Replies to post #54251 on Advanced Micro Devices Inc (AMD)

[alan81]

Somebody has to go first...
The hardware guys can wait for the software while the software waits for the hardware, and nothing will ever happen.
In the case of dual core, Intel did a good job getting the ball rolling with hyperthreading. The OS is ready BEFORE the silicon. Many important applications are already threaded. The one application I run that is too slow is my video editing software, and it is already threaded. The other case being running multiple applications, which threading helps with. I do agree multi-core will not be a benefit to gamers for another year or two.
The paths the two companies are taking make perfect sense given their unique circumstances. I think that has been well covered in the past, but for review...
AMD has limited capacity until fab 36 comes on-line. Thus, they should use that to make dual cores where they add the most value which is clearly in servers. The plan holds water.

Intel had excess capacity (not really anymore because they are making dual core products). Intel has just introduced new products into the server market that significantly improve performance over their prior products. Dual core in Intel servers with current 1M cache products will not improve performance very much due to memory bandwidth limitations. Intel plans to have the majority of their output converted to dual core by the end of next year. The plan holds water.
--Alan

[sgolds]

bobs10, I'm not sure your statement is technologically correct -

It looks like 64 bit games will appear in volume way before they're multi threaded.

The better games have been multithreaded for years, that is how they achieve smooth user input simultaneous with video and audio output. It would seem to me that games will get the best boost from multiprocessed systems, along with video editing programs.

I wonder if there are any dual socket gaming benchmarks to test my assertion.