RazberyBandit
Distinguished
[citation][nom]SchizoFrog[/nom]PMSL at the hypocrasy of the AMD/ATi fanboys...Why use GPUs with disabled cores? I'll just say... Phenom x3.[/citation]
Why not add all these to the list?
Kuma CPUs, Phenom II X2, Phenom II X3, Athlon II X2, and Athlon II X3
While you're at it, add the HD4830, 9600GSO (along with numerous other G92 variations), and my aforementioned GTX260.
This sort of thing should be of no shock to anyone who's paid attention over the last few years. It simplifies production tremendously. You design one chip, and one chip only. You take the ones that work perfectly and sell them properly as the flagship. You take the ones that are slightly imperfect, disable the imperfections, then sell them as lesser models.
The problem with nVidia's Fermi is that it doesn't seem as though they aren't getting any "perfect" yields, yet. Though, I can't imagine a card that has 32 of it's 512 cores disabled, yet already draws almost 300W of power remaining within the PCIe 300W standard when all 512 cores are enabled. Even if it barely manages to slip within the limit, there would be zero overclocking headroom.
Why not add all these to the list?
Kuma CPUs, Phenom II X2, Phenom II X3, Athlon II X2, and Athlon II X3
While you're at it, add the HD4830, 9600GSO (along with numerous other G92 variations), and my aforementioned GTX260.
This sort of thing should be of no shock to anyone who's paid attention over the last few years. It simplifies production tremendously. You design one chip, and one chip only. You take the ones that work perfectly and sell them properly as the flagship. You take the ones that are slightly imperfect, disable the imperfections, then sell them as lesser models.
The problem with nVidia's Fermi is that it doesn't seem as though they aren't getting any "perfect" yields, yet. Though, I can't imagine a card that has 32 of it's 512 cores disabled, yet already draws almost 300W of power remaining within the PCIe 300W standard when all 512 cores are enabled. Even if it barely manages to slip within the limit, there would be zero overclocking headroom.