Jayesh Badwaik

With his eyes closed, this rookie ventures boldly into the unknown.

[Jayesh Badwaik @jayesh_badwaik]

https://t.co/kRMo1finVb

[Jayesh Badwaik @jayesh_badwaik]

@MPECSInc @thurrott https://t.co/kZHKV6aulO

Actually not a single problematic patch from the experiment was accepted from what I read. Only good patches were accepted.

[FZ Jülich-JSC @fzj_jsc]

After our HPC storage upgrade from JUST5 to JUST6 finished in June, users already report remarkable I/O performance improvements: writing their large checkpoint files with nekRS got 10x faster (280s ↘ 25s). Thanks to our great storage team on today's #SysAdminDay! //

[Jayesh Badwaik @jayesh_badwaik]

https://t.co/fpeglafEWA

[Jayesh Badwaik @jayesh_badwaik]

@fasterthanlime Yeah, the cache download situation breaks down when your cache has huge number of combinations and at any given point you only need 1.

[Jayesh Badwaik @jayesh_badwaik]

In our @scalable_hpc project, we help accelerate Lattice Boltzmann Applications. Read my post about a poster about the project, including first results:
https://t.co/7PLvP4UOvm

#HPC #GPU

[FZ Jülich-JSC @fzj_jsc]

We are proud to host JUPITER, the first European Exascale system by @EuroHPC_JU!
The contract is awarded to a consortium of @Evidenlive and @ParTec_AG. JUPITER will consist of a Booster (NVIDIA technology) and a Cluster (SiPearl Rhea). #HPC #exa_jupiter https://t.co/zZZzjSyBpU

[Jayesh Badwaik @jayesh_badwaik]

@BarryRevzin @timur_audio Making attributes not ignorable seems like a much harder problem to solve. Substantial sustained efforts have been made to prevent giving attributes any required semantic meaning or even giving them diagnostic ability. I doubt this will change in the future.

[Jayesh Badwaik @jayesh_badwaik]

@BarryRevzin @timur_audio It is super compelling. It drastically improves both readability and diagnostic ability of the compiler, especially since attributes are ignorable, and there is nothing preventing attributes with spelling mistakes to be compiled.

[Jayesh Badwaik @jayesh_badwaik]

@timur_audio @JakeArkinstall @mknejp Actually not. Because imagine teaching a new student of C++
1. [[]] are attributes
2 [[gcc::]] is compiler specific attributed
3. [[pre:]] is a contract.

I'll not blame the learner if they refuse to use any of the three facilities altogether after learning this.

[Miro Knejp @mknejp]

@timur_audio @JakeArkinstall If it looks like an attribute people will think it's an attribute and attach the semantics of attributes to it. Most devs don't memorize grammar production rules. 10 years from now you'll still be clarifying this on the internet. Is it really worth this confusion?

[Miro Knejp @mknejp]

@MatRopert Not having a pointer is not the same as having a null pointer.

[Martin Hořeňovský @horenmar_ctu]

@MatRopert Using T* as bad optional is how you get "actually you can't associate a nullptr with key, or everything breaks", so they are correct. :-P

[FZ Jülich-JSC @fzj_jsc]

JSC employees will contribute to the #ISC23 with numerous talks, tutorials, and workshops. You can find all details on our website ➡️https://t.co/dUK7OrGVwV
@ml_marce1 @ChelseaMJohn @jayesh_badwaik @Claire_Vedette @bernd_mohr @MorrisRiedel @AndiH

[Jayesh Badwaik @jayesh_badwaik]

@gregcons @xray_matt Are there C++ conferences that already cover these costs? It would be nice to know about them.

[Martin Hořeňovský @horenmar_ctu]

Wrote up a new blogpost, this one is about Amdahl's law

https://t.co/UPWq5RucPn

[Jayesh Badwaik @jayesh_badwaik]

@MalwareMinigun @horenmar_ctu @awakecoding Wouldn't that be a temporary concern? As git moves to sha256?

[Jayesh Badwaik @jayesh_badwaik]

@__phantomderp RISC-V and ARM Neon have arbitrary sized vectorizarion. If VLA can support that case/optimization, then it might be useful.

[Jayesh Badwaik @jayesh_badwaik]

@david_picard @Michael_J_Black However, this is not often true as I have found. Different spoken languages actually have different logical models that are not compatible. For example, Germany doesn't have any real equivalent of present continuous. (There's gerade, but it's not really the same).

[Jayesh Badwaik @jayesh_badwaik]

@ProjectPhysX @FluidX3D Very cool 😎