Re: "you are confusing "invalidate transaction", which is
address-only transaction, with flushing of actual data.
The two operations were split explicitly to save
system bandwidth."
Nope. I am well aware of how some chipsets implement these as different transactions. The Ace's article, however, was talking about the former, while you criticize them because you think they are talking about the latter. You are mincing the word "flush", because according to you, it represents a different kind of transaction, when I have told you multiple times how I interpreted Ace's comments. All this time, I have been arguing against taking offense at the slight miscommunication in the Ace's article, but apparently, you want Ace's to read like a Harvard textbook. I don't know what you have in mind for the Ace's site, but I doubt it is any more than a few knowledgeable enthusiasts giving their point of view, and I still maintain that they got their point across, even if they happened to miss a few implementation specific details. Your diatribe is yet another example of your nit-picking obnoxious behavior, which gets more than a few people annoyed and wondering whether your inferiority complex requires you to be a jerk in real life, too, or just on the Internet.
wbmw |
