Originally posted by jamesfuh@Dec 23 2003, 09:07 PM Super Pi 32M當然結果立判...5-2-2-2效能技高一籌
問題就在為何Super Pi 1M時...11-2-2-2的組合能快一秒
在世界級競賽時...讓OS + 11-2-2-2設定組合...豈不很容易四捨五入...Ya Pee!
leonlianing的決勝Barton 2500+ Super Pi 39sec...不正是11-2-2-2如此來的! ^__^
Originally posted by 狂少@Dec 23 2003, 10:38 AM 多謝lawlawman兄資料!
以其作者之意:
"My original theory that it was because it was a dual-channel board and was using two sticks of RAM, it needed a higher tRAS latency to read all the data off of the Row, was disproved when the same results came up running in single channel mode, with one or two sticks. So as to why this happens, I am stumped. If anybody has any theories, please feel free to contact me, I would love to hear them."