I was just wondering, why didn't amd take the thuban, and do the following at the start of this year;
1.Remove the ddr2 mem controller, and am2+ support.
2.Support atleast upto 1600mhz ddr3, if 1866 isn't possible,
3.Remove 3dnow! instruction set, and add sse4.1 instead (or does this require a lot of chip redesigning and die area). No need of sse4.2, aes etc with this chip, let that come with bulldozer..
4.Introduce Turbocore 2.0, or somethin better than the current turbo!!!
5.Sell the chip as an AM3+
Also,why not manufacture it on 40nm (TSMC), just like the zacates, so that it gives off less heat = higher clock speeds. Why do intel and amd always scale by a fixed factor of around 1/√2.[130nm -> 90nm -> 65nm -> 45nm , and so on]
Don't u think such a 'Phenom III' sold from Jan, to say August would have certainly allowed AMD more time to get bulldozer, and 32nm tech right, rite? I mean,i don't think a PII 980 or the 1100T is really all that competitive right now!!
1.Remove the ddr2 mem controller, and am2+ support.
2.Support atleast upto 1600mhz ddr3, if 1866 isn't possible,
3.Remove 3dnow! instruction set, and add sse4.1 instead (or does this require a lot of chip redesigning and die area). No need of sse4.2, aes etc with this chip, let that come with bulldozer..
4.Introduce Turbocore 2.0, or somethin better than the current turbo!!!
5.Sell the chip as an AM3+
Also,why not manufacture it on 40nm (TSMC), just like the zacates, so that it gives off less heat = higher clock speeds. Why do intel and amd always scale by a fixed factor of around 1/√2.[130nm -> 90nm -> 65nm -> 45nm , and so on]
Don't u think such a 'Phenom III' sold from Jan, to say August would have certainly allowed AMD more time to get bulldozer, and 32nm tech right, rite? I mean,i don't think a PII 980 or the 1100T is really all that competitive right now!!