Bimos

I'm super confused. In your example, ++i is equivalent to (i+1) and the loop updates i itself. But with the syntax in the blogpost, i is only initialized, and it is the ++i that updates the value?

They came up with a Chinese name for you :)

柯亦安 Ke Yi'an. The given name "亦安" roughly means "also safe", it is a beautiful name.

If you do not have a social credit score, you will have to prepay a small deposit of 80¥.

It is acceptable if you guys are meaning that thing when mentioning "social credit score"..

A perceptron *is* the smallest brain you can build.

In 1958, a researcher named Frank Rosenblatt built a machine *he called* the perceptron.

It was *inspired* by a single brain cell, a neuron.

I'm not sure about whether this is a bottlenecking step in applications

In barrier method of solving linear programs, numeric factorization normally takes about 1/3 of the time of main loop (like after all symbolic/presolve steps).

That is, is there a (sparse) matrix representation which is used in gpu's?

We (COPT) did manage to write a gpu version of Cholesky factorization, and it's faster than our (pretty fast) cpu implementation. Although one can argue that the CPUs for benchmarking is not the fastest, or it is not as expensive as the GPUs.

Currently our barrier implementation is 67% faster on H200 than on i7-11700K [1]. Not all of the improvement contributes to Cholesky factorization.

Another GPU Cholesky factorization implementation we are aware of is CUDSS by nvidia.

And does it make sense to carry through the dag/tree construction as a sort of "prep" step (on cpu or gpu)?

For both CPU and GPU version, a symbolic phase is necessary, as we usually expect many factorizations on matrices with the same sparse pattern. CUDSS has (at least some part of) their symbolic phase on GPU.

[1] https://plato.asu.edu/ftp/lpfeas.html

The toilet also features a rotating heating element that evaporates urine and dries other waste.

Electrical cars got exhaust gas now, in a worse way.

a block of Rust threads that are properly programmed to take advantage of the vector processing by avoiding divergence

But the library is using a warp as a single thread

You didn't explain or prove there is a reason for it.

My car is 605km in CLTC and 492km in WLTP. And it runs for about 450km before battery totally die, so I have to charge for every about 400~420km. A bit cold and totally no hill here in Shanghai.

So for 900km CLTC, I think it will go like at least 600~650km easily.

I like the fact that the photos seem to be real rather than simulated.

Yeah, with all these colors, any single character should be enough.

Look at "Four Asian Tigers", people don't just constantly become rich for granted

I thought no guessing mode means the map is generated in such way that guessing is not needed to solve the whole board? (To those who didn't know the mode, it also forces you a starting point)

As an experiment, I tried clicking on a random position after clicking the starting point. And it is possible to that the position is safe. So I don't think the map in no guessing mode is dynamic.

Also in the help section: "In this mode, a starting position is provided, and you never need to guess to complete the board."

It looks like an unconstrained least square problem to me, which is convex and naive. Similar projects like reconstructing an image with transparent triangulars seems more interesting.

Same here.

Yeah but since it claims "any CSV file", and CSV files are widely known to be variate, I didn't expect it fails to work on edge use-cases.

“Even if the U.S. continues to impose higher tariffs, it will no longer make economic sense and will become a joke in the history of world economy,” the ministry said in a statement, according to a CNBC translation.

(x - m) / (εφm)

ε, was set at 4.

φ, is 0.25.

What a coincidence!

I would assume Guilin Noodles are cousin of Vietnam Pho, based on the similarity of their ingredients and their geographic positions.

Maybe add Chimera as well?

https://arxiv.org/pdf/2107.06925

I heard that it is possible to achieve better performance than cuBLAS using CUTLASS? I thought they chose the better one among cuBLAS and CUTLASS as baseline.

I edited it.

I think most AI players rely on high performance GEMM. But most people would be satisfied with cutlass or cublas, and the others implement gemm themselves, but not necessarily use undocumented features?

FFMA SASS interleaving

We observe a performance improvement in the CUTLASS FP8 kernel between NVCC 12.2 and 12.3. By comparing the compiled SASS, we discover that one bit in a series of FADD instructions is flipped in an interleaving pattern. After referencing some open-source CUDA assembler implementations, we identified that this bit controls yield, which may enhance warp-level parallelism (just a guess, yielding the current warp and let other warps work).

To leverage this, we develop a similar script to modify the FFMA instructions in the compiled binary. Besides simply modifying the yield bit, we also flip the reuse bit (registers cannot be reused if the warp is yielded). This adjustment improves performance (10%+ in some cases) for fine-grained scaling FP8 GEMMs by creating more opportunities to overlap MMA instructions with promotion FFMA instructions.

I would say it is really mind-blowing.

The PTX instructions they talked about in the tech report should be pointing to the code here?

How is it different from eliminating y?

As a Chinese I want to mention that foreign people don't usually call it "PPT".

Let's see how US reacts to this incident in Munich meeting

If GPUs Are So Good

Who ever claimed that?

In Shanghai, the Huangpu River divides old and new.

Pudong is new, but Puxi is not "the old one". It is mixed with old historical sites and modern buildings. It lives with its history, and it grows actively. It doesn't have to destroy all the old stuff, but it also doesn't have to slow down its pace for them. I believe it is the same for any city (e.g. in Europe) with history.