The rumours about AMD's new 7nm Navi GPU are continuing to spread, this time via The Inquirer. They have heard tell that the chips are currently being tested, which is good news for both AMD and consumers. We know that new Vega cards will be arriving in the near future but as of now we don't know as much about the release of Navi. If they are currently testing the silicon that could imply some sort of release next year, perhaps not a full lineup but quite possibly something we can take a peek at and see how it stacks up to Turing. Lets hope it is sooner rather than later.
"Touted to be the next graphics accelerators to take the fight to Nvidia with its Turing-based GeForce RTX graphics cards, Navi is set to use a 7nm process which promises to get more power out of GPU silicon."
Here is some more Tech News from around the web:
- 'We Expected VR To Be Two To Three Times as Big', Says CCP Games CEO @ Slashdot
- If you want to rent AMD Epyc bare-metal boxes in the cloud, Oracle hopes you see red @ The Register
- NAND flash prices may see larger drop in 2019 @ DigiTimes
- Zip it! 3 more reasons to be glad you didn't jump on Windows 10 1809 @ The Register
- Another Windows 0-day flaw has been published on Twitter @ Ars Technica
- Apple and Samsung fined over performance-throttling software updates @ The Inquirer
- Grocery Delivery in 2018: Does It Finally Work? @ TechSpot
It seems to me that the only
It seems to me that the only way we’ll see anything close to 2080ti, or even 2080 levels of perf from AMD is if they majorly rework GCN to the point where it really isn’t GCN anymore.
Especially if they want to keep the perf up and die sizes to a reasonable level to keep yields up.
Here’s hoping AMD can “pull a rabbit out of its hat”
I am going to bet that AMD is
I am going to bet that AMD is shooting more towards the 2070 range of performance. They deem the Navi as a mid range based card and the rumor mill is flying around with 1080-2070 range performance. Which is not to bad for a mid tier card and if the price is right as to what the rumors are saying I think was like $250-$300 US. If those prices are close and it does reach 1080-2070 range I think AMD would have a winner on their hands for sure.
Of coarse we will have those that say oh but it does not reach 2080-2080Ti speeds. My response would be guys be happy we got a card that performs great and does not have the Nvidia greed tax attached to it & if you want 2080 performance then get ready to spend at least $300-$350US extra on a 2080 and only get like 16-20% better performance.
If you want 2080ti performance then get ready to pay $900US extra for that extra performance which would only be about what 30-35% extra for $900US more. If AMD delivers on the performance of a 1080-2070 range and can do $300US-325US they have my money this time around.
If AMD would just spin up a
If AMD would just spin up a new Vega Base Die Tapeout with more ROPs and the higher Pixel Fill Rates that come with having more ROPs then AMD could compete with the GTX 1080Ti(88 ROPs) in raster oriented gaming.
AMD will need some years to engineer and certify any Ray Tracing cores IP but Tensor Cores are just glorified matrices math units so AMD could get Tensor Cores added rather qucikly. In Fact I’ll bet that MS(XBONE Next) and Sony(PS5) would like some Tensor Core based “DLSS” Like Upscaling functionality done via Trained AIs running on Tensor Cores. So AMD’s Semi-Custom Folks may already be looking into Tensor Cores for its next generation products for its console clients if they will fund the development costs.
What AMD needs to look into also is maybe getting some Bins built up of those Vega 20 DIEs that are sure to not have enough functioning units for any Radeon Instinct usage because of defects. So maybe a Dual GPU Die on one PCIe card consumer gaming flagship GPUs made from binned Vega 20 DIEs with plenty of ROPs, TMUs and Shader cores. AMD could wire these dual GPU DIEs on a Single PCIe card up Via its Infinity Fabric and like NVLink have the dual dies appear as one larger logical GPU.
So on the GPU’s PCB use the Infinity Fabric Protocol to Interface the Dual GPU DIEs and not PCIe that lacks the same levels of inter-GPU-Die coherency traffic capabilities that IP like the Infinity Fabric and NVLink support. Vega 20 will support xGMI(Infinity Fabric Based) for linking up 2 GPUs anyways so just create a dual binned Vega 20 DIEs based GPU SKU for gaming on a single PCIe card and wire then up via xGMI(IF Based).
Navi is currently only going to be used for Mainstream GPUs but who knows what AMD will do with Navi after AMD has had enough time and the proper ampunts of funding. If Navi supports xGMI also maybe AMD could just go back to using an xGMI bridge instead of xDMA over PCIe as was done for later versions of CrossFire.
Maybe AMD should take a
Maybe AMD should take a license from nVidia like GlobalFoundries took a license from Samsung… LOL