Vega Pro 56 vs 64 - Performance difference w/benchmarks

[Mac32]

Found some interesting numbers regarding Vega 56 vs 64 performance (w/8 core CPU), taken from these sources:

https://forums.macrumors.com/threads/bootcampdrivers-now-works-on-imac-pro.2101377/#post-25753668

3DMark Timespy
Vega Pro 56: 5797
Vega Pro 64: 6273
Performance improvement over Vega Pro 56: 8.2%

3DMark Firestrike
Vega Pro 56: 15603
Vega Pro 64: 16654
Performance improvement over Vega Pro 56: 6.7%

AFAIK these numbers are with Apple's fairly crappy Bootcamp graphics drivers. If this is indicative of these GPUs' performance, the difference is smaller than the desktop versions (around 14-15%). Maybe using a custom fan profile with Mac Fan Control would reduce throttling, and give Vega Pro 64 more performance headroom?

These are only gaming benchmarks, not professional software where the additional shading units and VRAM would have a bigger impact. This is the most specific benchmark comparison I've found so far.

Anyone with different results?

 

[ESA]

Found some interesting numbers regarding Vega 56 vs 64 performance (w/8 core CPU), taken from these sources:

https://forums.macrumors.com/threads/bootcampdrivers-now-works-on-imac-pro.2101377/#post-25753668

3DMark Timespy
Vega Pro 56: 5797
Vega Pro 64: 6273
Performance improvement over Vega Pro 56: 8.2%

3DMark Firestrike
Vega Pro 56: 15603
Vega Pro 64: 16654
Performance improvement over Vega Pro 56: 6.7%

AFAIK these numbers are with Apple's fairly crappy Bootcamp graphics drivers. If this is indicative of these GPUs' performance, the difference is smaller than the desktop versions (around 14-15%). Maybe using a custom fan profile with Mac Fan Control would reduce throttling, and give Vega Pro 64 more performance headroom?

These are only gaming benchmarks, not professional software where the additional shading units and VRAM would have a bigger impact. This is the most specific benchmark comparison I've found so far.

Anyone with different results?

Thats seems right..

But I don´t think the scores will be higher.

But the performance in games is a HUGE difference with REAL drivers.

It´s lika a GTX 1080 in performance in games.

 

[Mac32]

Good to know, will test my iMac Pro w/Vega 56 when I get it.

 

[anticipate]

Going from memory, All testing (barefeats) etc has shown at most a 18% speed increase at best, except for one anomalous 20% rating in Compressor. Under 10 is typical. Not worth the $600 to me, especially when the base is 30+% faster than the 2017 RX 580 - itself a relatively fast card.

This jives with desktop Vega card ratings.. the 64 is only marginally faster.

 

[sputnikBA]

Rather than the speeds of the cards, I’d really like to track down someone who uses the Vegas in a field where the 8GB vs 16GB VRAM is influential.

I’m very interested in what kinds of workloads those are and exactly what difference that makes, but unfortunately no reviews or user impressions seem to be analysing that specifically.

 

[BigJohno]

You could save 600 and down the road take the 600 and install whatever video card is supported through Thunderbolt.