Replies to post #28238 on Advanced Micro Devices Inc (AMD)

[dougSF30]

wbmw, in spite of your rude, childish reply, let me address a few points:

The specs *are* 1.6 GHz, 256K L2, 3000+. So if you trust the 3000+ rating, what does that imply about the IPC w.r.t. the standard K8 core? Did you put *any* thought into your answer?
The alternative, of course, is that this 3000+ is not the same as an A64 3000+ -- that is, a different target has been chosen for the comparison, like Celeron.

Regarding laptops and power dissipation, well the laptop dissipation is what really matters, isn't it? Furthermore, the two products seem to have similar components.

Regarding Banias v. Mobile A64, try 1.5-2x

You have no idea whether the Dublin core represents an improvement or not, and the specs & PR rating suggest there may well have been a nice IPC boost. If not, I'm glad to see them exposing the P4 Celeron GHz scam.

Regarding Odessa, that's the next 64-bit mobile part, so of course it is higher than a 32-bit only part. As for being the "bottom box", it's the lower of *two* lines in the mobile portion of the roadmap, so that's yet more sophistry from you.

Doug

[chipguy]

I just find it interesting that you somehow expect Dublin to be this miracle design improvement with lower power and higher performance, even in spite of the smaller cache and the fact that it represents no major core improvement on the roadmap.

Here is another thing to ponder. If Dublin has a 256 KB L2
then it will access memory on average twice as often as a
regular A64. A memory interface burns a *lot* of power for
a mobile chip because it generally doesn't scale down in
frequency with the CPU and it never scales down in voltage.
AMD seems to be going in the wrong direction in its mobile
offerings. The Dothan has 2 MB L2, twice the size of the L2
in Intel's 90 nm desktop chip.