Intel Celeron D
Written by Michael Larabel in Processors on 13 November 2004. Page 2 of 6. Add A Comment


Below are pictures comparing the Intel Celeron D against a Pentium 4 530 (3.0GHz), Pentium 4 2.4C, Pentium 4 1.6A, and AMD Athlon XP 2500+ Barton, all of which coolers are considered stock and included with the retail CPU.

We decided to include the AMD Athlon XP 2500+ Barton (Socket A) heatsink, simply for comparison of size between AMD and Intel coolers. It is interesting to note the changes that have occurred with the stock CPU coolers over time, with the oldest heatsink in the picture above taken off our old Pentium 4 1.6A testbed. This heatsink used entirely different mounting clips than what we are used to seeing today with the stock Socket 478 levers, which you pull on each side to lock them into place. The 1.6A cooler used simple metal clips which rest on the side of the heatsink and latch to the retention frame. On the bottom of the 1.6GHz HSF was only aluminum which was in contact with the few aluminum fins. Going in chronological order, next we see the Pentium 4 2.4C. This stock HSF continues to use aluminum for the base and fins, but now uses the two levers for latching the heatsink into place, which we are used to seeing. Next up, we see the Pentium 4 530 (3.0GHz) cooler. Being designed for the LGA775 processors, the heatsink is significantly different from what you will find with a Socket 478 HSF, due to its pushpin mounting and size. However, there are some common traits between this cooler and the Celeron D coolers. Finally, onto the Intel Celeron D heatsink. When looking at the heatsink, it appears to be a mix between a blend of the Pentium 4 530 and 2.4C coolers. The top of the heatsink with the fan and mounting levers are similar to the 2.4C counterpart whilst the bottom of the heatsink is similar to the Pentium 4 530 HSF with its radial fin design and copper base insert. It's interesting to see how Intel heatsinks have changed over the years for the Socket 478 processors. All of these changes have been significant improvements to better improve the dissipation of heat.

When looking at the Intel heat spreader on top of the processor, we see our processor was manufactured in the Philippines. At the bottom of the processor, we see all of the pins, and to our benefit, none of them were bent.


With the Celeron D encompassing a Prescott core with 90nm manufacturing, our standards were set high for overclocking. We knew the capabilities were there to overclock high on this budget processor when compared against its older siblings with Williamette core's which use 0.18 Micron manufacturing, and even the Northwood's with 0.13 Micron manufacturing, since they require more power to operate thus often producing extra heat than using a 90nm process. Right from the get go, after flashing our Abit IC7-MAX3 motherboard to BIOS version 17, we raised our FSB to 188, which would have been a 1GHz overclock, and increased our voltages slightly. However, at those speeds our system failed to boot. It would POST but would fail after displaying the BIOS information. We then backed off to 180Mhz, and the system successfully booted. This overclock resulted in the CPU running at 3.24GHz, a 0.84GHz overclock. We had to increase our CPU voltage to 1.40V. At these speeds, we could successfully boot our operating system and our system was stable by running CPU Burn-In v1.00. No matter how much we increased the voltages further, it did nothing but hinder our overclocking abilities.

Related Articles
Trending Linux News