Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.
Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.
I attended a conference, where a person hinted that core 2 duo runs two instruction per cycle. Is it true that core 2 duo runs 2 instructions per cycle and that's the major difference between Pentium D and core 2 duo?
The Pentium D module is actually two separate die cores on a multi chip module in the same package. It's like old school multiprocessor systems except both chips happen to be in the same physical package, the multiprocessor bus is just built into the module. The Core 2 Duo is a true dual core chip, where it's all on the same raw die. All things being equal, a Pentium D will run the vast majority of applications faster than a Core 2 Duo.
The reason this is a big deal is because two chips on the same die share the same L1 and L2 cache, where the Pentium D chips have separate caches. This is more a function of application development than hardware because if you have two chips on the same die and chip 1 has to talk to chip2 it's BAM there, they have the same bus. On the Penitum D, if Chip1 has to talk to Chip2 it has to go through an external bus (even though it's in the same package) to get to it, this requires an entire other level of hardware delay and processing. So a Pentium D would be better at running multiple separate applications, where a Core 2 will run an application specifically designed to run on dual core processors better.
Just as a comparison, I have a Core 2 Quad. Which is actually two separate dies in the same package, where each die is a true dual core chip. So it's two separate processors, bound together on a multi-core package, each processor just happens to have two cores.
A true quad core processor is all four processors on the same die sharing the same cache.
I personally think the parallel/serial configuration of the Core 2 Quad is ideal. I can set an application designed for (or otherwise able to use) a dual core chip to one processor (two cores), and set another application to use the other (two cores), and they'll never bump heads with each other, until they have to communicate off module, where some bus contention can occur and will influence things.
To date, I know of few applications that can take advantage of more than two cores. Having a dual dual core machine is better than quad cores in my opinion and given the proper external busing is true in the real world. The biggest problem is the bus system the processor, memory, and other I/O it's running on. Which is why raw MHZ cores and cache size doesn't always equal performance.
Structure is often as or more important than the raw materials that the system is made out of. Take a look at modern advancements in quantum materials, where the structure of the material transcends the raw materials basic functions as the structures causes quantum effects to predominate, such as 'nanomaterials'
This is juuuuuuuust starting to work it's way into industry. Where simple carbon molecules are looked at as an impurity much of the time, where as highly structured mircro crystaline carbon (basically powdered diamond) when properly mixed with an incorporated into another 'ordinary' material can increase wear resistance or corrosion resistance by orders of magnitude, with the same, or less base materials than are used in macroscopic environments.