If you'd said die, I wouldn't have posted :) You said chip though.
By a quarter.
Wow, literally in the memory: Samsung's New HBM2 Memory Has 1.2 TFLOPS of Embedded Processing Power | Tom's Hardware Today, Samsung announced...
Why do you use the term "simpler" and what do you mean by "simpler"?
Notice that NVidia prefers the smallest possible workgroup size (32) and AMD prefers the largest possible (128/256 - 256 if on console). Makes a...
GDDR bus size needs to scale with the count of graphics chiplets. An I/O chiplet providing GDDR doesn't do that.
Did we ever really understand this? PARALLEL MICROPOLYGON RASTERIZERS - Advanced Micro Devices, Inc. (freepatentsonline.com) Can anyone find a...
Here's 36+ 6800XTs sat on the shelves (add them to your basket) : AMD RX 6800 XT - Graphics Cards AMD | Ebuyer.com And currently there's 7 of...
If only cameras (or film, then scanned) were even remotely close to this ideal! Mr Mars' Camera Resolution Diatribe (warrenmars.com) We have...
TAA quality varies hugely by game.
This friend's pre-order of a 3080 at EBuyer, since November, wasn't fulfilled, so he had similar luck!
On my playroom's LG C9 77" at 8ft (245cm) viewing distance, I can't see all of the quality of 4K. On my PC's LG CX 48" at about 3ft (1m ish)...
More than 10 of these arrived in stock on Thursday morning: PowerColor Radeon RX 6900XT Red Devil 16GB Graphics Card | Ebuyer.com and there's...
I really want to play this. I love the feel of what I've seen of the game and the art style of Night City is intoxicating. I've decided to wait...
What kind of coherency-sorting would that be? What kind of scale, memory usage and latency characteristics would it have? Would the sorting occur...
Me neither. He has provided a broad overview of his process, implying it's a bot.
Ray tracing you mean? :lol: Yes I would tend to agree.
COMPUTE UNIT SORTING FOR REDUCED DIVERGENCE - Advanced Micro Devices, Inc. (freepatentsonline.com) This has been floating around for a while. I...
Whoops wrong thread. Here: https://forum.beyond3d.com/posts/2188583/
Another week: 1) RX 6900XT = 410 Units. 2) RX 6800XT = 395 Units. 3) RX 6800 = 205 Units 4) RX 5700XT = 160 Units. 5) RX 580 = 120 Units. Week 3...
Some guy on the internet is keeping track, somehow, of sales at a large retailer in Germany: Week 2 Mindfactory.de Total Gaming GPU Sales 3155...
Which games run their ray tracing effects at "full resolution" and full frame rate?
Dunno. Is there evidence that NVidia performance is improved by inline techniques?
The idea that RDNA 1 has utterly broken DXR support is fun. But I'm dubious. Though it might explain why RDNA 1 and RDNA 2 compute units appear to...
Welcome to my pain as a photographer. Films are another annoyance: two people are having a face-to-face conversation outside and the film shows...
The page I linked seems to imply that there's a variety of code that's produced, so perhaps not that surprising. I'm unclear if the ISA in the...
It looks like we have an ISA sample of the inner loop for BVH traversal from a TraceRays shader call: [IMG] which comes from: Radeon™ GPU...
Ooh, some interesting facts there: Box nodes on RDNA2 can have 4 boxes (and hence children), and come in fp16 and fp32 variants. The fp16 variant...
In the end, with Ryzen, it was about going way past the conventional reticle limit in terms of transistor count. It also helped yield. HBX will...
FWIW IMO the lighting from UE5 is clearly superior: there is graduation to the shadows which is simply missing from the UE4 render.
Part 3: DDGI Overview (morgan3d.github.io) DDGI is pretty nice and seems to successfully avoid the "artistic intervention required" problem.
You're thinking 10 or 20GB of BVH kept on disk? Streaming in a few hundred MB as needed? LOD strategies? In all these techniques, to do them in...
I wonder if the INT instructions are address calculations: need to convert box and triangle indices into real locations in memory?
I've realised that you can shift the storage of bounding boxes upwards by one level in the tree: Node { BBox[] childBoxes; NodeIndex[]...
It's cost too: silicon with TSVs is more costly. Though I'm not saying the packaging cost overhead is zero for this design. I think it's...
"Simultaneously" is strictly speaking incorrect, agreed. But all children need to be evaluated. In fixed function hardware, given that the...
There aren't TSVs in this design, because those pillars (212) you've identified are through a moulding (220). As to the way that a GPU is...
GPU CHIPLETS USING HIGH BANDWIDTH CROSSLINKS - ADVANCED MICRO DEVICES, INC. (freepatentsonline.com) Seems to be a design without TSVs and uses...
6800XT at DF's recommended settings: [MEDIA] struggling to hold 45fps in a small gunfight. The smeary "shadows" of legs as people walk around...
Dunno, you're being intentionally vague. As you know I don't consider "average" frame rate a valid assessment of playability, only 0.1% lows.
My theory is that NVidia made hardware accelerated ray tracing for the professional market and then went to Microsoft and said "hey this is...
Brute force, (e.g. "Psycho" setting in Cyberpunk 2077) which seems to be what we're seeing in existing games, is upended by bandwidth. We've gone...
I disagree. The explicit return of 4 results per query is specified, and since BVH traversal requires all child nodes to be queried simultaneously...
Close to the metal is a big deal in console world. AMD's documentation supports that.
10 years of console devs doing amazing things because of documentation...
I made a thread: https://forum.beyond3d.com/threads/rdna-2-ray-tracing.62161/ I'm very much looking forward to what console devs do.
It's interesting how that stuff will work with D3D and appears to provide some "close to the metal" functionality. I wasn't really expecting that,...
In theory the GPU and the driver can each notice these problems occurring and alter the way that the individual units of the GPU are being used....
Rys woz there: Updated getDefaultThreadCount() to better represent other vendor posi… · GPUOpen-LibrariesAndSDKs/cpu-core-counts@49a6e73...
If we were developers with the code at hand, I suppose we could go into more detailed profiling here, e.g. all of the different ray tracing...