SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?


To: Tenchusatsu who wrote (161347)3/7/2002 11:54:46 AM
From: Dan3  Read Replies (1) | Respond to of 186894
 
Re: Banias doesn't include an on-die memory controller. At least not that I know of.

Thanks for the correction. I can't find where I saw an indication that it did, perhaps I remembered incorrectly.

To me, that's bad news for Intel. I'm really surprised that Intel isn't going with any on board controllers. Considering how may cycles are lost when a cache miss occurs, reducing latency has got to provide a larger and larger performance increase as clock speeds increase. Obviously, the better the cache, the fewer the misses, but some misses are still inevitable, and cutting latency by 20 nanaoseconds would eliminate 50 stalled cycles in a 2.5GHZ CPU. If 1 in 100 memory accesses is a cache miss, then there is a cache miss every 40 nanoseconds.

Now, cache hits and misses aren't evenly distributed, so the impact of reducing latency won't ever reach the theoretical max, but it is clearly possible to save 20 nanoseconds every 40 nanoseconds, resulting in a huge performance increase.

Too bad banias (and P4, Athlon, and Itanium) doesn't have an on-board memory controller.