AMD Announces The Radeon Pro VII

[tomtomme]

Making a pro version of an already failed product which is also on the verge of being artificially moved to artificial legacy state due to being a too old architecture.

What could go wrong? Poor Volta I guess.

for gpgpu / workstation workloads this card is very capable. only the software stack makes it suitable only for a niche. the architecture albeit is very good for compute / pro customers. it only fails for gaming.

[nranger]

Making a pro version of an already failed product which is also on the verge of being artificially moved to artificial legacy state due to being a too old architecture.

What could go wrong? Poor Volta I guess.

Radeon VII was a poor gaming product, but for compute it was awesome. TFLOPS per die area was very impressive. As the last generation of desktop GCN, the architecture had bottlenecks in geometry processing and instruction scheduling when shader instructions couldn't fill a wave in the pipeline. But those issues go away when you're not trying to draw triangles or pixels on a screen.

There's a reason GCN is evolving in to "CDNA". Keep the big horsepower for double, half, and quarter precision in the HPC/Workstation products, go even bigger on compute/die area density. And hopefully that lineage means that AMD can continue to improve the software, since they are SO far behind CUDA as an ecosystem.

[eydee]

Being a good architecture for compute means nothing if AMD says the opposite and doesn't provide the drivers. Has everyone forgotten Terascale already? Even after a decade, OGL4 support is still blocked because AMD says so. Being open source changes nothing, AMD controls that part of Mesa. GCN has been out for more than what Terascale ever got. It can be shut down tomorrow. Even today. Just to boost RDNA sales. They don't even need a valid reason.

[tomtomme]

Being a good architecture for compute means nothing if AMD says the opposite and doesn't provide the drivers. Has everyone forgotten Terascale already? Even after a decade, OGL4 support is still blocked because AMD says so. Being open source changes nothing, AMD controls that part of Mesa. GCN has been out for more than what Terascale ever got. It can be shut down tomorrow. Even today. Just to boost RDNA sales. They don't even need a valid reaso.

There are no professional compute products with RDNA and there won't be, because RDNA is bad at comput. AMD does _not_ say the opposite. Shutting down GCN would mean to abandon the compute market.

[bridgman]

FWIW, nobody has ever said that RDNA is "bad at compute"... it's just that CDNA offers a compute-per-unit-area advantage over RDNA (and the competition IIRC), which is particularly important for datacenter compute.

[bridgman]

Making a pro version of an already failed product which is also on the verge of being artificially moved to artificial legacy state due to being a too old architecture.

With respect, what the heck does this even mean ?

Being a good architecture for compute means nothing if AMD says the opposite and doesn't provide the drivers. Has everyone forgotten Terascale already? Even after a decade, OGL4 support is still blocked because AMD says so. Being open source changes nothing, AMD controls that part of Mesa. GCN has been out for more than what Terascale ever got. It can be shut down tomorrow. Even today. Just to boost RDNA sales. They don't even need a valid reason.

Maybe I'm missing something here, but Terascale parts have been supporting GL 4.5 for a while (see MesaMatrix). Can you be more specific ?

What do you think we have been blocking via controlling "that part of Mesa" ?

How can we "shut down tomorrow" an upstream open source driver ? I didn't think any company was that powerful.

[deusexmachina]

How will it allow for multi-GPU (using passthrough) -> single display?

[bridgman]

How will it allow for multi-GPU (using passthrough) -> single display?

Do you mean Crossfire or something different ?

The links are general purpose - they just give each GPU the ability to access HBM on a linked GPU with very high performance - although I expect they will be used for compute at least as much as for graphics.

[coder]

Speaking of which, is that the only difference between this and the Radeon VII?

Radeon VII:

• is 300 W (this is 250)
• runs at higher clocks (boosts to 1750 MHz instead of 1700 MHz)
• has only PCIe 3.0 (instead of 4.0)
• has half of the fp64 performance
• has no over-the-top Infinity Link connector for direct card-to-card communication.
• costs $700 (this lists for $1900)

Also... wasn't the Radeon VII basically just a binned workstation GPU? If so, does that mean this is a bin of a bin?

Before now, you could only get Vega 20 in:

• Mac-specific Radeon Pro II & Pro II Duo
• MI50 & MI60 (which lack any display outputs)

So, this is the first workstation card to feature it.

That's not to say it wasn't binned at all, since it has only 60 CUs, yet the MI60 and Radeon Pro II have 64.

[Geopirate]

No, PRO cards don't support SR-IOV.

Wait, none of them? So AMD hasn't released a card with SR_IOV for 4 years?