What The Hell Is A Teraflop Anyway?

Megalith

Megalith

24-bit/48kHz
Staff member
Joined
Aug 20, 2006
Messages
13,000
The teraflop seems to be the hot, go-to spec for marketing how powerful a console is, but how much does it actually translate to gaming performance and graphical power? Well, there are certainly other things to look at, such as CPU power.

There is no direct, linear relationship between in-game performance and computational power as measured by teraflops, and while we can get some idea of relative GPU performance, this only really works when comparing different graphics hardware based on the same core architecture. And even then, a GPU with, say, a 40 per cent advantage over another will not see that advantage scale in a linear fashion in terms of pure performance. The capabilities of a graphics processor also rely on more than just computational power too - memory bandwidth in particular is of key importance.
 
Yes and no. While I agree it is FAR from the best or most reliable metric GFLOPS/TFLOPS can be an indicator of how well it performs.

GPUs do a couple things:
1. Memory operations
2. Math operations (ALU)
3. Lastly bus operations with the CPU (handling interrupts, taking request, DMA reads, DMA writes, GPU-GPU communications, sending data back etc. (Small fries unless you have a huge texture memory game or mGPU without a bridge)

Read, do math logic, write. Everything in the GPU/CPU world breaks down this way. But most operations are math operations.

Clever math, like early throw away of primitives, or 2->4 viewport renders in 1 pass, so additional math isn't done are critical also.

Bandwidth is a huge consideration. Things like bus width, memory rate, latency, and compression all affect memory performance.
 
Megalith said:
The teraflop seems to be the hot, go-to spec for marketing how powerful a console is, but how much does it actually translate to gaming performance and graphical power? Well, there are certainly other things to look at, such as CPU power.

There is no direct, linear relationship between in-game performance and computational power as measured by teraflops, and while we can get some idea of relative GPU performance, this only really works when comparing different graphics hardware based on the same core architecture. And even then, a GPU with, say, a 40 per cent advantage over another will not see that advantage scale in a linear fashion in terms of pure performance. The capabilities of a graphics processor also rely on more than just computational power too - memory bandwidth in particular is of key importance.
Click to expand...
It's only the "hot, go-to spec" because it's AMD, and AMD has an historical habit of flaunting their FLOP/s because mainstream publications always pick up and publicize it.
 
You must log in or register to reply here.
Back
Top