我是在MURC上看到的
原文是出自NVnews
我把他轉貼過來如下:

Just a few Parhelia thoughts, for what it's worth.

First of all, did anybody notice something different/missing? IE, there are *some* websites that are normally "all over" new product releases. Some of these guys weren't up to scratch, according to Matrox.

In truth, one can easily surmise that those websites that do NOT have Parhelia reviews posted this morning...Matrox has clearly given her opinion. Undoubtedly, Matrox feels that the bias exhibited by those websites is so extreme that they could be placed on the backburner for a few weeks. Anybody else feel as good about this as me?

To make matters more humerous, look at what some of these websites that aren't featuring Parhelia reviews this morning are saying to their readers. Clearly, they feel slighted. And in other cases, there's barely a mention of Parhelia. Funny how that works, eh? Under normal circumstances, a typical hardware site will splash links to all the other reviews, press statements, etc. When you don't get a review unit, no coverage is rewarded.

The fact of the matter is this: websites that did not feature a Parhelia review this morning are, in the eyes of Matrox, biased to the point where Matrox felt it better to distribute their shipping boards elsewhere. Instead of "sucking it up" and realizing their worst fears (IE sites bashing them because it's not a Ti4600 killer), they chose to ignore them entirely, and send review units to those websites which will give Parhelia a fighting chance.

Lastly, about the Parhelia performance results. If you were to evaluate this board based on nVidia's own testing methodolgy, you would come away crowning this chip a true champ. Why? Because, _nobody_ cares about standard benchmark results anymore. nVidia has already gone on record stating this fact countless times, and this will certainly continue with NV30. We're well past/beyond the days of bad-IQ performance. If I'm going to spend some money, I want to know that I can actually crank up the features and still play the game. In this regard, the Parhelia has won. In the vast majority of benchmarks, Parhelia came out on top, or close to it. Of course, Matrox's Antialiasing method is significantly better looking than anything else on the market as well.

__________________________________________________________

蠻有趣的想法，事實上其實也蠻贊同其中一些論點的
BENCHMARK的分數真的並不足以代表什麼吧...
且另外注意到的一點，不知小弟有沒有搞錯，不知為什麼，似乎目前的測試報告大部分好像都出在德國的網站？

總之算是幫Parhelia說點話吧，還是蠻期待他的出現的，雖然他有點貴，但誰知道台灣到時候會賣多少呢？

附註：剛剛又去看了一下，似乎不是全都在德國啦，只是覺得非英語系的報告似乎蠻多的..