Overclocking is the Dumbest thing you can do!

[imported_Halley]

Originally posted by: bryanW1995
I have 12+ years experience selling cars and I can definitely tell you that overclocking is fun as shit!!

It's fun to overclock. I know I have to spend more money to buy better RAM, better CPU cooler, better case, better PSU, better video card to have some satisfaction in working or play with my computer. My buddy spends more money on his wine but I 've never argued with him.
I doubt if my Q6600 would have a shorter lifespan @ 3.2ghz at 1.2875v. I got it in Feb 2008. Who cares if it fails after 4-5 years? I'm saving to build a 45nm Deneb rig next spring for my video editing hobby. I'm going to overclock it reasonably. Too bad if it can't take the heat.

 

[EliteRetard]

Its so !@#, even if I know what Im doing I cant overclock...from my barton 2500 to my S939 AMD 64 and Intel E6600 not a dang thing will freaking OC. I couldnt figure out how to get the barton even 1 MHz over and that was with a fancy cooler A7N8X-E Deluxe and Corsair PC3500 DDR. The AMD 64 went from 2.6 to 2.8 90% stable with an artic cooler MSI Neo4 and same DDR. And now the Intel with a Gigabyte DS3 and DDR2-800. That one I cant 1MHz either. GOD himself flat out refuses to let me overclock...thats all I can figure. Yes Im a retard, but Im not stupid...I know how everything SHOULD work, it just doesnt.

So thats why overclocking is the dumbest thing you can do.

 

[Drsignguy]

Originally posted by: EliteRetard
Its so !@#, even if I know what Im doing I cant overclock...from my barton 2500 to my S939 AMD 64 and Intel E6600 not a dang thing will freaking OC. I couldnt figure out how to get the barton even 1 MHz over and that was with a fancy cooler A7N8X-E Deluxe and Corsair PC3500 DDR. The AMD 64 went from 2.6 to 2.8 90% stable with an artic cooler MSI Neo4 and same DDR. And now the Intel with a Gigabyte DS3 and DDR2-800. That one I cant 1MHz either. GOD himself flat out refuses to let me overclock...thats all I can figure. Yes Im a retard, but Im not stupid...I know how everything SHOULD work, it just doesnt.

So thats why overclocking is the dumbest thing you can do.

No, driving in the wrong direction on a freeway, playing around a electric sub station, going skiing in shark infested waters, those are dumb, but over clocking? Uh huh?

Over clocking have tolerances, not to mention the throttling capabilities for power savings and cooling. If you stay within the tolerance levels that are given and can gain a substantial over clock, and is stable, all is good and pretty smart too. You just got yourself a price/performance boost. This also depends on the chip too.

 

[PliotronX]

I love it when people get all high and mighty over what others do with hardware they paid for. If anything in this day and age of every little facet of overclocking, tweaking, safe voltages and temperatures documented, overclocking is a lot less risky. It all depends on the user. Even if it was as dangerous as this poster asserts, who cares if they blow their stuff up? Don't we learn a lot of the time through T&E? The only possible downside is the frivolous RMAs chipmakers must process, taking up valuable resources from people with real hardware issues. I would love to be the one to stamp a giant red "VOID" onto the damaged hardware before it's returned to the user.

 

[videogames101]

He fails at life, CPUs are much more robust then most people think.

 

[gammaray]

that guy is like saying, dont put better heads on your engine youre gonna blow it....or dont increase the default factory timing.... oh god i swear

 

[Rubycon]

Of course it's dumb. People do dumb things everyday. Why should this be different?

If someone doesn't like it then fine. They can shut up already. The only ones that I don't like are the ones coming in talking about how they fried xyz hardware but they don't sweat it because they are getting it replaced under warranty. Now those kinds need to sit in the middle of a campfire for a while to think about how it affect prices and return policies for EVERYONE!

 

[OCGuy]

Originally posted by: Rubycon
Of course it's dumb. People do dumb things everyday. Why should this be different?

I think our definitions of "dumb" are slightly different. I think that people who pay $1499 for a 3.2ghz Yorkfield and leave it at 3.2-3.8 are dumb. You can get a 3.2-3.8ghz Yorkie for $270-$500 with a 10 min in BIOS.

 

[GrJohnso]

These are always amusing discussions..... CPU and system specifications are set to the lowest common denominators (crappy cooling, power, etc), to insure maximum lifespan and reliability. Is it reasonable to believe that with the right "improved" supporting components (better cooling, more stable power, etc) that you can move up from that lowest common denominator that had been holding a component back? Well, yes!!!! This is the science of over clocking. Some people may call it "art", but usually there is more logic than magic in basic over clocks... Identify the week links, remove/replace them, and move on...

Anyway, the guy being quoted probably has plenty of experience in his field, but I sure wouldn't want to rely on him for any innovation. Additionally I wouldn't want my degree from the same place he received his. He seems awfully constrained by "what he was taught" rather than "what he should have learned" over the years with all this alleged experience in the field.

 

[Rubycon]

Originally posted by: Ocguy31

I think our definitions of "dumb" are slightly different. I think that people who pay $1499 for a 3.2ghz Yorkfield and leave it at 3.2-3.8 are dumb. You can get a 3.2-3.8ghz Yorkie for $270-$500 with a 10 min in BIOS.

Perhaps that o/c is skewing your sarcasm meter mr. oc guy!

As for "getting" a certain frequency there are no guarantees at all. It's more of a lottery. In the late particularly with 45nm chips most are o/c into instability and the owner does not even know it. Prime, OCCT, folding etc. do NOT certify that a chip is error free, Intel does that. Some go further but that's a safety net in case a cooler is dusty or the a/c stops in the server room.

What's going to wind up happening is part of TCP (trusted computing platforms) software will check to see the CPU is NOT overclocked and if is will not install. Particularly in critical areas where errant data could be costly I can see this requirement logical.

Of course if it's prime stable [sic] and does not crash when playing your favorite game and you're happy that's all that counts, right?

I o/c personally but nowhere to the levels that I used to once I found the flaws in determining what is really stable and what's not.

 

[OCGuy]

Originally posted by: Rubycon

Originally posted by: Ocguy31

I think our definitions of "dumb" are slightly different. I think that people who pay $1499 for a 3.2ghz Yorkfield and leave it at 3.2-3.8 are dumb. You can get a 3.2-3.8ghz Yorkie for $270-$500 with a 10 min in BIOS.

Perhaps that o/c is skewing your sarcasm meter mr. oc guy!

As for "getting" a certain frequency there are no guarantees at all. It's more of a lottery. In the late particularly with 45nm chips most are o/c into instability and the owner does not even know it. Prime, OCCT, folding etc. do NOT certify that a chip is error free, Intel does that. Some go further but that's a safety net in case a cooler is dusty or the a/c stops in the server room.

What's going to wind up happening is part of TCP (trusted computing platforms) software will check to see the CPU is NOT overclocked and if is will not install. Particularly in critical areas where errant data could be costly I can see this requirement logical.

Of course if it's prime stable [sic] and does not crash when playing your favorite game and you're happy that's all that counts, right?

I o/c personally but nowhere to the levels that I used to once I found the flaws in determining what is really stable and what's not.

Well if it can pass 8 hours small FFTs, that is certainly my "stable." If you have to run your CPU @ 100% load for 48 hours to find an error, all you really did was waste power. Unless you are going to be able to mimic that kind of constant stress in the applications you run.

But you are right, I am talking about a home computer that is used for gaming and light media. If I had a critical business computer that needed to be 100% rock solid at all times, I wouldnt even mess with it. I dont think most people would....(I hope).

As far as a fan stopping, even without OCing you will hit the max thermal level and throttle down/shut off.

Most people I see on here fall into 1 of 3 categories, and ive passed through all three myself.

1. Theyve decided they want to try this great "overclocking" thing that they have read about in all those cool Newegg reviews for thier new E8400. However, they are still scared they are going to "fry" something, so they arent quite ready. They are pretty much browsing threads like this, trying to see if there is any real danger.

2. They figured out the whole FSB X Multi thing, and can add more Vcore to make it boot. They are still novice, and more than likely arent as stable as they could be, or are running incorrect voltages (too high or low). They maybe even use "Auto" settings for most. This is the group that can give OCing a bad name if they cant figure out how to clear CMOS after some bad settings, or use high Vcore (generally excepted as 1.4v max for 45nm 24/7)

3. Theyve got OCing down, and are knowledgeable and helpful to the others. The assumption that most of them are "ocing into instability" is probably very far off from reality. There are a ton of these people on this board.

But yea I guess I missed the sarcasm in your first post....its early!

 

[nerp]

If OCing was so inherently evil and stupid, don't you think we'd all still be setting jumpers and manually modding our gear? When there's a BIOS page to do it, I consider it widely accepted and frankly condoned by the manufacturers.

Intel and AMD could stop the OCing community in its tracks if it wanted. Especially with NVIDIA's chipset business shrinking by the day (go AMD!)

 

[Rubycon]

The problem with these popular tests is they do NOT exercise ALL cpu features. A system that seemed perfectly stable was producing errant output when encoding videos using SSE4 instructions! Whoops!

I'm not a programmer but until there is something that can do this the results from Prime, etc. just cannot be trusted. It's not how close either.

I had a chip doing 4.4GHz and erring out in a few min. Dropped to 4.2 ran OCCT all night. Encoding engines crashed once the queue started!

Dropped to 4.0GHz, same thing although the batch job did get a lot further. 3.8GHz worked however the output was checked and artifacts were found!

Finally at 3.6GHz it was completely error free. Remember this system worked perfectly with everything else at 4.2GHz! QX9650. Two QX9650's as well as QX9770's tried - the 3.2GHz 9770's were actually a little WORSE (well one of them was producing artifacts at a paltry 3.3GHz with default core).

This could be why Penryns never got above 3.2GHz.

 

[magreen]

Originally posted by: EliteRetard
Its so !@#, even if I know what Im doing I cant overclock...from my barton 2500 to my S939 AMD 64 and Intel E6600 not a dang thing will freaking OC. I couldnt figure out how to get the barton even 1 MHz over and that was with a fancy cooler A7N8X-E Deluxe and Corsair PC3500 DDR. The AMD 64 went from 2.6 to 2.8 90% stable with an artic cooler MSI Neo4 and same DDR. And now the Intel with a Gigabyte DS3 and DDR2-800. That one I cant 1MHz either. GOD himself flat out refuses to let me overclock...thats all I can figure. Yes Im a retard, but Im not stupid...I know how everything SHOULD work, it just doesnt.

So thats why overclocking is the dumbest thing you can do.

I LOL'ed. Maybe God really doesn't want you to OC.

For the rest of us, however, OCing is great.

 

[OCGuy]

Originally posted by: Rubycon
The problem with these popular tests is they do NOT exercise ALL cpu features. A system that seemed perfectly stable was producing errant output when encoding videos using SSE4 instructions! Whoops!

I'm not a programmer but until there is something that can do this the results from Prime, etc. just cannot be trusted. It's not how close either.

I had a chip doing 4.4GHz and erring out in a few min. Dropped to 4.2 ran OCCT all night. Encoding engines crashed once the queue started!

Dropped to 4.0GHz, same thing although the batch job did get a lot further. 3.8GHz worked however the output was checked and artifacts were found!

Finally at 3.6GHz it was completely error free. Remember this system worked perfectly with everything else at 4.2GHz! QX9650. Two QX9650's as well as QX9770's tried - the 3.2GHz 9770's were actually a little WORSE (well one of them was producing artifacts at a paltry 3.3GHz with default core).

This could be why Penryns never got above 3.2GHz.

Yes but that is your experience, with your chip, using the voltages that you tried.

The chips are binned at what they are for marketing. If a E8400 can run 4ghz stable for alot of people, I have a hard time believeing that they couldnt bin them @ 3.5ghz stock if they really wanted to. I also have a hard time believing a QX9770 couldnt do 3.3ghz without "artifacting." I would really have to see all the voltage settings you were using.

As far as the bolded part, the chip in my sig is a Penryn over 3.2ghz.

 

[VirtualLarry]

Originally posted by: Rubycon
I o/c personally but nowhere to the levels that I used to once I found the flaws in determining what is really stable and what's not.

You know, I wonder about that.

I have two E2140s @ 3.2Ghz, 1.425v (BIOS), 1.36v (CPU-Z), without the AC they ran at up to 9C from TJmax with CoolerMaster HyperTX2 coolers. Gigabyte P35-DS3R v1.0 mobo.

I also have an E4400 L2 stepping on an ASRock Conroe865PE, running at 2.8Ghz, default voltage (since there is no vcore adjust on the mobo at all).

The ASRock, I don't use, I just leave it running nonstop at a relative's house. I leave it for months at a time and it doesn't crash, although the ethernet switch I have it connected to seems possibily flaky, as I sometimes get messages from SeventeenorBust about no connectivity sometimes. I had to power-cycle the ethernet switch once to fix that.

But the Gigabyte boards I use, downloading, web browsing, MagicJack, etc. I also run SeventeenorBust in the background. But on those rigs, I don't have stable uptimes greater than a month. Either one instance of SeventeenorBust crashes, or I get a bluescreen. So far, there have been three major crashes, either SB quits or I get a bluescreen. One time the bluescreen might have been due to a thunderstorm, even though I have them connected to APC ES-550 UPSes.

So am I stable hardware-wise, if I can run for a month, or am I unstable, since I can't seem to run longer than a month.

I pass 24hr of Prime95 25.x, 24hr of Memtest86+, etc. I haven't run F@H on these rigs.

 

[VirtualLarry]

Originally posted by: Rubycon
The problem with these popular tests is they do NOT exercise ALL cpu features. A system that seemed perfectly stable was producing errant output when encoding videos using SSE4 instructions! Whoops!

This is why a system can be Prime95 stable, but (in the case of P4s, not integer stable), or not F@H stable. Prime95 exercises mainly the FPU. F@H exercises mainly the SSE2 units. The weak spot in Athlon and A64 systems seems to be the FPU (possibly SSE), the weak spot in P4s is the integer units, with the double-clocked ALUs (I think that's true). In Core2, it seems the weak spot is the SSE units.

There still isn't a "perfect" stress-test, one that fully exercises ALL of the pathways in the CPU. For another example. consider if the branch-prediction units were failing. The CPU wouldn't necessarily error, but branches would simply take longer than they should. This would be a nearly undetectable error condition.

 

[Spicedaddy]

Been doing it since the first computer I built and never damaged anything.

I don't OC to the bleeding edge though, I try to stay at or very close to stock voltage, but I'd feel kinda stupid running at stock speed and knowing there's that free extra 700-800MHz...

 

[Gillbot]

Originally posted by: Zap

Person1:

After working in electronics for 15 years and a bachelors degree in the field

Person1:

I have a Bachelors degree in EEE ive worked in the field for over 10 years exp if you must know.

So, is it 10 years or 15 years experience?

As SOON as someone starts spouting how many years of experience he/she has and what kind of qualifications they have... immediately tells me they know Jack and Shit.

HEY!!!! I've known Mr Jack Schitt for over 15 years now!!!

 

[karkas]

His post could have been just as dumb if he said

"Anyone who drives their car faster than the posted speed limit is stupid. All they are doing is putting additional wear and tear on their engine & car"

 

[imported_Woody]

Originally posted by: Fox5
Overclocking doesn't even really reduce the life time of the cpu as long as you don't increase voltage. Generally, it's safe to do whatever clock speeds within whatever voltage ranges the official cpus do, meaning you get more for cheaper at no penalty. The dangerous overclocks are the ones that add like half a volt and push it to 4.2ghz.

Exactly. It's the increase in heat that is the primary cause of increased wear and tear on the electronic components. Also the cycling of heat or rapid changes in heat cause wear and tear that can lead to eventual failure.

The fact is, if you put a massive heat sink on your CPU, increase the voltage only slightly if at all, and overclock it to the limits of it's stability, you may actually get more life out of it than if you just ran it at stock speeds on the stock heat sink because your temps and temp variations will be lower.

It all comes down to knowing what you're doing.

 

[shabby]

Originally posted by: Amart
Person1:

After working in electronics for 15 years and a bachelors degree in the field OC is the dumbest thing you can do end of story. You fail at life and science

LLLLLLL.................................OOOOO..............LLLLLLL....................
LLLLLLL.............................OOO.....OOO...........LLLLLLL....................
LLLLLLL..........................OOO..........OOO.........LLLLLLL....................
LLLLLLL........................OOO..............OOO.......LLLLLLL....................
LLLLLLL........................OOO..............OOO.......LLLLLLL....................
LLLLLLL........................OOO..............OOO.......LLLLLLL....................
LLLLLLL........................OOO..............OOO.......LLLLLLL....................
LLLLLLL........................OOO..............OOO.......LLLLLLL....................
LLLLLLL........................OOO..............OOO.......LLLLLLL....................
LLLLLLL........................OOO..............OOO.......LLLLLLL....................
LLLLLLL........................OOO..............OOO.......LLLLLLL....................
LLLLLLL........................OOO..............OOO.......LLLLLLL....................
LLLLLLL........................OOO..............OOO.......LLLLLLL....................
LLLLLLLLLLLLLLLLLLLLL....OOO...........OOO.........LLLLLLLLLLLLLLLLLLLLL
LLLLLLLLLLLLLLLLLLLLL......OOO.......OOO...........LLLLLLLLLLLLLLLLLLLLL
LLLLLLLLLLLLLLLLLLLLL.........OOOOOO...............LLLLLLLLLLLLLLLLLLLLL

How can someone with a bachelors degree be so stupid?

 

[dmw16]

Sure, you may shorten the life span of your CPU, but let's be clear...these things are designed to last many years beyond their useful life. Let's say they are designed to last for 15 years, and you overclock and shorten it to 7.5 years. That is still about 3 years longer than it has any real value from a performance standpoint.

 

[gammaray]

Originally posted by: dmw16
Sure, you may shorten the life span of your CPU, but let's be clear...these things are designed to last many years beyond their useful life. Let's say they are designed to last for 15 years, and you overclock and shorten it to 7.5 years. That is still about 3 years longer than it has any real value from a performance standpoint.

+1

 

[knght990]

Do we factor in that some CPU lines are 'rejects' from faster lines? Underclocked and sold as a lesser chip, you find them with sections of the die 'locked' but capable of greater cycles with no heat issues and persumably no greater wear issue than the original designs.

I've found quite a few engineers who lack the ability to think outside the constraints of their training. The rules of science are limitation because thoes who came before us were infalable. Failure in their education or outlook or somthing.