WraithPriest said:
http://charliedemerjianisadouchebag.blogspot.com/
Also, due to charlie everyone on the net is screaming and repeating "1.7%!" because he bungled a translation. Or alternatively "AMDs Competitive Analysis Team" bungled it and, as he works for AMD, he repeated it without checking it.
Charlie has got dozens of things wrong. The only thing he's ever been at all correct about was the defective materials in some nVidia mobile GPUs and even then he got the how/why and whens wrong.
Even a broken clock is right two times a day. Charlie -may- have alot of contacts with AMD/ATi (Due to working for them - coincidentally why TheInq fired him when they found out) but he has NONE inside TSMC or nVidia and has admitted so himself on semiaccurates forums.
Also, i saw you claimed that Charlie "leaked" the correct shader count for the 58XX. Its not really a "leak" if he posted it on semiaccurate the SAME DAY as ATi told everyone else and it appeared on their websites. I could claim i knew beforehand, but with nothing to back it up that claim is worthless.
|
Ah yes, I knew it would come to this. I'll respond to each point.
- Lol, that blog is dead and rants far worse and far more biased than Charlie himself does.
- Charlie says his source wasn't a translation; he checked with Nvidia people. If you assume he's credible (and he's posted a lot more detail in forum posts about how he sources this stuff than any other tech rumour journalist, even explaining why he dislikes Nvidia practrices but is friends with their engineers) then he isn't wrong yet. As I said, his credibility is the key point - if you don't believe he has any then you're free to say he's wrong and I won't argue.
- Really? Can you quote some specifics he got wrong? Sometimes it's Nvidia making an unexpected decision (e.g. he said dual GT200 wouldn't appear in quantity, because it would lose millions and Nvidia would be stupid to do it - but they did anyway and if you look at their financials it did hurt them.) And he admits and explains mistakes afterwards, if you ask him on the forum. He was correct in GT200 being far too big for its perfrmance; it has a 10% lead over RV770 for double the die area. They got the performance lead but at a big financial cost. He was also correct in GT212 and other 40nm GT200 derivatives being heavily delayed or cancelled; any company that wasn't having issues would have launched a 40nm shrink of their current chip before a new one (e.g. HD 4770 testing 40nm or 8800GT testing 65nm).
- We will soon find out. If Nvidia launches GT300 this quarter or has good yields early next then I will agree with you. If they do not, he must be right since everyone else running a rumour site on the web now is predicting a Q4 launch - Fudzilla, BSN, etc.
- Proof? Some links dated on the same day as the SA article would be appreciated, I did not know hat. If it is like that, then it's because he still has a business relationship wih AMD and Intel so doesn't leak stuff even when he knows it. He got cut off from Nvidia PR so he's free to leak as early as possible.
In the end, it doesn't matter whether Charlie is right. GT300 will still come out the same - either late and poorly yielding or a Q4 launch. If it is the former then Nvidia's financials will suffer further, and that's bad for both companies ($380 is overpriced for the 5870 IMO, lack of competition ironically).
I'm not biased, I just want to see a competitive CPU and GPU business. Right now, that means arguing for AMD CPUs when they are better value than people realise (I wouldn't argue for 965 BE over Lynnfield for example), and AMD GPUs until AMD's graphics segment is profitable and they have a marketshare that fairly reflects their current technology lead.
