After our initial article came out, we received a lot of e-mails and comments on the article.  While e-mails were commonly sent, we received a lot of discussion on our message board as well. Many of these turned into what really could be considered whole articles of their own.  One post of mine actually came out to be several pages long (at least when I copied it into word).  So, before getting started, we felt that you might be interested in having a look at these. I'll link to the specific forum threads directly below and I want to stress that all of them contain useful information and valid comments. So read them and join the discussion or even raise some more topics :

Questioning the T&L article Started by Newb

Do we really need T&L? Started by rookie

How fast can software T&L be?  Started by MTM

This forums seems Biased Started by Pat

Taking the load of the CPU... Started by Evil_Bob

That being said, there was also a very interesting article that raised some questions for us.  This was written by Milton Teruel and can be found here. While initially this was planned on just being some comments on his questions, it has quickly become an article of its own.  First, we’ll just look at his questions and give some answers to those, and afterwards we’ll look at some other points.

However, before we even get into that, I’d like to address some e-mails we received containing benchmarks results, which were higher than the ones posted in the original article. I want to make some notes about that. First of all, on average, these results still show that a P3 is faster when using 4 or 8 lights in that specific 3Dmark 2000 test. Most results e-mailed did show higher results for the 1 light case when using the GeForce’s hardware lighting. A second point is that many of these results were obtained on severely overclocked GeForce set-ups (not saying this is wrong, but we are going with default clocks here). Also, a low end CPU will not beat a GeForce in T&L no matter what optimisations. Now, none of those comments made the actual contents of the text invalid. If you missed our first article about this matter then you can read it right here.

Now lets get on with it...