alephnull said:
Squilliam said:
alephnull said:
Squilliam said:
Rainbird said:
Squilliam said:
MrBubbles said:
would this be a good deal for microsoft or are there better options? |
They could make Intel give them a pretty nice CPU as part of the deal and get it CHEAP. It would be quite an awesome deal if it did go through for both parties.
Intel gets one of the biggest/best software tool development companies in the world working on their chip and gains mass market adoption. Microsoft gets a cheap deal. So its a win/win if they can agree really.
|
Except it could put Microsoft in the position so well known from the PS3, with complaining developers. And as far as I understood, Larrabee is a combined CPU and GPU..?
|
Except Microsoft makes excellent tools and has an awesome developer relations. If they had made it, most likely the tools available in 2005 would have been much better.
|
This all assumes intel can even make a decent compiler for it. Remember the Itanic?
|
AFAIK it would be Microsoft creating the API, GPUs don't need compilers.
|
So, this thing aparently aparently uses the x86 ISA plus other stuff so you are going to get another C+intrisics compiler like cell , which btw has taken it's time in becoming stable. This thing is also supposed to magically maintain cache coherency between 32 cores with no massive overhead (or nasty tradeoff). This thing must be unicorn and dragon powered. I wish I could short sell just this division of intel.
|
Larrabee can hide the latency of incoherent reads as it uses L1 to accumulate the data before the thread resumes.
It seems to me that a core is defined as logic+L1+L2. Each L2 is only used by its core. Cores can only access foreign L2s under the cache-coherency protocol, which is effectively a request to fetch data to make a local copy
link
