CPU vs gpu clock speed
- Thread starter beany_bot
- Start date
The differences are a lot greater than simply clock speed. The GPU is designed to do a few things well, whereas the CPU can do a lot more.
This analogy is quite good, taken from here
One way to visualize it is a CPU works like a small group of very smart people who can quickly do any task given to them. A GPU is a large group of relatively dumb people who aren't individually very fast or smart, but who can be trained to do repetitive tasks, and collectively can be more productive just due to the sheer number of people.
This analogy is quite good, taken from here
One way to visualize it is a CPU works like a small group of very smart people who can quickly do any task given to them. A GPU is a large group of relatively dumb people who aren't individually very fast or smart, but who can be trained to do repetitive tasks, and collectively can be more productive just due to the sheer number of people.
Very nice explanation. Thanks. It does raise another question in my head though... If CPU's are so much "better" than GPUs. Why are GPU's more expensive?
With a gpu its not just the chip you are buying its all the circuitry round it too and that all costs development. Gpus sre also more complex.
Factor in nvidia for instance is relatively small compared to Intel and costs then scale also.
Factor in nvidia for instance is relatively small compared to Intel and costs then scale also.
Soldato
- Joined
- 27 Dec 2009
- Posts
- 2,727
- Location
- Gillingham, Kent
As Varkanoid said, when you buy a CPU you're just buying the chip itself. When you buy a GPU, you're in effect buying the chip, memory, cooler and motherboard all in one.
think of manufacturing cost much to do with of Die size, and transliterates.
the i7 skylake is 122mm^2 - maybe 1.5-1.6 billion transistors (cant find a source)
http://www.anandtech.com/show/9582/intel-skylake-mobile-desktop-launch-architecture-analysis
the nivdia 980ti is ~600mm^2 and 8+ million transistors
these transistors all need to be designed and go through many complex stages of simulation before you get a shiny new gpu. remember most designs take over 2-3 year process and we only get to see/here them towards the last 6 months.
even now Nvidia/amd will be starting internal planning for at least the next 2 gpu/cpu generations
Dan
the i7 skylake is 122mm^2 - maybe 1.5-1.6 billion transistors (cant find a source)
http://www.anandtech.com/show/9582/intel-skylake-mobile-desktop-launch-architecture-analysis
the nivdia 980ti is ~600mm^2 and 8+ million transistors
these transistors all need to be designed and go through many complex stages of simulation before you get a shiny new gpu. remember most designs take over 2-3 year process and we only get to see/here them towards the last 6 months.
even now Nvidia/amd will be starting internal planning for at least the next 2 gpu/cpu generations
Dan
Last edited:
Associate
- Joined
- 31 Oct 2012
- Posts
- 2,243
- Location
- Edinburgh
Quite the reverse, GPUs are massively more powerful. CPUs are easy to program for as most tasks are easily broken down into a series of steps, GPUs are very hard to program for as to make good use of them you need many things to be worked on at once.
CPUs are as a result easy to use for a broad range of tasks, while GPUs only excel at some - but when they do they are orders of magnitude faster. Many projects aim to rewrite computationally intensive problems to run on GPUs instead.
Forget clock speed, it's literally just a number. You can't compare two different products by clock speed unless they are the same in every other way. So two different AMD GPUs can't be compared with it, nor can they be compared with any NVIDIA GPUs nor any CPUs etc. Two AMD 390s can be compared, though even then there are other factors like power limit, cooling, vRAM speed, vRAM timings beyond just core clock speed (without even thinking about the rest of the system they're sitting in).
Last edited: