Reverse Hyper Threading.. Hmnn! Interesting!

[Reaps]

It's always the same you spend £££ on parts and something new is round the corner, i really hope they can do something for existing dual core users

[TheAnimus]

It's always the same you spend £££ on parts and something new is round the corner, i really hope they can do something for existing dual core users

You can, only buy stuff thats programmed well!

On all the stupidly easy algorithems/AI assignments i've done at uni (in my final year now) i've always taken time to talk about how you can design it to be multi CPU efficent. Often its quite hard because of accessing the same memory (Mutexs).

But as for cache improvement, I'm going to use half-life 1 here, because i played a lot with that (modding client bots server tricks and the like). Now most instructions that take lots of CPU time, use barly 8 pages of memory (32KiB). Now modern CPUs since P4 netburst style are good at reading memory into cache ahead, the Dothan, is simply fantastic at this, and as such you don't get much 'real world' performance increase going from DDR400 to DDR533. The dothan reads very far ahead, cahcing all options that the proccessor can goto.

As such 4meg of cache in the hands of the dothan team, could be incredibly impressive.

[Butcher]

In order to take advantage of this capability, wouldn't developers pretty much have to compile code with AMDs own compilers? Maybe there's some kind of reverse engineering that could be done, but surely at the expense of speed. If AMD compilers were required to get the most out of this, how would they get developers to use them? I've never seen seperate binaries for AMD before, and most things seem to end up being compiled through intel compilers anyway.

There is no AMD compiler. Intel makes one, but most stuff probably get compiled with microsoft's compiler in actual fact.