The development of the 848 came as no surprise since there was a gaping hole in an Intel offering that would compete against SiS and VIA in the low-cost chipset segment.
At issue was how to convince consumers to replace 845 processors with 865/Springdale and 875/Canterwood chipsets, which, needless to say, meant paying a lot more. In the lower and mid-range price segments, in particular, performance parameters such including FSB speeds, memory speeds and other features are seconded by clock speeds. System makers will thus look to reduce costs by using a chipset that is more affordable, such as the 848P, but slower than the 865 and 875 chipsets.
Already, numerous mainboard manufacturers had found they'd exhausted the possibilities offered by Intel's 845PE chipset, with its FSB533 and single channel DDR333. They also billed the 845PE as 800 MHz compatible, because most mobo makers found the chipset to be very overclockable and offered overclocking settings in order to run it at 800 MHz - which, of course, was not supported by Intel's 845 spec sheet.
These boards featuring a fast P4 for 800 MHz FSB speeds consistently caused problems in spot checks in our lab. Only a few boards were stable at this high speed, and even resorting to leading brands produced only limited improvements. These systems may work in individual cases, but as a rule we recommend leaving them alone. After all, they are quite literally overclocked.
Intel is aware of the dilemma: despite its ripe old age of almost two years, the 845 chipset is still a product that's sold in high numbers - despite every effort by Intel to see high-volume sales for its 865 and 875 chipsets. Indeed, the 845 early this year had already become the de facto successor to the legendary BX chipset, as already reported in several THG articles. In order to answer the demands the market has for a single channel chipset, Intel simply took the 865 and removed one channel. The result is today known as 848.