This class could be used instead of cd file https://catboost.ai/docs/concepts/input-data_column-descfile.html when creating Pool from filez. The class should have init function, methods load and save, and Pool init method should be able to use object of this class instead of cd file during initialization.

There are several internal things that make Emu's performance potentially suboptimal. This issue is a place to discuss them.

• wgpu::Device::poll is used here and right now it blocks in an async context. I'm not sure what the solution is but there is some discussion [here](gfx-rs/wgpu-rs#214 (comment)

Formulas for pricing a Barrier option under Black-Scholes model is of interest. (See, e.g., Section 26.9 of Hull(2018), Options, Futures, and Other Derivatives, 9th edition).

The module implementing this method should live under tf_quant_finance/volatility/barrier_option.py. It should support both puts (up-and-in put, down-and-out put) and calls (down-in call, up-and-out call). Tests should be

Collect coverage information at end of tests.

https://www.veripool.org/projects/verilator/wiki/Manual-verilator

1. Pass --coverage option to verilator in makefile
2. Modify c++ test harness to VerilatedCov::write into an output file. There will be one file per test, which should be collected somewhere (this probably needs to be done by test_harness.py)
3. Run verilator_coverage executable,

Potentially dump question - the documentation of this project is generated with Sphinx. Is there any specific reason why PDF is not explicitly supported as an output format in this project? Sphinx does support it out of the box and I could patch it into the documentation's makefile if there are no objections.

Repository Description:

This guide should help fellow researchers and hobbyists to easily automate and accelerate there deep leaning training with their own Kubernetes GPU cluster.

should be

accelerate their deep leaning training

The standard accelerate test suite, used by all the backends, can be quite slow. Several of the tests are significantly slower than the others, for example segmented folds and scans, which I believe is because the reference implementations are very inefficient. Writing some more efficient reference implementations (e.g. using Data.Vector.Unboxed) should help speed things up.

The docs generated for #431 at https://mratsim.github.io/Arraymancer/pca.html have broken formatting:

• Upstream bug on the type:
  
• Random bolding:
  and
  ![image](https:/

Examples to be removed:

• space_invaders/et_space_invaders.py
• space_invaders/minimal_space_invaders.py
• space_invaders/random_baseline.py
• asteroids/random_network_baseline.py
• datasets/davis.py
• datasets/spoken_mnist.py
• mnist/batch_conv_mnist.py
• mnist/minimal_mnist.py
• mnist/minimal_resevoir.py
• mnist/minimal_supervised_mnist.py
• [

The current behaviour of glossy and glossycoating is to behave like matte on the backface.
I can't think of a good reason for this, and it is very confusing for users, so I would like to request to rework these materials to show their coating on both front and back faces.

(intended for v2.4)

Want to back this issue? **[Post a bounty on it!](https://www.bountyso

Currently the output at runtime produces, for instance, the following:

PIConGPUVerbose PHYSICS(1) | Courant c*dt <= 1.66765 ? 1
PIConGPUVerbose PHYSICS(1) | species e: omega_p * dt <= 0.1 ? 3.56797e-08

The way the output is formatted I interpret it as a "question to the code", i.e., "Is c*dt <= ?" To which a natural answer would be "Yes/No" or "True/False" and not

make_ptr in multi-pointer class is not implemented now.

Implement proposed spec maintenance change: KhronosGroup/SYCL-Docs#42

Could you please reference the papers used for the example implementations? That could possibly clear up a lot of confusion regarding the energy specifications.

Thanks!

I have discovered this package and being a new user went straight to the docs to see how to use each of these functions. I tried as.gpuMatrix and naturally would think as..... is converting whatever I give it to a gpuMatrix so I tried:

gpuR::as.gpuMatrix(c(1:3),"vector")

This results in:

Error in (function (classes, fdef, mtable) :
unable to find an inherited method for

In the array example program, we're copying data into arrays and from arrays. While we're checking the CUDA runtime API return values, we're not checking that the copying results in the expected data. We may be printing information from which it's possible to deduce that, but not actually checking.

Let's do that... @codecircuit ?

This would make it easier for new users to find the denoised result.
Also better compatibility with Cycles compositing node trees.

Would be nice to implement #251 at the same time so the user can still see progress in endless render mode.

Want to back this issue? **[Post a bounty on it!](https://www.bountysource.com/issues/87091323-output-denoised-image-into-co

It seems that there is a bug with the call to gpufit within Matlab when including the user_info parameter. Using the included linear_1d model (which utilizes the user_info parameter), I created a simple program in Matlab to model the equation y=x from x=0 to x=10 and called gpufit on the data. This should return the parameters 0 and 1, but results in 4.3467 and 0.8711 instead.

Additionally, if