The integrated graphics processor (iGP) portion of the next-gen AMD Phoenix mobile APUs will reportedly be able to trade blows with Nvidia's RTX 3060 Max-Q discrete graphics, according to Twitter user and leakster Greymon55. This new rumor kind of makes the earlier estimation of the Phoenix APU performance seem far less impressive where the alleged 24 Compute Unit (CU) graphics engine on Phoenix was tipped to be performing around a mobile GTX 1060.
That's because the GTX 1060 mobile chip has less than half the single-precision compute prowess (FP32) compared to the Max-Q RTX 3060 mobile. The former packs 4.275 TFLOPS of FP32 while the latter comes with 9.846 TFLOPS. Estimates put the 24CU Phoenix APU at around 6.75 TFLOPS of FP32, though it must be kept in mind the TFLOPS aren't directly comparable across architectures in order to determine gaming performance. Phoenix will apparently be 35-45W parts which suggests that its performance against the 60W RTX 3060 Max-Q is mighty impressive, it's still not confirmed whether the iGP will be based on RDNA 2 or RNDA 3.
In terms of when to expect these Phoenix APUs, a leaked AMD roadmap that was apparently shared on Q1 2022 financial results day reveals that Pheonix is slated for landing in 2023. The chips are seemingly being designed for thin and light gaming notebooks.
.jpg)
Meanwhile, AMD also will have another family of Zen 4 mobile APUs codenamed "Dragon Range". These bigger parts will have a 55W+ TDP. However, not much is known so far about Dragon Range APUs.
Source: Greymon55 (Twitter) |Image: Dr. Ian Cutress (Twitter)
_small.jpg)
6 Comments - Add comment