Like the RX 580 has 8gb of VRAM, but has similar performance to a GTX 970 with 3.5/4gb of VRAM. Or like their processors with a higher core count that doesn’t seem to mean much when compared to Intel CPUs with the same core count.
In: Technology
VRAM doesn’t tell you how powerful a GPU is, technically you could have one with literally just 200mb that is faster than one with 50GB, in certain applications.
In videogames VRAM is more like a yes/no thing, if all of the games textures etc can fit within VRAM you’re fine, if they can’t, things will be slow, but the actual calculation is done by the GPU which isn’t as marketable.
In terms of CPU cores AMD has improved massively in recent years, today 1 AMD core is about as good as 1 Intel core, however with older generations like Bulldozer AMD made the bet to have a more cores that are individually weaker, for example in the Bulldozer Architecture the cores even shared a few componens, so if you were to do floating point arithmetic you would see similar performance to a 4 core even though you had 8.
It’s the difference between throwing more at the problem and optimizing what you have.
AMD’s strategy has been to add more cores, and Intel’s has been to optimize the work done by the cores it has. To use a rough racing analogy, AMD is like a NASCAR car, they just beef the engine and go to town. Intel is more like an F1 car, more care has been taken in the overall characteristics.
Because that’s just one number. A semi truck has more wheels than a sports car, why doesn’t it perform like it has more?
That’s the easiest number to measure, but it’s also the easiest number to make really big, so it wouldn’t surprise me if they focused on that number on purpose.
Adding more wheels to a car does not make it go faster.
Latest Answers