jra101

You're blaming NVIDIA for Sam Altman's dumb idea?

NVIDIA has an OpenGL extension that does just that [1].

[1] https://registry.khronos.org/OpenGL/extensions/NV/NV_fill_re...

Have you tried reducing the register count in your FP32 FMA test by increasing the iteration count and reducing the number of values computed per loop?

Instead of computing 8 independent values, compute one with 8x more iterations:

    for (int i = 0; i < count * 8; i++) {
        v0 += acc * v0; 
    }

That plus inlining the iteration count so the compiler can unroll the loop might help get closer to SOL.

This happens to me just using the adaptive cruise control, no Autopilot or FSD enabled and it's super annoying. Can happen on a completely empty road driving in a straight line.

PS4: ~117M sold in 9 years Switch: ~104M sold in 5 years

Switch is averaging 20M units/year and PS4 is averaging 13M units/year.

Thankful that Tweetdeck still works and has never tried to switch me to a non-chronological feed.

They are referring to the GPU part of the chip. There are two separate GPU complexes on the die but from the software point of view, it is a single large GPU.

Huh, I had no idea they used Vega 20 both as a codename and a product name. Confusing.

Vega 20 should be ~13.8 TFLOPs (single precision): https://www.anandtech.com/show/13923/the-amd-radeon-vii-revi...

The handling of bookmarks is terrible in the latest Firefox. So bad that I can only assume the developers don't actually use bookmarks and just search for websites by name?

Tile based Deferred Rendering came from PowerVR. Prior to moving to their own GPU architecture, Apple licensed various PowerVR designs.

Defold might be an option.

https://www.defold.com/

"Looking at the numbers, the GeForce GTX 1660 Ti delivers around 37% more performance than the GTX 1060 6GB at 1440p, and a very similar 36% gain at 1080p"

Perfect dies (64 CUs) get sold as a Radeon Instinct MI60 at a much higher price.

This is awesome, worked perfectly.

Only downside is if you have swipe to unlock, I couldn't actually get the swipe up to trigger the password input screen.

Intel integrated graphics are not executed by the CPU, there is an intel designed GPU on the same die as the CPU.

https://www.anandtech.com/show/9582/intel-skylake-mobile-des...

HBM is not exclusive to AMD, NVIDIA also ships GPUs using HBM (Tesla P100 and V100)

First Apple SoC using their own GPU design, no more PowerVR IP.

That's a feature of the graphics card, AMDs new Vega GPU architecture.

Gaming GPUs still more than 50% of revenue for NVIDIA:

http://www.anandtech.com/show/11361/nvidia-announces-earning...

Perfect, thank you.

What happen to Python support? It was supported in the release candidate.

Sounds like you are describing an NVIDIA Shield TV: https://www.nvidia.com/en-us/shield/shield-tv/

Intel's drivers may be beta but official NVIDIA and AMD GPU drivers both fully support Vulkan on Windows.

It is higher level than Vulkan as you don't need to allocate or manage GPU memory, the implementation handles that for you. You also don't need to worry about transitioning render surfaces between different states (D3D12 and Vulkan).

It does allow you to create pipeline state objects ahead of time like Vulkan and D3D12 but they only include a subset of state (shader + antialiasing state + color/depth attachment state). The rest of the state is set when recording the command buffer.

No mention of sensor size?

HTC One (M8), launched March 2014 can do this:

http://www.anandtech.com/show/7893/the-htc-one-m8-review/2

It's only ~50 lines of code to create a window:

http://pastebin.com/drrRJ6Kd

PFN_vkCreateDebugReportCallbackEXT dbgCreateDebugReportCallback = (PFN_vkCreateDebugReportCallbackEXT)vkGetInstanceProcAddr(instance, "vkCreateDebugReportCallbackEXT");

They don't ask "do you like this?" anymore. They ask something like "show more like this?" and I'm afraid clicking No will remove tweets from people that happened to be in the "while you were away" section.