SteamOS

[BlessedSquirrel]

Just noticed those Linux results are hidden, I’ll sort that in a mo, none of them are valid as 3DMark can’t access the Kernel to read the system info

Will include screenshots

I've run them just with all defaults, not changed anything, but you can see it seems to set random resolutions per test which is odd. Best I could manage though.

 

[TonyCarter]

I'd offer some comparisons from the new Mac version of 3DMark (none of the DirectX/RT ones like Firestrike, Time Spy, Port Royal, et al) but at least they're supposedly comparable to the Windows benchmark of the same name...but they're embarrassing

All about 3x lower than my 7800X3D/4090 gaming PC...although using about 4-5x lower power.

 

[BlessedSquirrel]

I'd offer some comparisons from the new Mac version of 3DMark (none of the DirectX/RT ones like Firestrike, Time Spy, Port Royal, et al) but at least they're supposedly comparable to the Windows benchmark of the same name...but they're embarrassing

View attachment 44683

All about 3x lower than my 7800X3D/4090 gaming PC!

I saw that today, was gonna give it a go later. I'm not at all surprised. BUT, it gives Apple some specific metrics to try to get better optimisation perhaps? We can only dream.

 

[TonyCarter]

I saw that today, was gonna give it a go later. I'm not at all surprised. BUT, it gives Apple some specific metrics to try to get better optimisation perhaps? We can only dream.

Well, if it scales properly, and you're only interested in METAL performance, then I suppose for 300W total you could get something with similar performance to my gaming PC.

Unfortunately it will cost £10k

Have just checked, and despite the higher-end M3 Ultra version having double the CPU & GPU cores, it only adds 33-54% to the scores:

• 81084 vs 60843 in Solar Bay (+33%)
• 5519 vs 3577 in Steel Nomad (+54%)
• 19586 vs 14391 in Steel Nomad Lite (+36%)
• 51896 vs 38253 in Wild Life Extreme (+36%)