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.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Mani1 who wrote (102402)	4/5/2000 7:06:00 PM
From: chic_hearne	Read Replies (1) \| Respond to of 1575706

Mani Re: chic, thanks for the link Mani, This link is much better: jc-news.com chic

To: Mani1 who wrote (102402)	4/5/2000 10:35:00 PM
From: Jim Davis	Read Replies (3) \| Respond to of 1575706

"chic, thanks for the link, I know Scumbria had talked about this "BRANCH MISS-PREDICTION PENALTY". I wish I had a clue about his stuff. I would really welcome a discussion about this topic by those who have some expertise." Simple! Consider it solving a maze. You come to a branch and have to guess which way to go. You pick the right way that's great and you just blast ahead. If you pick the wrong way and get blocked you have to back track to start over at the last branch. How fast you finish depends on guessing branches right, going down the paths fast and back tracking fast to go down the other branch. On a processor these are branch prediction accuracy, processing speed and BRANCH MISPREDICT PENALTY. When there is a branch mispredict the processor throws away all the work on the wrong path. Then it has to find the address of where it was really supposed to go, go to the instruction cache to get the instruction, decode it and issue it to an execution unit. The number of processor clock cycles required to do all that is the branch mispredict penalty.