my123

26 ms·

391.
by my123 4y ago
It does not. ARMv8.6 with PAC and MTE, but no SVE.
392.
by my123 4y ago
Small unit count availability is still problematic for now. For Ampere, there has been the AVA Developer Platform ($5450, 32-core) recently: https://www.ipi.wiki/products/ava-developer-platform but that's not a ch
393.
Broadcom to 'focus on rapid transition to subscriptions' for VMware(theregister.com)
13 points by my123 4y ago|3 comments
394.
by my123 4y ago
Not just NVIDIA, ZFS too. From the GPL perspective, any non GPL compatible license is considered the same way. That said, that part of the GPL on library linking was never attempted to be enforced in court. For additional entertainment: red
395.
Amazon Graviton3 vs. Intel Xeon vs. AMD EPYC Performance(phoronix.com)
9 points by my123 4y ago|0 comments
396.
Broadcom Agrees to Purchase VMware Shaking Up the Industry(servethehome.com)
1 points by my123 4y ago|0 comments
397.
by my123 4y ago
That's indeed a quite raw abstraction, but is way too powerful performance-wise to not expose...
398.
by my123 4y ago
The C++ standard encourages, but doesn't require concurrent forward progress. However, a guarantee for forward progress for diverged threads in a warp is provided for Volta.
399.
by my123 4y ago
> Modern CUDA uses explicit programmer-managed masks, which is powerful and takes advantage of their hardware specifics. But mis-using the mask can cause a deadlock, as divergent threads could simply never participate in a subgroup opera
400.
by my123 4y ago
OpenCL 1.2* with very few extensions, pretty much stuck in the past. Don’t even think about using unified memory from there. SYCL: due to the above, not on the GPU. Just on the CPU. * x86_64 apps also get to see the OpenCL CPU backend in ad
401.
Running PyTorch on the M1 GPU(sebastianraschka.com)
92 points by my123 4y ago|42 comments
402.
by my123 4y ago
Virtual PC for Mac did get an update to run on the G5.
403.
by my123 4y ago
For proprietary code, shipping object files often gives out details that the vendor explicitly doesn't want to be released. So unless the vendor has a specific toolchain for stripping info from those (like NVIDIA for their kernel modul
404.
by my123 4y ago
> so, uh, what's wrong with g++ -static foo.cpp glibc is LGPL, so license incompatibility might ensue if you aren't using musl. And it's far more than just OpenGL as an issue, there's PAM too among other things. Windo
405.
by my123 4y ago
> The Mac Studio isn't some crazy exotic hardware like datacenter class GPUs, but definitely has some exotic capabilities. Datacenter class GPUs are expensive yeah, but are quite easy to buy, even in a single unit amount. example:
406.
by my123 4y ago
> - It needs extremely high-bandwidth controllers, which severely limits the amount of memory you can use (Intel Macs could be configured with an order of magnitude more ram in it's server chips) In the first half of 2023, NVIDIA Gr
407.
by my123 4y ago
The neural engine is only exposed through a CoreML inference API. You can't even poke the ANE hardware directly from a regular process. The interface for accessing the neural engine is not hardened (you can easily crash the machine fro
408.
by my123 4y ago
> but they're already way ahead on energy efficiency 1) Nope. For neural network training not the case: https://tlkh.dev/benchmarking-the-apple-m1-max And that's with the 3090 set at a very high 400W power limi
409.
by my123 4y ago
Yeah that's within the M1 family, but get within dGPUs and it doesn't even come close. 30Tflops for a 3080 for vector FP32, but 119Tflops FP16 dense with FP16 accumulate, 59.5 with FP32 accumulate, and if you exploit sparsity then
410.
by my123 4y ago
I'd say that there has been quite some sharing between both back and forth. Evolutions in compute stacks shaped modern graphics APIs too. Texture units are indeed a part that is useful enough to be exposed to GPGPU compute APIs directl
411.
by my123 4y ago
AMX is indeed very nice for FP64 where customer GPUs aren't an alternative at all. However, for lower precisions (which is what deep learning uses), you're much better off with a GPU.
412.
by my123 4y ago
That terminology isn't used at all in GPGPU compute APIs specifically tailored for that purpose, which use quite different programming models where you can mix host and device code in the same program. And there are "GPUs" to
413.
by my123 4y ago
Apple doesn't have a separate API tailored towards compute only, but a single unified API that makes concessions to both. Concessions towards compute: a C++ programming language for device code (totally unlike what's done for most
414.
by my123 4y ago
Low. Apple doesn't have matrix math accelerators in their current GPUs. The neural engine is small and inference only. It's also only exposed by a far higher level interface, CoreML. Where it could still make sense is if you have
415.
by my123 4y ago
MISRA C or MISRA C++ are used, enforcing much stricter guarantees than what the C or C++ specs provide.
416.
by my123 4y ago
> And I understand NVIDIA is even worse re: GPU virtualization. Nope, it’s much better on the nvidia side actually. The latest AMD GPU with a publicly accessible OSS GIM driver is the AMD S7150, which was released in 2016. ( https:/
417.
by my123 4y ago
And very few AMD GPUs even support it... (and good luck finding a remotely recent AMD GIM driver)
418.
by my123 4y ago
btw for the Oracle vs Google lawsuit, OpenJDK was licensed as GPLv2. It was very much a GPL (purported) violation case.
419.
by my123 4y ago
And display outputs too. What happens if the user plugs in multiple displays to the hub could be an important scenario to support... (and for example, macOS doesn't support DisplayPort MST, requiring an encapsulation over Thunderbolt t
420.
by my123 4y ago
Snapdragon 7c Chromebooks do indeed outperform the Pentium-based machines… Says more about Intel’s Atom line than anything tbh.

More ›