Intel caught in graphics shocker
G55 actually won't suck
IF YOU ARE not already sitting down, do so now, because this news is not for the weak or easily scared. That said, the new graphics in Intel's upcoming CPUs are not going to suck.
You may be familiar with the Intel integrated graphics saga, a tale of woe and dread rivaling a romance novel with the last few pages cut out. It all started with the G965, a chip that basically never worked right. Its successor, the G35, also never worked right, and the G45... well, to be charitable, came much closer to working. It was never considered fast, or even anything more than barely usable for office work, but it wasn't totally embarrassing. The current score, zero for three.
So, how many tries does it take for Intel to get graphics right? Four. No, really, I am not joking. The word from early testing is that Intel's integrated graphics - essentially a G55 integrated into the CPU package, not the chipset - are not going to suck this time around. They not only got it right, but gave it the horsepower to, well, stand proudly above those that don't offer bare minimum functionality.
The open question now is not raw horsepower, but drivers. Intel has been saying the drivers would function right for three-plus years now, but it never quite hit the mark. Given the sustained effort it is putting in, they just might get there. The lack of major architectural changes to the GPU also is a potential positive.
In the end, after three or four years of graphics that would not cut it on demanding games like The Sims, Intel may be stepping out of the dark. Its graphics will never run Crysis at full speed, but this time, I am hopeful that they will work, and do so at acceptable speeds. Believe it or not, this is real progress. µ
G55 actually won't suck
IF YOU ARE not already sitting down, do so now, because this news is not for the weak or easily scared. That said, the new graphics in Intel's upcoming CPUs are not going to suck.
You may be familiar with the Intel integrated graphics saga, a tale of woe and dread rivaling a romance novel with the last few pages cut out. It all started with the G965, a chip that basically never worked right. Its successor, the G35, also never worked right, and the G45... well, to be charitable, came much closer to working. It was never considered fast, or even anything more than barely usable for office work, but it wasn't totally embarrassing. The current score, zero for three.
So, how many tries does it take for Intel to get graphics right? Four. No, really, I am not joking. The word from early testing is that Intel's integrated graphics - essentially a G55 integrated into the CPU package, not the chipset - are not going to suck this time around. They not only got it right, but gave it the horsepower to, well, stand proudly above those that don't offer bare minimum functionality.
The open question now is not raw horsepower, but drivers. Intel has been saying the drivers would function right for three-plus years now, but it never quite hit the mark. Given the sustained effort it is putting in, they just might get there. The lack of major architectural changes to the GPU also is a potential positive.
In the end, after three or four years of graphics that would not cut it on demanding games like The Sims, Intel may be stepping out of the dark. Its graphics will never run Crysis at full speed, but this time, I am hopeful that they will work, and do so at acceptable speeds. Believe it or not, this is real progress. µ
