get over 2.8 on the stock cooling, and here I am...I've been burning in over the past few days, but my temps are fairly high, so I don't really want to raise the voltage much: high 40s-low 50s, depending on which mobo sensor (if either :rolleyes: ) is right. A8N-SLI dlx for the mobo (BIOS is 1012.001), Patriot PC3200 RAM...Any help? 2.7 is my original goal, so if I can even hit that, I'll be happy. I've got a thread over at OC Forums with the same user name ( link (http://www.ocforums.com/showthread.php?t=400811) ).

Again, any help would be great.

I popped the IHS off on my 4400 just to improve temps (about 9~11C). I have a thremaltake Big Typhoon 120mm on it and it works great. Once I get my TVT card to work I can start OCing.

[Ref]

Not all chips are made equal. Most people will hit 2.6ghz with relative ease (1.45v-1.5v) but beyond that it really is pot luck. My duff core is really holding me back. The good core, Core 0 will run @ 2.6ghz prime stable on 1.39v whereas the bad core will only do the same @ 1.48v. Bit of a bummer really. I haven't got the grapefruits to remove my IHS. £380 chip going down the pan as I sneeze with a razor blade in my hand doesn't appeal to me. My watercooling does a decent enough job anyway.
It is very tempting to remove the IHS and see big gains in OC (especially with X2s) ... you know you WILL remove it once ... so why wait? if you do it properly there is virtually no risk at all... :toast:

[smids]

if you do it properly there is virtually no risk at all... :toast:
On a scale of 1 to incompetant, I'd say I ranked in at about incompetant. I was removing the RAM from my board two weeks ago and almost broke the stick, the socket and my graphics card - it was well jammed in so I applied extra force, only for the thing to shoot out, me hitting all wires under the sun and cutting my palm on the corner of my already strained graphics card (w/c tubing and heavy block). I might remove it towards the end of it's life - or one which belongs to a friend.

[burningrave101]

I personally dont think its worth it to remove the IHS. You might get an extra 200Mhz or so but who cares? You can't tell the difference when actually using the PC. You'll gain a few points in benchmarks but you wont have any 3 year warranty on the retail version any more and you stand a good chance of cracking the core if your using a heatsink like the XP-90 or XP-120. I wouldn't waste my time taking off the IHS unless i had extreme watercooling or phase change and was just trying to push the chip to its limits whether i killed it or not.

[Absolute_0]

I personally dont think its worth it to remove the IHS. You might get an extra 200Mhz or so but who cares? You can't tell the difference when actually using the PC. You'll gain a few points in benchmarks but you wont have any 3 year warranty on the retail version any more and you stand a good chance of cracking the core if your using a heatsink like the XP-90 or XP-120. I wouldn't waste my time taking off the IHS unless i had extreme watercooling or phase change and was just trying to push the chip to its limits whether i killed it or not.

That describes me. I mean, 200 Mhz doesn't make a noticeable difference in windows, and neither does being at 2.95 instead of stock. No windows difference. But what's the fun in running stock. We go the the limits for the thrill of the OC :toast:

[highoctane]

Its all in good fun, some people build model rockets to launch up for no good reason, some model airplanes, some race cars, some build bird houses, and some of us just have fun with computers.

Everything doesn't have to make "sense", if it did the purpose of this site would be largely dimished.

Its not like the people removing the IHS are calling the guys who don't have the balls to pull off the IHS pussy's, to each his own.

[pcdoc1]

Its all in good fun, some people build model rockets to launch up for no good reason, some model airplanes, some race cars, some build bird houses, and some of us just have fun with computers.

Everything doesn't have to make "sense", if it did the purpose of this site would be largely dimished.

Its not like the people removing the IHS are calling the guys who don't have the balls to pull off the IHS pussy's, to each his own.It's not "all in good fun" when someone trashes a $1000.00 cpu for little to no gain...... There's been a few people (and only a few), that have had a gain of any significance by removing IHS, yet it continues to be hot topic! :) So carry on.........

[highoctane]

It's not "all in good fun" when someone trashes a $1000.00 cpu for little to no gain...... There's been a few people (and only a few), that have had a gain of any significance by removing IHS, yet it continues to be hot topic! :) So carry on.........

Well that someone would have to be pretty reckless plain and simple, and whats this "few" do you have an actual number count or is this just your educated guess, was there a poll taken?

Oh well theres always some haters/pesimists, I'm just glad I didn't spend $1000 for cpu.

[Absolute_0]

It's not "all in good fun" when someone trashes a $1000.00 cpu for little to no gain...... There's been a few people (and only a few), that have had a gain of any significance by removing IHS, yet it continues to be hot topic! :) So carry on.........

I've read the better part of the IHS removal thread and many people make significant gains, 2-11 C.

[pcdoc1]

I've read the better part of the IHS removal thread and many people make significant gains, 2-11 C.Equate that 2-11 to your OC and let me know how it turns out...... I promise you it won't result in the 150-200MHz increases being claimed......

[Absolute_0]

Equate that 2-11 to your OC and let me know how it turns out...... I promise you it won't result in the 150-200MHz increases being claimed......

i dont remember people claiming 200 Mhz, although I am sure that i have read 2 claims of 100 Mhz. And i usually assume people are telling the truth. All i want is 40 Mhz myself.

At this point, all i need is about 4C off my load temps and i can bump voltage up a notch and hopefully stabilize 3 Ghz.

[giantjoebot]

I know that this is an old thread, but I'm having a problem with my Opty 165.

I ran a thread on it on the AMD forum (http://forums.amd.com/index.php?showtopic=75626&pid=1363965519&st=0&#entry1363965519)

I think that the heat spreader is the problem after reading this thread. It either that ofr my Zalman 9500 isn't working. I'm going to give removing it a try in a couple of months when I'll have the cash to replace it if it breaks. Does anyone know any good How To's on removing the heat spreader?

[arisythila]

Mine is doing pretty well. 2.8 ghz, 1.3v ect ect. AM2 tho.

~Mike

[giantjoebot]

Not only is your processor not giving you any problems it the latest one out.

Just kidding. So how is the AM2 anyways?

Yeah I haven't heard about very many people that had these issues. I've found a few threads where people had the same thing happen to them, but they never said how it ended up. I think they just gave up, and stopped posting. Well I'm not that type of person unfortunately. I obsess over things like this, and won't stop until it kills me to fix it. I went through the trouble of getting a new mobo, ram, and heatsink to try and solve this problem. I reinstalled my heatsink 4 times already, and the first time I did it was with a different type. I've followed the arctic silver instruction, and there really seems to be no other explanation. I'm just unlucky I guess. But maybe with a little more hard work I can fix my luck, and turn it around. I bet if I pull off the heat spreader it will take off 11C if not more. Then agian with my luck I'll either break it, or it will end up even hotter than before.