Intel Counter Attacks!

Throughout 1998, Intel saw their market share steadily decline. The most recent reports show that Intel’s share of the retail market has dropped below 50%, with their overall share of the microprocessor market dropping below 80%. Most of this erosion has come at the hands of rival AMD, whose K6 processor line garnered much media attention and was the catalyst for the sub-$1000 systems that have been selling in huge quantities for the past year. It is apparent that Intel was unprepared for this onslaught, and it has taken them a full year to respond properly – but respond they have (see the table at the bottom of this page).

When the K6 was introduced in mid-1997, Intel was ignoring AMD and pushing the Pentium II processor in the belief that Socket 7 had been all but buried. By the beginning of 1998 it became obvious that the K6-233 was priced right and had enough power to convince many low-end users to try AMD for a change. Intel’s response was to release the Celeron processor with no L2 cache, which was soundly trounced in the popular press. Though Tom’s Hardware Guide had correctly identified the Celeron as a strong contender at the low-end for gamers as well as for overclockers, the damage had already been done and the Celeron was tagged as a poor alternative for the K6 or K6-2 CPUs.

The first indication that Intel had recognized a serious threat to their dominance was the release of the Celeron 300A and 333, which were based upon the Mendocino core and sported 128KB of on-chip L2 cache, effectively killing the original Celeron as well as the low-end Pentium II processors. AMD’s response was to release faster K6-2 CPUs, with their 400MHz announced in December. In the meantime, while AMD was making noise about the upcoming Sharptooth (K6-3) and K7, Intel was quietly altering their roadmap to fire directly back at AMD.

As reported in a previous news report, Intel pushed the release of the Celeron 400 up from mid-’99 to Jan.4. In addition, they have added several additional Celeron processors to the mix. As can be seen in the table below, the next several months will see the introduction of a 433MHz and 466MHz Celeron (using a 66MHz FSB) in both SEPP and PPGA packages. By mid-year, the roadmap shows a 500MHz Celeron with a 100MHz FSB. All of this while AMD is still trying to get their Sharptooth to market. Most reports indicate that AMD will initially release the Sharptooth at 400MHz, followed quickly by the 450MHz, however it appears that by the time this happens, Intel will have regained the title of ‘fastest low-end processor’ with the 466MHz Celeron. By the time the Sharptooth is available at 500MHz, Intel will likely have their 500MHz Celeron to market as well.

How well Intel is able to stop, or even reverse, their market share loss depends upon whether AMD can convince users that a 100MHz system bus and L3 cache is better than the superior floating point operation and 66MHz FSB of the current Celeron line. Tests have shown that with the L2 cache running at full processor speed, the system bus speed is of little consequence, however it remains to be seen how much of an effect the L3 cache will really have on performance.

Of course, pricing will also play a big part in this processor war. AMD has consistently shown the resolve to accept lower margins for additional market share, and with good effect as earnings have steadily improved over the past 4 quarters. All of this has not been lost on Intel, as they have introduced their 400MHz Celeron with fairly aggressive pricing at under $200 each.

In the final analysis, it is the end user who benefits from this war. While those on both sides of the battle are wishing for a ‘knockout blow’ to be delivered, it is in the best interests of the consumer that this doesn’t actually happen. Many articles have been written, and passionate newsgroup threads have flourished about the relative merits of the Celeron, the Sharptooth, the K7 and the Pentium III – with each author proclaiming one or the other as ‘the best processor’. What this actually indicates is that the ‘free market’ concept of competition is working relatively well at the present with many choices available and the lowest prices ever. It remains to be seen if this parity will continue to exist into the 2nd half of 1999.

Intel 1999 Roadmap (as of Jan ’99) for desktop processors categorized by intended system prices
  Q1 ’99 Q2 ’99 Q3 ’99 Q4 ’99
Basic (<$1.2K) Celeron 300-400MHz (66MHz FSB) Celeron 333-433MHz (66MHz FSB) Celeron 366-466MHz (66MHz FSB) Celeron 400-466MHz (66MHz FSB), Celeron 500MHz (100MHz FSB)
Performance ($1.2K – 2.5K) PII 350-450MHz (100MHz FSB) PII 400-450MHz, PIII 450-500MHz (100MHz FSB) PIII 450-500MHz (100MHz FSB), PIII 533MHz (133MHz FSB) PIII 500MHz (100MHz FSB), PIII 533MHz, Coppermine 600MHz (133MHz FSB)
Professional (>$2.5K) PII 450MHz (100MHz FSB) PIII 500MHz (100MHz FSB), PIII 533MHz (133MHz FSB) PIII 533MHz, Coppermine 600MHz (133MHz FSB) Coppermine 600-633MHz (133MHz FSB)

Be the first to discuss this article!