O3marchnative

Out of curiosity, did you try out any other portable simd crates before landing on fearless simd?

This is quite amazing. Fearless SIMD is what finally allowed my FFT crate, PhastFT, to run on stable Rust [0]. PhastFT previously depended on Rust’s std::simd (as well as multiversion), which still requires nightly [1].

One of the main contributors to v1.0 of fearless_simd started out by helping me evaluate portable SIMD crates for PhastFT. Once we landed on fearless_simd, he ported PhastFT from std::simd to fearless_simd. We did find that more functionality was needed than fearless_simd offered at the time. So, he started contributing significantly to fearless_simd. It’s really rewarding to see how working on a hobby open source project can help improve the Rust ecosystem.

[0] https://github.com/smu160/PhastFT

[1] https://doc.rust-lang.org/std/simd/index.html

That makes the issue worse as the actual app/device likely just relies on querying some NTP pool for accurate time. Then that would mean the actual stratum 1 servers are not working right.

I was thinking of this in terms of aerial refueling.

It's always better to take off with less than a full tank of fuel. Once you're airborne, the aircraft can top off its tank to complete its journey. Of course, it may even be more economical to top off at multiple legs as the journey proceeds. To your point, a "reverse" refueling would also be ideal in order to lower the weight of the aircraft.

In certain contexts, it's already being done. It'd likely need to be much safer in order for civilian aircraft to carry out aerial refueling though.

I'm reminded of the OPM breach back in 2015 [0]. Practically everyone that even applied for a security clearance was compromised. In addition, millions of sets of fingerprints were recovered by the entity that carried out the hack.

[0] https://en.wikipedia.org/wiki/2015_Office_of_Personnel_Manag...

I've been considering upgrading from 1 Gbps to 5 Gbps. Have you noticed any significant benefits for any particular use cases?

The author mentions Rust's portable SIMD library [0]. The only issue with portable SIMD is it's only available on nightly. I used it in my FFT crate, but we had to switch to the fearless_simd crate in order to get a portable SIMD solution that works on stable [1].

[0] https://doc.rust-lang.org/std/simd/index.html

[1] https://github.com/linebender/fearless_simd

This argument usually depends on how you quantify defense spending, no? [0] In terms of PPP, the U.S. is falling behind. To be perfectly clear, that doesn't somehow absolve the size of U.S. defense spending. If anything, it's damning of all these wunderwaffe projects defense contractors take on and the taxpayer funds.

[0] https://www.economist.com/graphic-detail/2026/05/11/by-one-m... (no paywall: https://archive.ph/X9zzc)

Going back to Clausewitz, war is a continuation of politics through other means. Moreover, victory is achieved by using force to compel one's opponent to carry out one's will.

> intact and combat capable

On its own, this claim is likely true. Nevertheless, this does not mean one's opponent is going to carry out one's will. Hence, one cannot declare victory. It follows that anything short of achieving one's goal is a loss. A loss doesn't have to be complete and utter destruction. It's fair to say that Argentina did not achieve its strategic objectives during the Falklands War, no?

> Losing is what happened to Germany after both World Wars.

Yes, but Germany is not as good of an example in this case as Japan during WWII. Japan would have likely accepted something short of unconditional surrender much sooner than August 1945, but the U.S. insisted on unconditional surrender. As such, the U.S. did not win until Japan agreed to unconditional surrender.

I always remember it as the opposite of what I think is typically called the granny knot. In the case of the square knot, it's just right over left, then left over right.

Have you taken a look at Iran's targeting capabilities and the assets they were able to destroy? More importantly, have you considered who facilitated those targeting capabilities?

https://www.hudson.org/national-security-defense/war-above-w...

This one and the square knot are indispensable

> This is a foundational principle of computer science

How exactly is this a foundational principle of computer science?

Wow. I was not expecting that topic when I clicked on the link you provided.

When I went through such a selection process years ago, we took all sorts of tests before and even after the selection process. Towards the end, the head instructor told us they don't really have a good way to measure who will make it through. What he did tell us though is that top physical fitness test scores were not indicative that a candidate will make it to the end.

Is there a PDF version or instructions for building your thesis? I'd like to read it.

> There were engineering questions around how to purify uranium and how to actually construct the weapon etc. But the phenomenon was known.

FWIW, constructing a weapon with highly enriched uranium is, relatively, simple. At the time, the choice was made to use a gun-type weapon that shot a projectile of highly enriched uranium into a a "target" of highly enriched uranium. The scientists were so sure it would work that the design didn't necessitate a live test. This was "little boy", which was eventually dropped on Hiroshima.

Fat Man utilized plutonium which required an implosion to compress the fissile material that would set off the chain reaction. This is a much more complex undertaking, but it's much more efficient. Namely, you need much less fissile material, and more of that fissile material is able to participate in the chain reaction. This design is what allows for nuclear tipped missiles. The same principles can be applied to a U-235 based weapon as well.

The implosion based design is super interesting to read about. One memorable aspect is that the designers realized that applying a tamper of uranium (U-238) around the fissile material allows for significant improvement in yield. The chain reaction is exponential, so the few extra nanoseconds that the uranium keeps the fissile material together leads to significant increase in yield.

https://en.wikipedia.org/wiki/Little_Boy

https://en.wikipedia.org/wiki/Fat_Man

The Royal United Services Institute (RUSI) has an updated tally on defensive and offensive munition expenditures. It's likely not 100% accurate due to the sensitive nature of those figures.

> 11,294 munitions in the first 16 days of the conflict, at a cost of approximately $26 billion.

Several detailed tables are in the link below.

https://www.rusi.org/explore-our-research/publications/comme...

Author here. The $75M is specifically for Ground Based Interceptors (GBIs). This is the U.S.'s ICBM mid-course interceptor. There are other interceptor types in the current U.S. arsenal:

Patriot PAC-3 (~$4M): Nations burnt through 600-800 in the first few days of Operation Epic Fury. There are reports that they're being used for drone defense.

SM-3 (~$10-30M): Ship-launched

SM-6 (~$4-5M): Ship-launched

THAAD (~$12-15M): Terminal phase, high altitude

GBI (~$75M): intended for interception of ICBMs (reported as the hardest type of missile to intercept)

Each type of interceptor is optimal for certain type of threats, which is yet another constraint on the optimization problem.

The recently announced "Golden Dome" project intends to get around this issue by putting a vast constellation of satellites into orbit. Each satellite would likely need a serious source of power in order to use its laser. Assuming that's just an engineering problem, then the issue becomes coverage. That is, depending on the adversary's capabilities, you'd need an absolutely massive constellation in orbit [0].

[0] https://en.wikipedia.org/wiki/Golden_Dome_(missile_defense_s...

> For ICBMs, one idea was to use orbital, nuclear powered lasers to hit the missile on the boost phase.

Author here. Thank you for your insight.

I took some time to read about the recently proposed "Golden Dome" defense system, and what you laid out seems to be the end goal [0]. It's difficult to tell how realistic this actually is. The size of the constellation of satellites needed seems prohibitive, to say the least.

[0] https://armscontrolcenter.org/fact-sheet-golden-dome/

[1] https://www.defenseone.com/ideas/2026/02/space-based-interce...

That's also great for letting the compiler unlock auto-vectorization opportunities without delving into the world of manual SIMD.

Even storing something simple such as an array of complex numbers as a structure of arrays (SoA) rather than an array of structures (AoS) can unlock a lot of optimizations. For example, less permutes/shuffles and more arithmetic instructions.

Depending on how many fields you actually need when you iterate over the data, you prevent cache pollution as well.

Hi,

One of the authors of PhastFT here. Thank you for your interest.

We went out of our way to configure FFTW for AVX-512. The Rust bindings don't do it, but the FFTW itself in the benchmark does.

It's worth noting that with FFTW you have to choose between building it for your CPU and making it non-portable, or targeting the lowest common denominator of CPU features so that it runs everywhere but much slower. Meanwhile PhastFT detects the available CPU features at runtime, and will utilize the fastest CPU features without sacrificing portability.

Lastly, we are currently working on support for interleaved format [1]. That should ship in the next release.

[1] https://github.com/QuState/PhastFT/pull/27