Chip firm Intel presented a 65 nanometre dual core multithreaded Xeon with a shared 16MB on die level three cache at the Solid State Conference. The Xeon has 1.328 billion transistors, a 1MB unified L2 cache per core, and has a die size of 435mm2. It delivers 3.4GHz at 1.25 volts and 150 watts TDP, and comes with a 667 and 800MT/s three load front side bus interface. This chip is compatible with existing chipset designs and Intel claims it has the largest cache and device count for an X86 processor.
Is that debranded hyperthreading? Or are they just repeating ‘dual core’ in different words?
Yeah, “multi-threaded” just means hyperthreading. I don’t know why the headline focuses on that aspect, since it’s not new. The new part is the huge cache.
Decisions, decisions… obsolete cores with huge cache (Tulsa) or cool new cores with merely large cache (Woodcrest)…
Or how about AMD procs? Arent AMD Opterons better than anything in Intel’s current arsenal? That is the first time I have heard of a processor with over a billion processors!
Yeah, “multi-threaded” just means hyperthreading. I don’t know why the headline focuses on that aspect, since it’s not new.
True, but lets hope they stick with “multi-threaded”, because the word “hyperthreading” is an annoying and meaningless invention by Intel’s marketing department.
Hey – Whatever gets the Job done, right?
As long as it’s generally faster than predecessors, reasonably energy efficient, priced competitively, and runs my software without a hitch (stable)….
Then they can call it whatever they want…or build it however they want.
Edited 2006-02-08 16:37
In this case, I think Tulsa will be expensive and energy-inefficient. It’s probably a niche product for apps that can really benefit from the 16MB cache.