Intel’s Ivy Bridge HD4000 Overclocked by 40%

Intel could have offered us a much better iGPU in the Ivy Bridge. The die size was small enough to allow something like this and the architecture is so simple now that doubling the functional units would have resulted in literally double the performance.

Intel is famous for not caring about the client's needs and for concentrating mostly on fractionizing any performance improvements it offers in many small increments, for which the company will charge “an arm and a leg.”

If it can get away with such lackluster performance, why would it diminish profits with more than 10% to increase the customer’s benefit by 100%? It’s likely that if it were possible, the company would charge 100% profit for just 10% consumer benefit.

Apple wanted Intel to offer it a truly powerful iGPU, but Intel couldn’t be bothered to build such a CPU for only some OEMs.

Having just this much to play with, Charlie’s colleague over at semmiacurate.com, Thomas Ryan, put Ivy Bridge’s iGPU on the test.

Using Intel’s own CPU and GPU monitoring tool, he managed to overclock the iGPU to a sky-high 1,600 MHz.

The voltage increase was substantial and it's probably not a safe bet to keep the system running at these frequencies 24/7, but it’s doable.

You can also overclock without any additional voltage. The iGPU managed to reach 1,300 MHz using stock voltage. That’s a healthy increase of 13% on stock voltage, but with better cooling.

Unfortunately, increasing the voltage is not enough. The CPU has a TDP cap and if you don’t modify it upwards, it will likely throttle the iGPU or CPU, and the obtained results will be lower.

After modifying this cap, the 1,600 clocked iGPU managed to offer an average of 15% performance improvement, that’s quite disappointing when you think that the frequency increase was almost 40%.

Posted in: Hardware, Preview