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

matthewpomar

macrumors member
Original poster
Oct 27, 2010
81
12
I recently updated my Mac Pro from 138.0.0.0 firmware to 141.0.0.0. Before I did this, I used Unigine Heaven Benchmark 4.0 to get my video card performance before and after the upgrade.

Based on this forum thread, updating the firmware to 140.0.0.0 should increase the PCIe bandwidth from 2.5 GT/s to 5.0 GT/s. I would expect to see better video performance with 5.0 GT/s PCIe bandwidth vs 2.5 GT/s PCIe bandwidth, but the results contradict this:

TEST 1
Mac Pro running Windows 8.1 with MVC Flashed GTX 980 and 138.0.0.0 firmware:

- FPS: 127.5
- Score: 3211

TEST 2
Mac Pro running Windows 8.1 with PC GTX 1070 FE and 138.0.0.0 firmware:

- FPS: 196.9
- Score: 4959

TEST 3
Mac Pro running Windows 8.1 with PC GTX 1070 FE and 141.0.0.0 firmware:

- FPS: 153.8
- Score: 3874

TEST 4
Mac Pro running Windows 10 with PC GTX 1070 FE and 141.0.0.0 firmware:

- FPS: 170.9
- Score: 4304

Note: In Windows 8.1 the NVIDIA Driver version is 419.67 for these tests.

The evidence suggests that upgrading the Mac Pro firmware from 138.0.0.0.0 to 141.0.0.0.0 decreased video performance in Windows 8.1 and Windows 10.

Any thoughts on why I would be seeing worst results with the latest Mac Pro firmware?
 
Last edited:
The other variable is the operating system. In TEST 2 I was running Windows 8.1. In TEST 3 I was running Windows 10. I don't believe that made a difference though.

I DO believe this is the difference. The only way you can tell is by running the EXACT same OS/Software for all 3 tests. Changing up the OS is MAJOR factor in my personal opinion.
 
  • Like
Reactions: bsbeamer
I recently updated my Mac Pro from 138.0.0.0 firmware to 141.0.0.0. Before I did this, I used Unigine Heaven Benchmark 4.0 to get my video card performance before and after the upgrade.

Based on this forum thread, updating the firmware to 140.0.0.0 should increase the PCIe bandwidth from 2.5 GT/s to 5.0 GT/s. However, the benchmark results I got appear to contradict this information:

TEST 1
Mac Pro running Windows 8.1 with MVC Flashed GTX 980 and 138.0.0.0 firmware:

- FPS: 127.5
- Score: 3211

TEST 2
Mac Pro running Windows 8.1 with PC GTX 1070 FE and 138.0.0.0 firmware:

- FPS: 196.9
- Score: 4959

TEST 3
Mac Pro running Windows 10 with PC GTX 1070 FE and 141.0.0.0 firmware:

- FPS: 170.9
- Score: 4304

Based on the benchmark results, it appears updating the Mac Pro firmware from 138.x to 141.x decreased video card throughput. I ran each test 2 times and got nearly identical results, so the results are not anomalous.

The other variable is the operating system. In TEST 2 I was running Windows 8.1. In TEST 3 I was running Windows 10. I don't believe that made a difference though.

Any thoughts on why I would be seeing worst results with the latest Mac Pro firmware?
You can’t just say that you don’t believe that running a different OS doesn’t affect your results.
One could just as easily ask “why am I seeing worse results with Windows 10?”
 
  • Like
Reactions: crjackson2134
I DO believe this is the difference. The only way you can tell is by running the EXACT same OS/Software for all 3 tests. Changing up the OS is MAJOR factor in my personal opinion.

I'll throw the old AHCI drive back in with Windows 8.1 and run the test again. Hopefully I can get it to boot to that again. Once I do, I'll update the stats on my original post.

Thanks.
 
If you want your testing to be conclusive, you should also note the NVIDIA driver version being used. They have been tweaking/modifying their drivers in Windows 10 with GRD and CRD drivers. It makes a difference which one you're using. Also would note your Windows build numbers whenever possible. Windows 10 is almost becoming as broad of a definition as OSX was.
 
  • Like
Reactions: crjackson2134
I'll throw the old AHCI drive back in with Windows 8.1 and run the test again. Hopefully I can get it to boot to that again. Once I do, I'll update the stats on my original post.

Thanks.

Also, understand that as @tsialex stated above... The PCIe connect speed didn’t change between fw v138 and v141. It changes between v0089 and v138.
 
Noted about when the PCIe connect speeds changes based on firmware.

I updated my test results with an apples-to-apples comparison in Windows 8.1 (test 2 and test 3).

It appears going from firmware 138 to 141 has caused video performance to degrade. It doesn't appear to affect Windows 10 as much, but I still don't get the speeds I did when I was running 138.

Thanks for everyone's help.
 
Noted about when the PCIe connect speeds changes based on firmware.

I updated my test results with an apples-to-apples comparison in Windows 8.1 (test 2 and test 3).

It appears going from firmware 138 to 141 has caused video performance to degrade. It doesn't appear to affect Windows 10 as much, but I still don't get the speeds I did when I was running 138.

Thanks for everyone's help.
Lot's and lot's of people have been benchmarking 138/139/140/141.0.0.0.0 and no one noticed such a big difference like you.

People with VEGA GPUs have been benchmarking non stop each firmware update and seen better results with each firmware and macOS update, not worse. Maybe it's a fluke or something else is going on.
 
  • Like
Reactions: crjackson2134
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.