From one of the people that actually had a big hand in writing the code for PhysX, Pierre Terdiman.
The history of PhysX
I came across his blog whilst searching for stuff on the internet, and I was surprised to see him quote a thread that I started on Anandtech last year in October.
Here is the thread.
Apparently, Mr. Terdiman was kind of ticked off at some of the unfair criticisms in that thread leveled towards PhysX, especially PhysX 2.xx and he set the record straight in his blog in a very detailed manner.
He clarifies why PhysX 2.xx never had SIMD support, and why merely running it through a compiler with the SSEn flag ticked would not have been worth it at the time.
It's an interesting summary for those that are curious about such things.. Anyway, it just reinforces what I said in that thread in October. The new focus now for NVidia is on software physics, and PhysX 3.xx is a huge upgrade over 2.xx in terms of speed and features.
The history of PhysX
I came across his blog whilst searching for stuff on the internet, and I was surprised to see him quote a thread that I started on Anandtech last year in October.
Here is the thread.
Apparently, Mr. Terdiman was kind of ticked off at some of the unfair criticisms in that thread leveled towards PhysX, especially PhysX 2.xx and he set the record straight in his blog in a very detailed manner.
He clarifies why PhysX 2.xx never had SIMD support, and why merely running it through a compiler with the SSEn flag ticked would not have been worth it at the time.
It's an interesting summary for those that are curious about such things.. Anyway, it just reinforces what I said in that thread in October. The new focus now for NVidia is on software physics, and PhysX 3.xx is a huge upgrade over 2.xx in terms of speed and features.