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.
Politics : Formerly About Advanced Micro Devices
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: milo_morai who wrote (90498)1/30/2000 3:14:00 PM
From: Tenchusatsu  Read Replies (1) of 1572726
 
Milo, <So my point was Made FP matters the Most.>

Remember that when Merced and Willamette are released this year. <BG>

I can say right now that Merced takes a sledgehammer approach (no pun intended) toward floating-point. I guess when you have huge register files, a lot of silicon real estate to work with, and new freedoms enjoyed by moving some of the guesswork to the compiler, you can afford to brute-force your way into FP. Integer should also do very well, although I would agree with Scumbria and say that we are fast approaching the limits of integer IPC (instructions-per-cycle). (Disclaimer: I have absolutely no idea how well Merced actually performs, so don't ask me.)

I don't know exactly how Willamette approaches FP, though. All I know is that everyone working on Willamette (including my roommate) is pretty excited about its potential FP performance.

Tenchusatsu
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  

HomeMailHotSubjectMarksPeopleMarksKeepersSettings
Terms Of UseContact UsCopyright/IP PolicyPrivacy PolicyAbout UsFAQAdvertise on SI
© 2025 Knight Sac Media.  Data provided by Twelve Data, Alpha Vantage, and CityFALCON News