You're chugging along on that P166 and you know its time to upgrade. You could bite the bullet for a PII 350 or 400, but add the price of the CPU to the new mainboard and your wife will skin you alive! What are you gonna do?? How about a 450 MHz Pentium II CPU for $225 US?

With the BX 100MHz bus now in the mainstream and memory prices low and 3d accelerators taking us where no man has gone before, it's near to nirvana for gamers. To add to the fun, August 24th saw the release of two new versions of the low-cost Celeron processor and a 450-MHz Pentium II.

But hang on.. there is more here than meets the eye! This is not just another pretty Celeron face. The new Celeron A (formerly code-named "Mendocino" is important to you if you are about to make a CPU upgrade. What the heck is the Celeron, let alone the "Celeron Pro?" I'm glad you asked!

Remember the 486SX, the early 32 bit processor with the disabled math coprocessor? The Celeron is a similar idea, a Pentium II core chip designed to work with secondary cache memory, but Intel removed the cache to reduce cost, leading to a degradation in performance.

Running at 300 MHz and 333 MHz, the new Celeron A greatly improves upon the performance of the first Celeron chips. These new CPUs contain 128K of integrated "cache" memory. Integration means that the cache runs at full CPU speed, further improving performance.

MS CFS Flak Field.

All this sounds great, but how do you get from 300 MHz to 450? I journeyed the road myself this week, so here I am, still alive and kicking, to share the story....

It Was a Dark and Stormy Night

It was quite a stormy process getting this hardware happening, and I have to back up three weeks to tell the tale. I first received my new CPU three weeks ago, popped it promptly into the socket, and attempted to fire up my ASUS P2B at 3x100 MHz. The power supply was heard whirring eerily in the silence of my office. Not another sound did I hear. Not that lovely telltale beep, and my Viewsonic 19 stared blankly back at me.

Hmm. Obviously, this would require some work. I assumed that the problem was the clock locked multiplier on the Celeron, so I rolled up my sleeves, reclocked to 4.5 x 66, and tried again.

Still nothing. Conclusion: dead CPU or dead mainboard. I transferred the likely offender into the slot of my main system, clocked at 4.5 x 66, hoping that my spanking new ASUS mainboard was the offender. To my chagrin, the same symptoms manifested on my old ASUS P2L97 mainboard.

Dead CPU, I concluded matter-of-factly. My wife heard the whimpers from my workshop and called out supportively, "Told you it wouldn't work!" I popped another valium and sat down to consider my plight.

The next morning I shipped the CPU back to the vendor, and then late this week a new Celeron 300A arrived at my door. Eagerly I unwrapped the precious cargo. Vaulting over the obstacles on the stairs I ran to my workshop and popped the new CPU into its slot. Nada. Same as before.

But this time I had a possible solution in mind. I had already downloaded the BIOS update for the ASUS P2B, since the older BIOS did not have support in place for the Celeron A. Unfortunately, this meant I now had to cannibalize my main system so that I could boot the machine to a DOS prompt to flash the new BIOS. A bit of a pain, but a necessary step. I pulled my PII 300 CPU and popped it into the ASUS board.

I then inserted the ASUS CD, copyed the AFLASH utility to a bootable disk, and restarted. I flashed the BIOS with the BX2I1005.AWD update, loaded the defaults, and shut it all down again. I booted WIN98 to be sure all was well, then shut down, removed my PII CPU, and installed the Celeron. I reset the jumpers to 4.5x 100. Fingers, toes, and eyes crossed, I prepared to restart.... ten, nine, eight, seven...

I then inserted the ASUS CD, copyed the AFLASH utility to a bootable disk, and restarted. I flashed the BIOS with the BX2I1005.AWD update, loaded the defaults, and shut it all down again. I booted WIN98 to be sure all was well, then shut down, removed my PII CPU, and installed the Celeron. I reset the jumpers to 4.5x 100. Fingers, toes, and eyes crossed, I prepared to restart.... ten, nine, eight, seven...

Click to continue . . .

About the time I hit seven tragedy struck. You knew this wouldn't be simple, didn't you?

Yes, just as I hit seven two things happened simultaneously. First, a round rolling object came into my workshop. Second, the cat that was curled up on the chair behind me, being sensitive to unidentified rolling objects and having been roused from a deep sleep, shot straight up into the air and landed on my back, claws extended.

Shouting and diving lest there be Commandos making an attempt on my life, I smacked my head into the clean and new mid tower case housing the guts of my new system.

As I pondered the meaning of life the machine sprang into action, and I caught the process in time to hit DEL and get to the BIOS screen. Lo and behold, I now had a system running at 450 MHz!

The next step was to exit the BIOS and attempt to load WIN98. This is where the second snag caught me, although not completely unaware. WIN98 hung shortly after the splash screen.

Oh-oh! Have I exceeded the limits? Have I fried my CPU? My life flashed before my eyes. My hopes and dreams were crushed. I contemplated eternity with a PII 300.

But I also checked the CPU. Hmm. It was good and warm, but not really hot. If overheating wasn't the problem, and others have successfully overclocked the CeleronA to 450, what was up here?

Pushing a CPU this much past its native speed requires more juice. The Celeron series are built at .25 microns and run at 2.0 volts, an incredibly low power consumption. But this also restricts their power output. I had read this from a message base on the ASUS web site.

I had to find a way to boost CPU voltage from 2.0 to 2.2 volts. The problem now lay with the ASUS mainboard itself. Unlike the Abit BX6 mainboard, there is no way to change the voltage supplied to the CPU. Abit is currently the only manufacturer that supplies a SOFTMENU utility that actually allows one to alter the voltage up to +15% over default.

The only way to accomplish a boost from 2.0 to 2.2 volts on the ASUS P2B is to actually alter the pin connections on the Celeron CPU itself. Is this a difficult process? If you have ever applied a bit of tape to a small wire, you already have the technical prowess necessary to accomplish this feat of micro engineering.

Nevertheless, I was intimidated. Now I was definitely voiding my warranty. I was preparing to enter the VOID. I was facing DEATH at 450MHz. Ah what the heck. My readers need me.

Actually, there are a number of ways that one can accomplish the pin masking task, and I found the discussion at Ultimate PC. One can use teflon tape, or the tape used to splice a cassette tape. Some have tried model airplane dope or even nail polish, but I don't recommend the latter since I had one report of an attempt that failed. Apparently when the system heated up one of the polish came off one pin enough to allow contact and the CPU suddenly went from 2.2 volts to 2.6!

In order to change the voltage on the CeleronA CPU one need only mask three pins, as shown below...

The pins that must be masked on the front side of the CPU board are A119 and A121. The front side is the side with the CPU core protruding. Only one pin needs to be masked on the back side (the side with the pin holes only): B119.

Go to Part II