Silicon Investor (SI) -- The First Internet Community

STOCKTALK

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: Scumbria who wrote (119243)	11/26/2000 11:26:35 AM
From: Saturn V	Read Replies (2) \| Respond to of 186894

Ref < Lists and pointers are problematic for P4....This type of software is problematic for P4, because the instruction flow and data accesses are highly unpredictable. > Scumbria I strongly disagree. I have written lots of code using lists and pointers. Yes, the data in large lists may not be easily predictable. But the code can still be predictable. I have written simple loops for accessing and searching lists. The complaint against P4 is the branch mispredict penalty, i.e code which cannot be expressed in simple loops. The branch target for manipulating lists should be highly predictable, since the code will fit inside a small loop. The dataset while manipulating the list may not be predictable, and the data may fall outside the data cache. This problem is common to all processors, and there is nothing which about P4 which makes it worse. If anything the higher bandwidth of the P4, makes it faster to recover from data being outside the data cache. The other weakness of the P4 is the lack of barrel shifting which makes it slow on multiple shifts. That is not relevant to lists and pointers. Thus P4 relative weakness on Office has nothing to do with list processing. I confess that I do not understand the reasons for its weakness on Office applications.