Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

FToWnPlAyA

macrumors newbie
Original poster
Sep 12, 2012
25
5

Attachments

  • IMG_0001.PNG
    IMG_0001.PNG
    624.9 KB · Views: 412
  • IMG_0002.PNG
    IMG_0002.PNG
    61.1 KB · Views: 90
  • IMG_0003.PNG
    IMG_0003.PNG
    66.1 KB · Views: 93
Last edited:
I tried this with 3D Mark.

Left iPhone 5S and right iPhone 5

It doesn't make sense for the physics test to regress. I wonder if 3D Mark hasn't been optimized for this yet.

I like how on the iPhone test, it said it is one of the most powerful devices and yet it said the same thing on the iPhone 5S when it is 3 times as powerful.
 
What about those 2 points where the iPhone 5 wins? Weird

Physics is usually CPU. The test may not be optimized for the stuff that A7 can do. However, Anandtech's benchmark does show that A7 is not that much improved over physics.
 
Physics is usually CPU. The test may not be optimized for the stuff that A7 can do. However, Anandtech's benchmark does show that A7 is not that much improved over physics.
Actually, physics is GPU. All modern GPUs support hardware accelerated physics.

The iPhone 5S uses a new GPU architecture from Imagination Technologies and drivers for that architecture are very new / immature. They are essentially v1.0 drivers, and they provide very limited physics acceleration. Future driver releases, which Apple would incorporate into 7.0.X or 7.X, should greatly improve physics performance.

Edit: The 3DMark physics test disables GPU physics acceleration. Thus, Mikhail's post is correct.
 
Last edited:
Actually, physics is GPU. All modern GPUs support hardware accelerated physics.

The iPhone 5S uses a new GPU architecture from Imagination Technologies and drivers for that architecture are very new / immature. They are essentially v1.0 drivers, and they provide very limited physics acceleration. Future driver releases, which Apple would incorporate into 7.0.X or 7.X, should greatly improve physics performance.

Unless 3d mark has recently changed their benchmark suite, THIS physics test is CPU bound. If it's not, it would be the first time a 3d mark physics test was not using strictly the CPU in the nearly 15 years I've been using 3d mark.
 
Unless 3d mark has recently changed their benchmark suite, THIS physics test is CPU bound. If it's not, it would be the first time a 3d mark physics test was not using strictly the CPU in the nearly 15 years I've been using 3d mark.
You are absolutely right. I stand corrected.

Futuremark.com:

"The Merry Go Round Air Show continues in this heavy physics stress test for your CPU. This time, the smoke-trailing planes collide with various cloth and soft-body obstacles, each other, and the ground. The smoke spreads and reacts to the planes passing through it. As this test is designed to stress the CPU physics acceleration on the GPU is disabled by default."
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.