Beside being completely out of the context of clock speed, workload conditions, and a comparable measure when comparing the triangle rate of different GPUs, armchair analyses always assume mobile GPUs are clocked at a speed that at least remotely approaches the desktop. Nothing is further from the truth.
First generation OpenGL ES 1.1 devices were clocked closer to 50 MHz, with speeds reaching around 83 MHz in popular platforms like OMAP2 and Apple's aggressive 103 MHz in the first iPhone platform. Later iPod Touches and systems outside of the mobile market, like Sega Sammy's Aurora arcade device, would of course push ES 1.1 parts higher.
65 nm enabled ES 2.0 GPUs to clock around 110 MHz - 133 MHz, with Apple once again aggressive at 150 MHz.
With 45 nm and the like, expect around 200 MHz and above and even a high performance mode for tablets and PMPs at 400 MHz, but these ideas of 200 MHz ES 1.1 parts and 600 MHz 2.0 GPUs are just too far from the reality of phones.
First generation OpenGL ES 1.1 devices were clocked closer to 50 MHz, with speeds reaching around 83 MHz in popular platforms like OMAP2 and Apple's aggressive 103 MHz in the first iPhone platform. Later iPod Touches and systems outside of the mobile market, like Sega Sammy's Aurora arcade device, would of course push ES 1.1 parts higher.
65 nm enabled ES 2.0 GPUs to clock around 110 MHz - 133 MHz, with Apple once again aggressive at 150 MHz.
With 45 nm and the like, expect around 200 MHz and above and even a high performance mode for tablets and PMPs at 400 MHz, but these ideas of 200 MHz ES 1.1 parts and 600 MHz 2.0 GPUs are just too far from the reality of phones.