my123

73 ms·

871.
by my123 6y ago
They went to Samsung instead to get more capacity this gen.
872.
by my123 6y ago
NVIDIA doesn’t own any fabs, they don’t really have a choice.
873.
Qualcomm Completes Acquisition of Nuvia: Immediate Focus on Laptops(anandtech.com)
18 points by my123 6y ago|0 comments
874.
Qualcomm Completes Acquisition of Nuvia(qualcomm.com)
6 points by my123 6y ago|0 comments
875.
by my123 6y ago
Intel and AMD also did the exact same thing with x86. But yet, people still use it. (and Arm also has its fair share of patents, but is a licensable architecture by others)
876.
by my123 6y ago
Intel and NVIDIA have a patent cross-licensing agreement at the time, which is still ongoing for all patents filed on or prior to March 31, 2017. https://www.sec.gov/Archives/edgar/data/1045810/0001193125
877.
by my123 6y ago
This chapter of the NV driver manual details how: https://download.nvidia.com/XFree86/Linux-x86_64/460.56/READ... On Turing and later, you can also enable this: https://download.nvidia.com/XFr
878.
by my123 6y ago
Right-click + Open has to be done twice to work. The anywhere option is hidden by default, use sudo spctl —-master-disable from Terminal instead, which flips Gatekeeper settings to Anywhere. Edit: there's also this path: https:/&
879.
by my123 6y ago
System Preferences -> Security and Privacy. You’ll find an override toggle there. You can also use Right Click -> Open when launching the app. And no, you don’t need to disable SIP to disable Gatekeeper which can be done through sudo
880.
by my123 6y ago
Using Vulkan for compute is still a much worse developer experience, it isn't really a reasonable alternative.
881.
by my123 6y ago
OpenCL as implemented by AMD is definitely not very usable. AMD themselves aren’t really committed to OpenCL given the state of their OCL 2.x drivers and went to their own AMD-only path (ROCm/HIP), which is broken... (and only supporte
882.
by my123 6y ago
It pre-dates Device Tree support on Linux for Arm. Both have a common ancestor (OpenFirmware, notably used for PowerPC Macs)
883.
by my123 6y ago
It can be disabled freely by the user in System Preferences -> General on macOS.
884.
by my123 6y ago
https://github.com/simdjson/simdjson as an example that fits in the "outside-of-a-conventional-SIMD-workload" mold.
885.
SK hynix Starts Mass-production of LPDDR5-6400 chips with 18GB of RAM(videocardz.com)
2 points by my123 6y ago|0 comments
886.
Apple M1 Microarchitecture Research(dougallj.github.io)
237 points by my123 6y ago|48 comments
887.
by my123 6y ago
Hello, Intel GVT-g is not supported on Gen11 and Gen12 gen Intel Graphics. It's only supported on Gen9 and earlier. About Intel discrete graphics: https://github.com/intel/gvt-linux/issues/160 > Inter
888.
by my123 6y ago
The BeagleV ( https://beagleboard.org/beaglev ) also has an NVDLA in the SoC. (1 core w/ 2048 MACs at 800MHz)
889.
Nvidia Deep Learning Accelerator (NVDLA): free open inference accelerator (2017)(nvdla.org)
62 points by my123 6y ago|25 comments
890.
by my123 6y ago
GPGPU with the graphics APIs isn’t a comparable developer experience at all, and comes with quite some major limitations. About DirectCompute, C++ AMP is in practice dead on Windows, stuck at a DX11 feature level with no significant updat
891.
by my123 6y ago
On Xavier, the dedicated AI inference block is open source hardware. Available at http://nvdla.org/
892.
by my123 6y ago
“Making science accessible to everyone is important to us. That's one of the reasons why GeForce cards can run CUDA.” — Bryce Lelbach at Nvidia AMD meanwhile considers GPU compute as a premium feature... which is a different approach.
893.
by my123 6y ago
It does not. And even then, almost all AMD customer GPUs are not supported. (Everything but Vega) APUs? Nope too.
894.
by my123 6y ago
You have https://github.com/NVIDIA-AI-IOT/torch2trt as an option for example to use your own models on TensorRT just fine. And https://github.com/tensorflow/tensorrt for TF-TRT integration.
895.
by my123 6y ago
They made new compute cards, but they aren’t available to customers. (Only businesses, under the Radeon Instinct brand) With the price to match for those...
896.
by my123 6y ago
Jetson Xavier NX, but that comes with a high price tag. It’s much more powerful however.
897.
by my123 6y ago
RX580 is out of support for ROCm nowadays, which only supports Vega and Instinct now. https://github.com/RadeonOpenCompute/ROCm/issues/1353#issuec...
898.
Reverse-engineering Rosetta 2 Part 1: Analyzing AoT files and the runtime(ffri.github.io)
121 points by my123 6y ago|17 comments
899.
by my123 6y ago
In fact, OpenGL on Apple M1 is implemented through AppleMetalOpenGLRenderer.
900.
by my123 6y ago
It does, and it's called Arm SystemReady. UEFI + ACPI is required on Arm if you want to boot Windows or RHEL/CentOS, which do not even try to boot on non-standardised 64-bit Arm systems.

More ›