Observations About Overclocking

Pages: 1 2

In the Beginning…

It was almost 1 year ago that I first became exposed to overclocking. We had just put our website up, and I was making my acquaintance with the various Usenet groups and Web sites that pertained to our business. I was reading one post and was shocked at the questions that were being asked about running an Intel Pentium 100 CPU at 133MHz. My first thoughts were of smoke and flames, damaged equipment and a distraught user. Apparently, many others thought the same way because the next few days revealed literally dozens of replies that severely chastised this individual for even *thinking* of such a heinous act. After that, I noticed that some people would meekly (or even boldly) post their questions just hoping that out of the myriad of replies, one or two would reveal some secret information that would let them successfully overclock.

Some of the replies would actually be helpful, but also would be cautious. I can recall seeing more than one post from somebody who would admit to having destroyed at least one processor while trying to figure out the right combination of jumper settings and components that would let him finally get the ‘maximum’ performance from his system, perhaps at the expense of it’s longevity. At that time, the expectation was that you might get only a year or so out of the processor, but the idea was that the savings that were gained from being able to purchase a lower speed processor would offset the reduced life, and would help to pay for the next chip somewhere down the line.

As I became more familiar with the concepts of overclocking, electro-migration, thermal considerations, faster bus speeds, etc., I began to warm to the idea of overclocking, and even did some cautious playing around during some free moments in our tech lab. While never a big fan of benchmarks, I did run a few just to see what my gains were when I was able to successfully get a stable system up and running (which seemed surprisingly easy). While I understood the issues involved with overclocking, and could not officially condone it, I also began to feel that there was little danger in overclocking if the individual was prudent and cautious.

It was not much later when I became aware of Tom’s Hardware Guide and much of the wonderful information it contained. Like so many others, I would visit frequently hoping to see what new information he might provide about bus speeds, memory timings, motherboard capabilities, etc. About that time, we decided to answer his call for sponsors so we could support his efforts, and perhaps gain some benefit for ourselves in additional customers

Overclocking begins to Heat Up

Well, as it turned out, we were very surprised at the number of people who ordered from us just because we sponsored Tom’s site. We congratulated ourselves on our ‘genius’, and watched our sales climb. We set our policies so that we would be ‘overclocking friendly’, and offer a little slack when providing warranty service for people who were admitting to overclocking.

While this seemed like a good idea at first, we soon began to realize that our RMA numbers were beginning to climb rather rapidly. At first we blamed it on some of the components (non-ASUS cache modules, slow memory, boards released too early, etc.) and frantically searched for suppliers of only the best hardware, in some cases completely abandoning suppliers that were less expensive, but couldn’t offer the high quality we were seeking.

It didn’t take long to realize that this approach was not working either. A little bit of analysis revealed that the problem was not with the components, but with people who didn’t understand the limitations of the hardware they were trying to push so hard. Many had read the information on Tom’s Hardware Guide, and the suggestions that were now beginning to become the ‘norm’ on Usenet, but were ignoring the many caveats and cautions that were also being given by Tom and others. In essence, these individuals were jumping into overclocking without understanding the built in limits of the components, and in some cases did not even know that the settings that they were using were not supported by the manufacturers.

The solution to this was to put into place some stricter policies on warranty service, and to provide some information about the issues regarding overclocking. In addition, some motherboard manufacturers were also becoming more ‘overclocking friendly’, and were producing boards that were capable of running Intel processors at 200MHz clock speeds relatively reliably. Most notable was ABIT, who used high quality components, aggressive memory timings, and easy to set user controls (such as the SoftMenu).

The Chip Hits the Fan

For a few months, our problems appeared to be minimized, other than those who insisted on trying to overclock Cyrix processors, which were notorious for running hot, and were already being pushed pretty hard at the manufacturer’s recommended clock speeds anyways. Granted, there were still some who insisted on trying to run 150MHz processors at 225MHz, and when something gave, would try to get replacements by insisting that the component was somehow defective in design, but the percentage was low enough to not get too worried about.

Sometime around April or May, we began to notice the number of RMAs were rising again. We were puzzled because few people were admitting to overclocking, but the symptoms sure made it appear that’s what the problem was. As we began to quiz people a bit more closely by asking for jumper settings instead of clock speeds, and by investigating the jumper settings on returned boards, we discovered that many of the problems were, indeed related to overclocking. We again began to charge people restocking fees and threaten to void warranties, but were shocked this time when the response we got was not an apology or admission of responsibility, but instead was anger, threatening statements and attempts at chargebacks.

A number of these people truly believed that running a 166MHz processor at 2.5x83MHz was perfectly legitimate, and since Tom (or whomever their favorite overclocking guru) could make it work, their motherboard/CPU/memory must be defective. Many began to insist that we pay shipping both ways – even when they were on their 3rd or 4th motherboard!! The only way to combat this was to tighten our policies and begin charging more restock fees, and institute a tech support fee for products that ran perfectly well at standard clock speeds, though they were unstable at faster speeds. We were rewarded for this by people who would post messages on Usenet or on some of the web surveys about how we were ripping them off, and claiming they were being charged for product that they ‘knew’ was defective. Most tried to (unsuccessfully) charge these fees back to us.

This practice has continued to rise, and almost nobody will admit to overclocking. Of course, we know that a great percentage of our customers come from Tom’s Hardware Guide, and are insisting on boards that run at 75 or 83MHz. If overclocking was not in their game plan, this would not be an issue. But just as nobody reads the National Enquirer, nobody overclocks either (at least not if the product doesn’t boot or is unstable). With the introduction of processors that run faster than 200MHz, the thermal issues are becoming increasingly important even when *not* overclocking, and have created another area of concern, since many people are not aware of how sensitive these components are to even a small amount of excess heat over time

The Usenet posts for the last several months have seen people actually advocating overclocking for everyone, and claiming that nobody has ever damaged a component by overclocking (yes, I have seen these statements several times). The worst case, so these people say, is that the system simply doesn’t boot, and you set the jumpers back to a slower speed and off you go!! It is not unusual to see postings from people claiming to be running 166MHz processors at 250MHz and higher! The incredible naivete of this is astonishing, and the implications are potentially dangerous, even to cautious overclockers.

Pages:   1 2  Next »

Be the first to discuss this article!