News:

Willkommen im Notebookcheck.com Forum! Hier können sie über alle unsere Artikel und allgemein über Notebook relevante Dinge disuktieren. Viel Spass!

Main Menu

The new 3DMark Wild Life cross-platform benchmark is reporting some strange results where Apple's A13 Bionic GPUs beat the latest AMD Vega 8 iGPUs

Started by Redaktion, October 28, 2020, 16:37:21

Previous topic - Next topic

Redaktion

PCWorld tested the Apple A13 Bionic SoC against a few popular laptop processors, and it turns out that the iPhone 11 chip appears to be faster than iGPUs like the Iris Plus and the Vega 8, but it is also significantly lagging behind the new Tiger Lake Iris Xe iGPUs. Are we even supposed to mix mobile GPUs with laptop iGPUs?

https://www.notebookcheck.net/The-new-3DMark-Wild-Life-cross-platform-benchmark-is-reporting-some-strange-results-where-Apple-s-A13-Bionic-GPUs-beat-the-latest-AMD-Vega-8-iGPUs.499667.0.html

bq

PCWorld comment is really nonsensical. Wild Life is the first "real" cross platform benchmark — doing exactly the same work across different platforms and APIs. It should be treated as any there synthetic benchmark, keeping in mind it's intended use case and limitations. Real games obviously might perform better — because they are doing many other things at the same time and because their utilization of the GPU is different. On the other hand, if you just want to compare raw GPU performance for a common run-on-the-mill deferred shading scenario, Wild Life is exactly what you need.

But PCWorld's  conclusion that "you should not compare an iPhone GPU to a laptop GPU because there is no way that an iPhone is faster" is void of any logic.

mfaisalkemal

iphone gpu start from A11 is double FP16 performance (FP32:FP16 = 1:2) and simultaneously double texture fillrate. if we compare with same core (256 FP32 core) tegra x1 only has 16 tmu vs 32 tmu on A12, A13 and A14. a13 GPU fp32 GFLOPs around 660GFLOPs. after compare A10 and A13 score on 3dmark wild life, i think 3dmark wild life shader code utilize 90% FP16 code and 10% FP16 code so speed up around 1.8x vs desktop 100% FP32 shader code.

Quick Reply

Warning: this topic has not been posted in for at least 120 days.
Unless you're sure you want to reply, please consider starting a new topic.

Name:
Email:
Verification:
Please leave this box empty:

Shortcuts: ALT+S post or ALT+P preview