Transfer speed on the hardware won't be a problem. We'll probably have faster SD memory cards than console SSDs by the middle of next gen! It's a matter of latency and access that the PC needs to work around.
No CPU has sufficient PCIe lanes to allow a 16x GPU + either and 8x or 16x SSD. The best you could get would be an 8x SSD but then you'd be halving the bandwidth to your GPU.
I mean, surely to god MS has been anticipating this for a while now right? They've got to know that they can't just sit on this stuff for forever. From the very conception of the Series X, they've had to know that super fast storage would be a defining factor for next gen games, and that they were becoming I/O bound. Right then would have been the time to put the gears in motion to do whatever they could to improve I/O overhead and latency on PC.Transfer speed on the hardware won't be a problem. We'll probably have faster SD memory cards than console SSDs by the middle of next gen! It's a matter of latency and access that the PC needs to work around.
One of the "leaks" on YouTube said Nvidia is working on a RAM/SSD caching system. I wouldn't put it past them to put dedicated hardware on the GPU and use their MDF to drive adoption. Also, if studios are going to standardize on Kraken/DirectStorage then GPU companies would be smart to follow.
I believe that DirectStorage, and future Windows 10 optimizations will go a long way to improving the situation on PC, as well as Nvidia/AMD drivers/technology.
What? 24/32 core Zen 2Threadripper has 64 PCIe 4.0 lanes and the 64 core model has 88 lanes. All the Zen1 Threadrippers have 64 PCIe 3.0 lanes. Skylake X models have 44 lanes.No CPU has sufficient PCIe lanes to allow a 16x GPU + either and 8x or 16x SSD. The best you could get would be an 8x SSD but then you'd be halving the bandwidth to your GPU.
What? 24/32 core Zen 2Threadripper has 64 PCIe 4.0 lanes and the 64 core model has 88 lanes. All the Zen1 Threadrippers have 64 PCIe 3.0 lanes. Skylake X models have 44 lanes.
Even the consumer Zen2 models have a total 24 PCIe 4.0 lanes (only 4 dedicated to the chipset), which is the equivalent of 48 PCIe 3.0 lanes.
With a Zen2 + X570, one could have a 16x PCIe 4.0 GPU and a 4x NVMe PCIe 4.0, effectively getting a theoretical storage transfer speed that is higher than the PS5's, before data decompression.
Nvidia already accelerates storage access in their DGX workstation and it's called GPUDirect Storage:One of the "leaks" on YouTube said Nvidia is working on a RAM/SSD caching system. I wouldn't put it past them to put dedicated hardware on the GPU and use their MDF to drive adoption. Also, if studios are going to standardize on Kraken/DirectStorage then GPU companies would be smart to follow.
A "gaming" version will eventually come to mainstream PC if Nvidia thinks it will be necessary...
I think both AMD and Nvidia understand the necessity of directly accessing fast storage from the GPU, now more than ever. I would not be surprised to see both Ampere and RDNA2 offer the functionality. Surely MS is working with both of them. Sony and AMD are also working together and technologies designed for PS5 will be coming to PC with their RDNA2 gpus. Cerny made a point to explain that it wasn't just Sony picking AMD technologies, but technologies they've developed together would be making their way to PC as well.And that's the thing i think, is it necessary for the coming years? NV (and AMD) probably know more, but atm it doesn't seem a top priority. Prices will be higher too, but on the other hand seeing what people are happily paying for RTX products..... It has become the Apple of GPU's it seems. I remember buying a Ti500 for around 400/500 dollars (i think), which backthen was considered the most expensive GPU price. Now that gets you abit over mid-range with high end going over a 1000 dollars.
You have any idea why some outlets got the idea epic noted that not even sata ssds would be sufficient and you'd need an nvme ssd? Was that a misunderstanding, or was there reason to believe that?Yeah that was precisely how it was - I listened to a recording of the meeting that DF and other outlets had representatives at! IIRC it was the PC Gamer representative who asked questions about the nature of the PC (RTX 2070 was the example used) that could run it.
Nvidia already accelerates storage access in their DGX workstation and it's called GPUDirect Storage:
https://devblogs.nvidia.com/gpudirect-storage/
A "gaming" version will eventually come to mainstream PC if Nvidia thinks it will be necessary...
However, I expect SATA to be too slow and wouldn't be surprised if NVMe was a requirement for some UE5 games.
"You could render a version of this [demo on a system with an HDD], it would just be a lot lower detail," said Sweeney.
I wonder if that will ever happen.
The same way UE4 games don't really stop you from using an iGPU (it just gets unplayable-y slow), It looks like UE5 won't be able to require NVMe at all:
The game will just get super slow or look super terrible.
I mean they can put a NVMe in the minimum requirements' list, but those are usually just very relaxed guidelines and many times they're just silly.
Most of the time there's a DirectX version compliance check and that's it.
Why fast storage changes everything
The specs on this page represent only the tiniest fraction of the potential of the storage solution Microsoft has engineered for the next generation. In last year's Project Scarlett E3 teaser, Jason Ronald - partner director of project management at Xbox - described how the SSD could be used as 'virtual memory', a teaser of sorts that only begins to hint at the functionality Microsoft has built into its system.
On the hardware level, the custom NVMe drive is very, very different to any other kind of SSD you've seen before. It's shorter, for starters, presenting more like a memory card of old. It's also rather heavy, likely down to the solid metal construction that acts as a heat sink that was to handle silicon that consumes 3.8 watts of power. Many PC SSDs 'fade' in performance terms as they heat up - and similar to the CPU and GPU clocks, this simply wasn't acceptable to Microsoft, who believe that consistent performance across the board is a must for the design of their consoles.
The form factor is cute, the 2.4GB/s of guaranteed throughput is impressive, but it's the software APIs and custom hardware built into the SoC that deliver what Microsoft believes to be a revolution - a new way of using storage to augment memory (an area where no platform holder will be able to deliver a more traditional generational leap). The idea, in basic terms at least, is pretty straightforward - the game package that sits on storage essentially becomes extended memory, allowing 100GB of game assets stored on the SSD to be instantly accessible by the developer. It's a system that Microsoft calls the Velocity Architecture and the SSD itself is just one part of the system.
"Our second component is a high-speed hardware decompression block that can deliver over 6GB/s," reveals Andrew Goossen. "This is a dedicated silicon block that offloads decompression work from the CPU and is matched to the SSD so that decompression is never a bottleneck. The decompression hardware supports Zlib for general data and a new compression [system] called BCPack that is tailored to the GPU textures that typically comprise the vast majority of a game's package size."