hyperthreading vs hypertransporting technology

Status
Not open for further replies.

robina_80

Banned
Messages
902
Location
LONDON
i know what HT means but not really HTT is HTT the actual speed at which all the hardware talk to eachother so is HT better?!?
 
thats what i thought its amds new FSB and i know intel use QPI but what about intels HT has amd come up with there own technology like that coz thats making me move towards intel as it gives it 4 more virtual cores over amds just quad core
 
thats what i thought its amds new FSB and i know intel use QPI but what about intels HT has amd come up with there own technology like that coz thats making me move towards intel as it gives it 4 more virtual cores over amds just quad core

As I've heard it, AMD is in the development stages of something similar to Hyperthreading, though I don't think they can implement it on their current architectures.

Now, Intel's Turbo boost, AMD has already announced something similar I believe called C-state performance boost. Or something like that. As far as I know when you're using a single threaded app, it will actually turn off un used cores AND overclock the single core that's left on.
 
thats what i thought its amds new FSB and i know intel use QPI
HyperTransport is the name of the bus (communication technology) between the CPU and RAM.

the Front Side Bus was used between the CPU and northbridge, which traditionally contained the memory controller, which communicated with the RAM.
AMD moved the memory controller to the CPU with their K8 architecture (Athlon 64), so that it could communicate directly with RAM
Intel recently moved the memory controller to the CPU with their i series processors.

So now there is no FSB with any AMD CPU for about 7 years, or with Intel's i series CPU's.
but what about intels HT has amd come up with there own technology like that coz thats making me move towards intel as it gives it 4 more virtual cores over amds just quad core
Look up AMD's Bulldozer architecture.
It doesn't do the same thing as HyperThreading; but basically AMD thinks it can achieve even better multithreaded performance than HyperThreading would give with very little die size increase over what they have.

From my understanding, HyperThreading is used as an attempt to keep execution units busy whenever there's a cache-miss.
It could really help the Pentium 4 mainly because it had a very long pipeline, and cache-misses were very costly for performance.
With the Core 2, Intel didn't think it really needed that band-aid, since it has much shorter pipelines.
Each core of the i series chips have more logic circuitry added specifically so that HyperThreading can improve efficiency. I'm not exactly sure how.

From the user's perspective, it looks like it's making one CPU core appear as if it's two, and allowing more threads to be executed at once.
Really, it can still only execute one thread at a time per execution unit. But if one thread is stuck because of a cache-miss, it can execute a different thread in the meantime, and then go back to executing the previous thread once the correct data has been loaded into cache.
 
Status
Not open for further replies.
Back
Top Bottom