Did you use ECC RAM? How do you know it was the CPU not the RAM?
Both (I had systems that used ECC and others which did not). I was also able to duplicate the same portfolio of symptoms on both my AMD and Intel based rigs.
I knew it was the CPU because the "problem" went away when I reduced the overclock.
The issue here, as is the case with all silent data-corruption, is detection. Just because I was able to reduce the overclock and make a known corruption symptom go away does not mean that I made all the unknown corruption issues go away. False negatives are a bitch.
You only know to look for corruption where you know to look for it. When it happens somewhere that you aren't looking then it goes undetected.
So you start bolstering your detection methods, that takes up more time and resources (the two things you were attempting to gain by overclocking in the first place). And it still doesn't ensure you are not at risk, at best you are merely reducing the known risk but you are still in the dark on the unknown risks.
But it is like this - this is a competitive world and as much as I am altruistic in sharing my educated opinions on things with my fellow enthusiasts I also have no problems letting my competing engineers and academicians undermine themselves by overclocking their systems and trashing their results (whether or not they catch onto it when it happens under their own noses is their problem).
OC'ing your rig when doing anything that matters on it IS a classic
Tortoise and the Hare situation.
And the moral of the tortoise and the hare story is that you need idiots (Hare's) who foolishly run around (OC'ing) thinking they are gaining some advantage over everyone else (ego mixes with arrogance, a toxic cocktail) in order for the tortoise to eventually win out in the situation.
In other words, without the hare the turtle is just one slow-ass turtle. It takes the foolishness of the hare to make the turtle into a winner.
And for that reason I have had many successes in my life, once I stopped being the Hare to other people's tortoises and became a tortoise myself. That was a key turning point in my career, and since making that transition I have watched it happen like a fly on the wall in many other people's careers as if it is some sort of universal learning curve (and not everyone makes it up the climb).
Slow down and get shit right the first time and people will value you for your consistency more than they will for your random one-hit homeruns that come once a season.
So to all you Hare's out there I say "keep it up! don't stop, sure you know what you are doing, you are more clever than everyone else, other people don't do it because they aren't as clever and crafty as you, you are special, please keep OC'ing your hardware and making your results all the more likely to be crap! the rest of us are counting on you continuing to be you
"
Now don't confuse my position here as being one of anti-OC'ing in all regards. I do OC, as a hobby on my home computers for which I am doing stuff that doesn't really matter when it comes to my livelihood (or someone elses livelihood).
There is amateur vs. professional OC'ing and there is responsible vs. irresponsible OC'ing.
Amateur OC'ing can ONLY be responsible OC'ing if the amateur restricts the application of OC'ing to those situations in which the rig won't be used for anything critical. Don't OC your computer if you are a freelance consultant or running your own business, etc. You will race ahead like the hare only to crash and miss a critical deadline (or worse, turn in junk results and lose all credibility for the future).
It is VERY irresponsible for an amateur to OC a rig for a friend or family member, or to sell a pre-OC'ed rig, because the amateur OC'er has zero control over the types of computations and applications the OC'ed computer is tasked with.
Someone might be attempting to do their home finances on that OC'ed rig, or it silently corrupts all their digital photos of their honeymoon (happened to me) or baby photos of their kids. Had they been made aware of the real risks that come with OC'ing then they might not have been so willing to let their nerd friend OC their computer. Deciding for someone that the risk is acceptable is the very definition if irresponsible.
A professional can do responsible OC'ing but many people confuse professionals with being amateurs who have a lot of experience OC'ing in a DIY manner. The difference between amateur and professional is not a matter of experience (time spent OC'ing), rather it has to do with training and awareness. Knowing what to be looking for, how to quantify risk, mitigate it, detect it, etc.
Anand Lal Shimpi said:
Glen took me on a tour of Centaur's simulation lab. To say it was a different experience would be an understatement. While some machines were racked, there were a lot of desktop motherboards running Core i5s and Core i7s running out of cases:
The systems that were in cases were water cooled Core i7s, overclocked to 5GHz. There are two folks at Centaur who build each and every one of these machines, and overclock them. You and I know that overclocking both Nehalem and Sandy Bridge results in much better performance for the same dollar amount, but this is the first time I've seen overclocking used to speed up the simulation of microprocessors.
Source
When you design and validate CPU's for a living you know exactly what to do (and what not to do) when OC'ing your hardware. The folks at Centaur know what they are doing, they are professionals. But that is because of the uniqueness of their profession (designing/validating x86 processors for a living).
You don't get to be an aeronautical engineer by working at Boeing pushing a broom for 20yrs (loads of experience in the aerospace industry)...the guy who just graduated college with a 4yr degree knows more about designing airplanes on his first day on the job than the guy who has pushed a broom for 20yrs.
Same with the difference between professional and amateur OC'ing. I am not a professional OC'er, I merely know enough to know that I am an amateur without question. It took me a while to gather my wits about me and realize this. As such you won't see me compiling or running simulations on Oc'ed hardware anymore.
But I'm counting on young whippersnappers being foolish and shooting themselves in their own feet, its what makes the rest of us tortoises invaluable (highly compensated
) to our employers :sneaky:
Facebook
Twitter