Benchmark results v25000

[Kees499]

 

Imac M4 16GB  10c cpu 10c Gpu

[MikeTO]

I've updated the table and added it to the first post of this thread

https://forum.affinity.serif.com/index.php?/topic/213944-benchmark-results-v25000/

[Komatös]

Here are further measured values of a Mac mini M4 in the 16 GB / 256 GB version.

[MikeTO]

3 hours ago, Komatös said:

Here are further measured values of a Mac mini M4 in the 16 GB / 256 GB version.

10-core CPU and 10-core GPU?

The first number is a bit on the low side, you might try it again.

Thanks

[Komatös]

8 hours ago, MikeTO said:

10-core CPU and 10-core GPU?

Yes! 

I have now run the benchmark again directly after the restart. The single CPU value was 791, but with each subsequent run I get values between 720 and 780, and the other values also fluctuate.
This benchmark is therefore not meaningful for me. Except that the best values are achieved shortly after switching on.

[flar]

AMD Ryzen9 7950x

[Eisbar]

imac 2020, intel i7 3,8 ghz, 40gb ram

[MikeTO]

I've updated the table in the first post of this thread

https://forum.affinity.serif.com/index.php?/topic/213944-benchmark-results-v25000/

 

[z.f.]

This is on a Mac Pro 5.1 2x6 cores @3.06GHz (2x Xeon X5675) and a AMD Radeon Pro VII GPU on MacOS 12.7.6 thanks to OCLP:

And the same mashine running under Windows 10:

[z.f.]

On a MacBook Pro 15" 2015, 2.5GHz (i7 4870HQ,  4 cores) equiped with a second GPU, MacOS 12.7.6:

On a MacBook Pro 13" 2016, 2GHz (i5 6360U, 2 cores), MacOS 12.7.6:

On a MacBook Pro 13" early 2011, 2.3GHz (i5 2415M, 2 cores), MacOS 12.7.6 thanks to OCLP:

 

Edited January 23 by z.f.

[MikeTO]

3 hours ago, z.f. said:

On a MacBook Pro 13" early 2011, 2.3GHz (i5 2415M, 2 cores), MacOS 12.7.6 thanks to OCLP:

For this one, there's no GPU numbers. Was the Intel HD Graphics 3000 disabled by turning off hardware acceleration, or was hardware acceleration on and Affinity is just unable to use this GPU?

Thanks

[z.f.]

9 hours ago, MikeTO said:

For this one, there's no GPU numbers. Was the Intel HD Graphics 3000 disabled by turning off hardware acceleration, or was hardware acceleration on and Affinity is just unable to use this GPU?

Thanks

The GPU is definitely "on" as most graphic effects in MacOS 12 work fluent. But this is due patches of OCLP. The GPU is not reported as Metal in the system report. Affinity seems just not to to use this GPU.

[MikeTO]

I've updated the table in the first post of this thread with all the benchmarks from @z.f. as well as an iPad Air M2.

https://forum.affinity.serif.com/index.php?/topic/213944-benchmark-results-v25000/

Cheers

[Jeff Storms]

Mac Pro (2019)
3.2GHz 16 core Xeon
96GB DDR4
Dual AMD Radeon RX 6900 XT

[z.f.]

On 1/28/2025 at 3:00 AM, MikeTO said:

I've updated the table in the first post of this thread with all the benchmarks from @z.f. as well as an iPad Air M2.

https://forum.affinity.serif.com/index.php?/topic/213944-benchmark-results-v25000/

Cheers

You put the wrong GPU in the table for the Macbook Pro early 2011. Its a Intel HD Graphics 3000.

[MikeTO]

2 hours ago, z.f. said:

You put the wrong GPU in the table for the Macbook Pro early 2011. Its a Intel HD Graphics 3000.

Thanks for the catch, I've fixed it.

[Jeff Storms]

I have several Mac Pro's so I wanted to see how well my old 4,1/5,1 did compared to my 7,1. Not bad for a 14 year old computer.

Mac Pro 4,1/5,1 2009
Dual X5690 @ 3.46GHz 12 cores
96GB 1333MHz DDR3
Saphire Pulse RX 580 8GB

Edited January 31 by Jeff Storms

[hed]

Windows 11 Pro, Ryzen 7 9700X, NVIDIA GeForce RTX4060 Ti 16 GB, 32 GB RAM DDR5

[MikeTO]

4 hours ago, hed said:

Windows 11 Pro, Ryzen 7 9700X, NVIDIA GeForce RTX4060 Ti 16 GB, 32 GB RAM DDR5

What is the second GPU, the AMD Radeon model?

Thanks

[MikeTO]

There's something curious about this newest benchmark report and a couple of the other existing benchmarks. Why would the Raster Multi GPU score ever be lower than the Raster Single GPU?

[hed]

9 hours ago, MikeTO said:

What is the second GPU, the AMD Radeon model?

Thanks

It is not dedicated GPU, it is integrated with CPU. AMD does not specify model, they describe it generally as "AMD Radeon Graphics" (https://www.amd.com/en/products/processors/desktops/ryzen/9000-series/amd-ryzen-7-9700x.html).

[z.f.]

On 2/14/2025 at 1:51 AM, MikeTO said:

There's something curious about this newest benchmark report and a couple of the other existing benchmarks. Why would the Raster Multi GPU score ever be lower than the Raster Single GPU?

That seems only to happen if the second GPU is less powerful, as the MacPro 7.1 benchmark with two equal GPUs suggests. I suppose it is some problem of load balancing. The fast GPU is then finished and the slow one is still doing something. But finding the precise reason will most likely not be possible without some access to the code and debugging.
There are more strange things. For the combined GPU test, the other MacPro 5.1 benchmark with the RX580 is much faster than the Radeon VII, although I see no single point where the RX580 outperforms the Radeon VII in the specs.

 

[Jeff Storms]

9 hours ago, z.f. said:

That seems only to happen if the second GPU is less powerful, as the MacPro 7.1 benchmark with two equal GPUs suggests. I suppose it is some problem of load balancing. The fast GPU is then finished and the slow one is still doing something. But finding the precise reason will most likely not be possible without some access to the code and debugging.
There are more strange things. For the combined GPU test, the other MacPro 5.1 benchmark with the RX580 is much faster than the Radeon VII, although I see no single point where the RX580 outperforms the Radeon VII in the specs.

 

My dual CPU setup is faster. I have 2 x X5690 @ 3.46GHz whereas the other 5,1 has 2 x X5675 @ 3.06GHz.

You can see the Vega outperformed the RX580 in Raster but when testing the CPU alone and the combined CPU/GPU, my system is faster. I also have 96GB DDR3 @ 1333MHz which raises my combined score also. I don't know how much RAM the other 5,1 has but what I do know is that if it's maxed out at 128GB it incurs a performance penalty because the 5,1 only has a triple channel memory controller. So by only filling 6 of the 8 available memory slots I gain multi-core performance.

My Geekbench scores for the 5,1 will be better also. My Geekbench scores for the 7.1 are also more than double the M4 max because of the dual RX 6900's. As you can see from my raster benchmark score with the Dual RX 6900's are almost double the M4 Max.

That's why I love MacPros, you can upgrade them and tweak them to be performance monsters.

My Geekbench score for the 7,1 with the dual 6900's.
 

 

Edited February 16 by Jeff Storms
Recent GeekBench

[z.f.]

I reproduced similar RX580ish results, actually a RX590, on my 5.1. That was to be expected, a nearly double GPU bench performance at "combined" should not come from a 10% faster CPU.
In my real applications the Radeon VII clearly outperforms the RX590, frequently by a factor of 3, or so, with heavy computations.

 

[Jeff Storms]

On 2/16/2025 at 2:59 AM, z.f. said:

I reproduced similar RX580ish results, actually a RX590, on my 5.1. That was to be expected, a nearly double GPU bench performance at "combined" should not come from a 10% faster CPU.
In my real applications the Radeon VII clearly outperforms the RX590, frequently by a factor of 3, or so, with heavy computations.

 

All I can say is everything else appears as it should in a side by side comparison.