Arc Alchemist GPUs are finally out in the open, but it quickly became apparent that something was up with its low clock speeds. Intel’s Tom Peterson cleared the air during an interview with HotHardware, explaining the differences between Intel’s clock speeds and competitors like Nvidia and AMD.
Instead of offering up a base clock speed, Intel opts for a graphics clock, which functions differently than what many consumers are used to. Unlike a base clock, the graphics clock is the average clock speed in a “TDP-constrained worst case.” This translates to the lowest clock speed users should experience in an application, which leaves a lot of unknown headroom for the best-case scenario.
Peterson used Intel’s A370M as an example. He noted the graphics clock of 1,550MHz and its relation to the TDP range of 35 to 50 watts. The relationship here is that the graphics clock of 1,550MHz is the average clock speed in a low power envelope of 35W in a standard application. Peterson went on to note that in lighter applications like CS:GO, those clock speeds will likely increase beyond 2,000MHz.
In several benchmarks by Benchmark Lab and Digital Testings via KitGuru, Intel’s A370M doesn’t quite hit the speeds Peterson talks about in his interview with HotHardware. In fact, when testing the A370M in CS:GO, the GPU doesn’t even crack its graphics clock. On the other hand, a GPU-Z screenshot obtained by VideoCardz via KitGuru places the A350M at 2,200MHz. This seems more in line with what Peterson claims in his interview and was corroborated by VideoCardz after reaching out to GPU-Z’s developer. Still, it’s a mixed bag in terms of early testing.
Intel launched its Arc Alchemist mobile line of GPUs on March 30 during which it released a teaser for a limited edition desktop GPU. Discrete cards are due out sometime in Q2 2022, while Intel’s mobile solutions will be available in partner devices from ASUS, MSI, Acer, and more.