"But who writes direct to assembly any more?"
Answer, those that grew up flipping the front panel switches on the Dec 11/10 smart terminal computers in the basement of Bolter Hall IIRC at Berkeley. I still have my Cardac cardboard slide device to teach you how to increment the program counter by one and execute an instruction or read the data depending on where you are in the cycle of things. Guys like the late Dave Bargin knew what I'm talking about. As your post alludes to the higher level languages that are abstractions of real machine language are never a efficient as the real thing. With my limited knowledge in the area I would liken it to learning to drive an automatic first. Rarely do those people go back and learn the nuts and bolts of a manual transmission. You might ask why should they? But like the Dead Heads say, "If you gotta ask....."
--- In Avid-L2@yahoogroups.com, Philip Hodgetts <philip@...> wrote:
>
>
> On Jun 10, 2013, at 11:43 AM, Dom Q. Silverio <domqsilverio@...> wrote:
>
> > GPU is not addressed yet. Maybe a hybrid of on board chipset (laptop)? The
> > bare card (no PCI bracket, custom cooling) is installed in factory? No info.
>
> Dual AMD, soldered in. Not replaceable apparently.
> >
> > OpenCL is great but as far as I understand it, the acceleration benefit
> > does not go as deep as platform specific solutionw like CUDA.
>
> OpenCL is not as fast as going direct to the GPU via CUDA, just like writing assembly language is faster than Objective C. But who writes direct to assembly any more? Plus OpenCL is cross all GPU cards, not just NVIDIA.
>
> philip
>
| Reply via web post | Reply to sender | Reply to group | Start a New Topic | Messages in this topic (13) |
No comments:
Post a Comment