Jim, Rambus isn't the only Intel-supported technology that Tom was against. Tom was also against Intel's move to Slot 1 and to AGP. Tom was also gung-ho on 3DNow! and ho-hum about Pentium III's SSE instructions. You should have seen all of the arguments against each. If I remember correctly:
Slot 1:
- Intel is artificially forcing the market away from Socket 7. (No, Socket 7 was unable to handle off-chip L2, which was necessary at the time)
- AMD was able to take Socket 7 to 100 MHz. (Big deal, AMD never made any money off of Socket 7 processors. Only when AMD switched to a Slot 1 wannabe, the Slot A form factor, did they finally realize profits.)
- AMD's K6, a Socket 7 processor, will be faster than Pentium II, (which never held true except in a few cases which Tom trumpeted.)
AGP:
- No graphics card will ever need the texturing ability of AGP. (This was true for the first 1.5 years after AGP release, but not true anymore. Even Tom's latest tests shows the value of AGP.)
- Voodoo2, a PCI card, trounced the AGP cards at the time. (Voodoo2 is ancient history now; almost all graphics cards these days are AGP.)
- AGP was yet another attempt by Intel to control interfaces. (And the alternative was what, Tom?)
Pentium III's SSE:
- 3DNow! had a nine-month headstart. (Big deal, SSE now has tons more support than 3DYesterday.)
- The industry doesn't like monopolies, which is why 3DNow! will triumph over SSE. (Again, the truth is that SSE has triumphed over 3DNow!.)
- AMD K6-2 outperformed a Pentium II at 300 MHz running Quake2. (K6-2 failed to keep up the performance as clock speeds went higher. Nowadays, a Celeron is able to outperform a K6-2 even in Quake2, and Celeron has neither SSE nor 3DNow!.)
So Tom has a long history of being anti-Intel and of conjuring arguments that can convince any casual reader of his web site. But one serious look at his history shows that Tom is wrong more often than he is right about Intel.
Tenchusatsu |
