 |
|
Message boards : Cafe CPDN : Which processors can\'t do maths?
Message board moderation
| Author | Message |
|---|---|
|
Send message Joined: 16 Oct 04 Posts: 692 Credit: 277,679 RAC: 0 |
Different processors are getting different results. Josti found one model which was identical, which involved the unit being crunched by an Athlon XP and and Athlon 64. Thanks to UK Nick I have done another check with his Athlon XP against my Athlon 64 3500+ S939, which also turned out identical. I am not saying that Intels processors cannot do maths, but AMD's processors seem more consistent between different chips. Visit BOINC WIKI for help And join BOINC Synergy for all the news in one place. |
|
Send message Joined: 1 Feb 05 Posts: 4 Credit: 15,948 RAC: 0 |
You mean to say Intels still have the division bug? I recall reading that when the Pentium was introduced, NASA refused to use it and relied on 486's until the was fixed. I think it only effected the very first generation of the pentium Remember the old sig line - I am Pentium of Borg, you will be approximated. |
|
Send message Joined: 16 Oct 04 Posts: 692 Credit: 277,679 RAC: 0 |
I don't think it is anywhere near as bad as the division bug. In this case, occasionally different chips round estimated solutions to an equation slightly differently (probably only on the least significant digit). IIRC the division error would result in something like the 5th or 6th decimal place being wrong. Visit BOINC WIKI for help And join BOINC Synergy for all the news in one place. |
©2024 cpdn.org
