In a recent live stream by YouTube channel HotHardware, Intel fellow Tom Petersen discussed the company's new Arc Alchemist graphics clock specification and how it differs from Nvidia's clock specifications for its GPUs.
Intel does not list a base clock on its GPUs, Petersen explained. Instead, Intel has designated its "graphics clock" specification as a general estimation of Arc's clock speeds under the most thermally and power-constrained environments.
This is due to the way Intel's Arc Alchemist GPUs use boosting, allowing the GPUs to push clock speeds as high as possible. As long as there is power, voltage, and temperature headroom to spare the GPU will boost higher than its rated clock speeds, akin to AMD and Nvidia's modern GPU architectures. Petersen says we can expect clock speeds upwards of 2 GHz on Arc under lighter gaming workloads such as Counter-Strike: Global Offensive.
This means most users will never see an Arc GPU running at its exact graphics clock specification and will see clocks well above what is listed. However, it doesn't mean Intel's Arc GPUs will not go below their designated clock speed either in some exceptionally demanding workloads (like Furmark).
In many ways, Intel's new graphics clock specification is similar to Nvidia's Boost Clock and AMD's Game Clock specifications. These are general estimations of clock speed behavior from the GPU when it is pushed close to its limits.
On the flip side, this also means Intel's Arc GPUs don't have a minimum specified frequency at all. For instance, this means Arc GPUs cannot, technically speaking, thermal throttle -- since these GPUs can run at any frequency and still be within the listed specs.
