Intel vs Nvidia

Practical and theoretical implementation discussion.
Post Reply
Chris Morris
Posts: 6
Joined: Sun Mar 18, 2012 5:25 am

Intel vs Nvidia

Post by Chris Morris » Fri Mar 07, 2014 3:48 am

After building out nearly every whitepaper I can get my hands on, I don't see a clear winner CPU vs GPU looking forward. GPUs are generally faster right now in the simpler cases, but CPUs might scale better to complex scenes/materials looking 5 years ahead. So scrapping the CPU vs GPU argument, I have to look compare the companies that make them.

Who do you have more confidence in looking ~5 years down the road from now, Intel or Nvidia?

Posts: 206
Joined: Fri Dec 02, 2011 8:00 am

Re: Intel vs Nvidia

Post by Dade » Fri Mar 07, 2014 10:47 am

Chris Morris wrote:Who do you have more confidence in looking ~5 years down the road from now, Intel or Nvidia?
It seems the perfect topic for a flame war :D

"No one has been ever fired for having bought Microsoft" or Intel, IBM, etc. If you just compare the size of the companies, there seem to be no contest.

However, in my opinion, you can not scrap the CPU vs GPU argument, it is all about the the CPU vs GPU argument. Just compare the scaling of CPU and GPU performances in last years.

Posts: 56
Joined: Wed Dec 21, 2011 6:08 pm

Re: Intel vs Nvidia

Post by hobold » Fri Mar 07, 2014 10:53 am

Both processor types will continue to converge somewhat. CPUs will continue to evolve more powerful SIMD ALUs (though not necessarily wider ones), and GPUs will further improve in terms of ALU utilization for a larger set of less regular algorithms. They will never ever meet in a single point, because CPUs won't compromise single threaded performance of spaghetti code, and neither will GPUs compromise realtime graphics throughput.

On an even playing field, with identical silicon fabbing capability and comparable engineering prowess, no type of processor would ever clearly win. Each machine would continue to have its strengths, where it cannot be touched by another kind of processor.

So, IMHO, if there ever turns out a clear winner, it won't be for technical reasons, but because of economical constraints. I find it difficult to call a winner yet, as all players, Nvidia, Intel, and AMD, are struggling due to the shrinking PC market. I don't think they can continue feeding all their chip designers indefinitely, even if they manage to participate in the mobile gadget revolution. ARM reigns supreme there, and their processor cores are tuned for low cost and power efficiency, not for performance. Absolute performance is just not a selling criterion in that space, as long as it is "high enough".

Posts: 146
Joined: Fri Mar 02, 2012 7:01 pm

Re: Intel vs Nvidia

Post by Geri » Sat Mar 15, 2014 3:40 am

integration is the key. and the fact that 10 million of existing program utilizes cpu, and there is only like 500 gpgpu capable software, from the 500, in reality, maybe 2 is working, becouse they was not even able to make it to work in the last 7 years, alreday declares a winner in fact. imho gpu have no future.

Post Reply