Intel Plans to Launch Its Discrete GPU Lineup Starting at $200

IdiotInCharge said:
Really, really doubtful that it would benefit games in the near future, but there are plenty of possibilities here going forward.

As to consumers, if one can slap on an Optane (or similar SSD designed to be abused) for the purpose of providing local resources for content creation workloads- or even for game streaming!- I can definitely see a market here.
Click to expand...
Exactly.
It may be a trivial boost to gamers but as we've learned over the years, videocards are being used for more than games. Mining, imagine processing, large scale artificial intelligence, supercomputing, streaming and likely some things people will find a use for if they implement a couple empty slots for. Sure, a low end card doesn't need it, but Intel wants to come out swinging. Maybe something they could reserve for midrange and top end cards.

I'm not saying its faster than video ram or even system ram, but not having to share system memory? That's a bonus. Having 2TB memory on top of the HBM2 memory is a bonus.
They wouldn't need to preinstall any SSD like, just 2 empty the slots, let the consumer figure out if the want to add it.

https://www.amd.com/en/products/professional-graphics/radeon-pro-ssg

Video cards are for more than just games.
 
oldmanbal said:
Intel is just not going to be competetive on the high end, period. Amd was once competitive and look where they are now. I dont get all the people that think intel will magically dethrone nvidia who are a decade ahead of them in r and d. Its nice to have more options but thats all its gonna be for years
Click to expand...
It wasn't that long ago that Nvidia came out of nowhere and now 1 company is gone, and several others don't make GPUs anymore. So anything is possible
 
Factum said:
So you have no clue if it will help games...but still want it?
(System RAM is faster and cheaper FYI)
Click to expand...

And now Nvidia is adding 2 NVMe slots in raid to their cards...
https://devblogs.nvidia.com/gpudirect-storage/

https://blocksandfiles.com/2019/08/06/nvidia-gpus-direct-access-to-nvme-storage/

"Using GPUDirect™, multiple GPUs, network adapters, solid-state drives (SSDs) and now NVMe drives can directly read and write CUDA host and device memory, eliminating unnecessary memory copies, dramatically lowering CPU overhead, and reducing latency, resulting in significant performance improvements in data transfer times for applications running on NVIDIA Tesla™ and Quadro™ products"..

https://bit-tech.net/news/tech/graphics/nvidia-details-gpudirect-storage-performance-gains/1/
"Explicit data transfers that don’t fault and don’t go through a bounce buffer are also lower latency; we demonstrated examples with 3.8x lower end-to-end latency; avoiding faulting with explicit and direct transfers enables latency to remain stable and flat as GPU concurrency increases; use of DMA engines near storage is less invasive to CPU load and does not interfere with GPU load; the ratio of bandwidth to fractional CPU utilisation is much higher with GPUDirect Storage at larger sizes."

Maybe not many of today's games, but future games? Certainly.
Intel could just add 2 empty slots...
 
Last edited:
griff30 said:
And now Nvidia is adding 2 NVMe slots in raid to their cards...
https://devblogs.nvidia.com/gpudirect-storage/

https://blocksandfiles.com/2019/08/06/nvidia-gpus-direct-access-to-nvme-storage/

"Using GPUDirect™, multiple GPUs, network adapters, solid-state drives (SSDs) and now NVMe drives can directly read and write CUDA host and device memory, eliminating unnecessary memory copies, dramatically lowering CPU overhead, and reducing latency, resulting in significant performance improvements in data transfer times for applications running on NVIDIA Tesla™ and Quadro™ products"..

https://bit-tech.net/news/tech/graphics/nvidia-details-gpudirect-storage-performance-gains/1/
"Explicit data transfers that don’t fault and don’t go through a bounce buffer are also lower latency; we demonstrated examples with 3.8x lower end-to-end latency; avoiding faulting with explicit and direct transfers enables latency to remain stable and flat as GPU concurrency increases; use of DMA engines near storage is less invasive to CPU load and does not interfere with GPU load; the ratio of bandwidth to fractional CPU utilisation is much higher with GPUDirect Storage at larger sizes."

Maybe not many of today's games, but future games? Certainly.
Intel could just add 2 empty slots...
Click to expand...

But the time games require such OBSCENCE mounts of DATA...onboard RAM amound and speed will have increased, making this solution look retarded ;)
This is for CORNER NICHE cases...and will never be used for gaming.
 
Factum said:
But the time games require such OBSCENCE mounts of DATA...onboard RAM amound and speed will have increased, making this solution look retarded ;)
This is for CORNER NICHE cases...and will never be used for gaming.
Click to expand...

Well, Intel is plugging Optane NVMe directly into DIMM slots- also for niche cases, but at the same time, they're pushing that 'persistent RAM' thing again and are closer than anyone else has gotten.

The use case for games would be 'obscene' amounts of world assets; more than would be feasible to put in system RAM, if that's still a problem, and more than you'd want to stream over PCIe, if that's still a problem. Hard to imagine that being a consumer application though, even if used for games, outside of boutique setups.
 
You must log in or register to reply here.
Back
Top