marshallward

5 ms·

1.
by marshallward 2y ago
> I hope you can list all of the groups of order < 8 without having to think too hard. Welp, guess I'm out.
2.
by marshallward 2y ago
Here is more recent video that beats those records: https://www.youtube.com/watch?v=jPyf5kIGoKA Except! Even this is now out of date, and I think BlueScuti has broken some of these records! The Tetris leaderboards are ver
3.
by marshallward 2y ago
> “The pyramids seem like pretty monumental work” You don't say...
4.
by marshallward 2y ago
I ran with OMP_NUM_THREADS=1, but your point is well taken. As for the original post, I felt a bit embarrassed about my original comments, but I think the compilers actually did fairly well based on what they were given, which I think is wh
5.
by marshallward 2y ago
I'm sure there's more to it, but just comparing the profile output shows aggressive use of prefetch and broadcast instructions.
6.
by marshallward 2y ago
I just did a test of OpenBLAS with Intel-compiled BLAS, and it was about 6 GFLOP/s vs 150 GFLOP/s, so I must admit that I was wrong here. Maybe in some sense 4% is not bad, but it's certainly not good. My faith in current c
7.
by marshallward 2y ago
You are right, I just tested this out and my speed from BLAS to OpenBLAS went from 6 GFLOP/s to 150 GFLOP/s. I can only imagine what BLIS and MKL would give. I apologize for my ignorance. Apparently my faith in the compilers wa
8.
by marshallward 2y ago
I don't disagree, but where are those techniques presented in the article? It seems like she exploits the particular shape of her matrix to align better with cache. No BLAS library is going to figure that out. I am not trying to say
9.
by marshallward 2y ago
How do you feel about Nvidia endorsing do concurrent migration to GPUs? Would that be classified as parallelization?
10.
by marshallward 2y ago
Fair enough, this is not meant to be some endorsement of the standard Fortran BLAS implementations over the optimized versions cited above. Only that the mainstream compilers cited above appear capable of applying these optimizations to t
11.
by marshallward 2y ago
There is an implication here that the Fortran implementation of `SGEMM` is somehow inadequate. But any modern Fortran compiler will quite easily apply the AVX and FMA optimizations presented here without any additional changes. Both GNU a
12.
by marshallward 3y ago
Unfortunately I don't think the compilers will add it in time.
13.
by marshallward 3y ago
I have only looked at the GCC Fortran frontend. I would not say it is necessarily difficult but rather very ad-hoc and disconnected from the rest of gcc. I don't see many tools for specifying a grammar or tokens. Instead, there are
14.
by marshallward 3y ago
I know someone who took that course. They did not have fond memories of it. My impression is that it can be a very frustrating way to learn mechanics if you don't have much interest in functional programming.
15.
by marshallward 3y ago
It's an excellent essay, and the Fortran community owes you a major gratitude for promoting these issues. But surely there loops can be safely parallelized if the iterations do not interact, e.g. per-element array arithmetic, and a co
16.
by marshallward 3y ago
Very true, and it opens the door to new optimizations which are not possible when going through a library (at least in the case of MPI).
17.
by marshallward 3y ago
I agree that it is not a parallel construct. But it can nonetheless be parallelized if certain criteria are met.
18.
by marshallward 3y ago
Do concurrent loops can also be parallelized.
19.
by marshallward 3y ago
I remember going to jai alai games in Miami as a kid. I was never very into sports, but it was fun to watch. As far as I could tell, the adults were there to bet on the games.
20.
by marshallward 3y ago
What is it about?
21.
by marshallward 3y ago
I have not yet seen anything like this, but I have observed the Intel compiler produce different bytecode on Intel and AMD machines. In one case, the compiler would use masking vector instructions on Intel but not on AMD. It was also the c
22.
by marshallward 4y ago
We work hard to retain floating point reproducibility in climate models. I have a presentation on this, if anyone is interested. https://www.marshallward.org/fortrancon2021/#/title-slide https://m.yout
23.
by marshallward 4y ago
True but every kind is considered a type.
24.
by marshallward 4y ago
Fortran is no longer in ALL CAPS.
25.
by marshallward 4y ago
Fortran supports both of these, with `mod` as the C-like truncated modulo and `modulo` as the floored modulo. Having both is convenient, but you do get errors from people who don't realize the difference
26.
by marshallward 4y ago
I think the statement has different meanings to different people, but one "unsolved" problem is the ability to model the dynamics of small scales (the "turbulence") in terms of the large-scale "observed" flow.
27.
by marshallward 4y ago
We use biharmonic diffusion operators in oceanography as a basic turbulence model, which is effectively a fourth order derivative in space. It is like an accelerated diffusion which aggressively dissipates small scales but preserves the la
28.
by marshallward 4y ago
I had no idea this was a thing, but I did notice Chris Lattner (of LLVM) is a woodworker. https://nondot.org/sabre/Woodworking.html
29.
by marshallward 4y ago
Same for me, it was laughably bad.
30.
by marshallward 4y ago
I find autoconf to be good at its job - building a standalone script which runs micro tests - and still use it, but that automake and libtool add very little value for the reasons you described. I do agree that m4 is rather unpleasant to w

More ›