- Joined
- Apr 19, 2011
- Messages
- 2,198 (0.44/day)
- Location
- So. Cal.
i seriously doubt that for the forseeable future, processor integrated graphics can be faster than a low-end card.
igps will get better, but so will discrete gpus. both are using the same technology, architecture and production process. heat density limits will always limit IGP performance.
i see the big advantage in massive cost savings. why should a 250 mm² silicon die with cpu+gpu cost significantly more than a 250 mm² silicon die with just a cpu?
Working from the hypostasis... that it's about gaming or fast CPU computations is the misnomer.
First will AMD still be developing/offering IGP motherboards? Will they necessarily offer low-end discrete cards that are intended to compete with their APU line. While could Nvidia make any money offering what will become low-end discrete (approx GT430) that betters APU performance (that will improve)?
IGP’s will get better?... I think that boat sailed. Nvidia would need to get serious at building a competitive IGP mobo and then price that package competitively; it wasn’t happening before this, why would the picture be any rosier going forward? Intel is behind offering less competitive (siamese) on-chip graphic’s. It appears they have yet to find that path over at least the next two years to offer anything more than GMA. They could catch-up; although they’ll need a (new) graphic architecture, which they’d need to totally integrate that onto an ever smaller die to be a challenger.
Is it priced competitively… no, as that would hurt only AMD. They’ll need to work down stock of CPU’s and IGP motherboards before they start really turning up the heat in the aftermarket, but you can bet they are hyper-aggressively working the OEM’s. If they hit the 250 mm² and have the process/foundry in full swing before Intel get on the tracks the low end computer market will be AMD.
Last edited: