Post

More from @lauriewired

LaurieWired

@lauriewired

Jun 3

A squadron of F22’s was once taken out by an imaginary line.

On a mission to Japan, an unforeseen software bug occurred crossing the international date line. Longitude swaps from W179.99 to E180 degrees.

Navigation, comms, and even fuel management went down!

This wasn’t a simple "turn it on and off again” fix; something was seriously wrong. Reboots weren't helping.

According to Maj. Gen. Sheppard:

“…all systems dumped and when I say all systems, I mean all systems...they could have been in real trouble."

Thankfully, the squad of 12 F22’s were accompanied by a KC-10 tanker, who they followed visually back to Hawaii safely.

Details are sparse, but these issues aren't uncommon in aviation!

It's extremely difficult to notice subtle bugs over millions of lines of code.

Read 4 tweets

LaurieWired

@lauriewired

May 27

Want to recognize a song from just a few seconds of distorted audio?

Use Constellation Maps.

The math is brilliantly simple.

With just a handful of bytes; discarding 99% of the waveform, you can recognize a unique fingerprint across hundreds of millions of tracks.

First, chop up the audio into few-second windows.

Take an FFT of the waveform, then extract the local peaks. Each maximum becomes a “star” on an xy plot of time vs frequency.

Pair nearby stars into clusters and hash the result. Boom, a noise-resistant fingerprint.

Constellation maps were the basis for Shazam.

Shazam today is a polished iOS/Android app, but the tech actually started back in 2003!

The early marketing was kind of hilarious; dial a phone number and get your Nokia to "listen". The ID'd track was sent back via SMS.

Read 4 tweets

LaurieWired

@lauriewired

May 23

I miss the insanity of 80s processor design.

Intel’s iAPX 432 was a “micromainframe”.

It had no general purpose registers, supported object orientation *directly*, and performed garbage collection on-chip.

It was also 23x slower than an 8086. Here's why it failed.

Intel targeted Ada so aggressively that C support was an afterthought.

Problem was, particularly at the time, the Ada compiler was extremely untuned and immature.

Scalar instructions were basically never used; *everything* was huge object-oriented calls.

The “micromainframe” moniker wasn’t just marketing. One I/O chip could stitch together 63 CPUs on a single bus.

Essentially memory safe in-hardware; dangling pointers were impossible at the ISA level.

Partners like BiiN suggested using the CPU for nuclear-reactor control.

Read 4 tweets

LaurieWired

@lauriewired

May 22

NTIRE is the coolest conference you’ve never heard of.

Deleting motion blur? Sure.
Night Vision? No problem.

Every year, labs compete on categories like hyperspectral restoration, satellite image enhancement, even raindrop removal (think car sensors)! Some highlights ->

Low-light enhancement is always popular.

Retinexformer, shown here got 2nd place in the 2024 contest.

A *TINY* transformer-based model, it runs in about 0.5 seconds for a 6K image on a single 3090. Only 1.6M parameters (<2MB weights at INT8)!

Maybe motion blur removal is more your thing.

UAVs are often used to examine wind turbine blades for early failure warning. Movement of drone + rotational velocity pose a challenge.

Here’s the 2021 winner DeblurGANv2, taking ~0.19s of processing per image.

Read 5 tweets

LaurieWired

@lauriewired

May 21

What if an OS fit entirely inside the CPU’s Cache?

Turns out we’ve been doing it for decades.

CNK, the OS for IBM’s Blue Gene Supercomputer, is just 5,000 lines of tight C++.

Designed to “eliminate OS noise”, it lives in the cache after just a few milliseconds of boot.

Kernels that “live” in the cache are common for HPC.

Cray’s Catamount microkernel (~2005) used a similar method for jitter free timing.

Huge Pages, Statically Mapped Memory, and a lack of scheduling are all typical aspects of these systems.

What about the modern era?

Modern CPUs are *insane*.

L3 sizes exceed GIGABYTES per socket (see Genoa).

Many HPC labs run the hot path in light kernels (LWKs), outsourcing file I/O and syscalls to separate nodes; all with the intent of reducing µs-level jitter. Determinism is the name of the game.

Read 5 tweets

LaurieWired

@lauriewired

May 13

TDP (Thermal Design Power) of CPUs is a garbage metric that misleads consumers.

In the Pentium era, a 89W TDP meant just that; expect to dissipate 89W of heat in the worst case.

With Alder Lake, a 125W CPU can draw ~241W indefinitely!

Here's the goofy math:

CPU’s didn’t really know how to idle until the early 2000s. They just kinda ran full bore all the time.

With the introduction of C-States, various parts of the processor could be shut down, saving power when the computer was doing nothing.

Of course, this was HUGE for laptops.

If you temporarily downclock a CPU, why not upclock it as well?

Around ~2008, Intel came up with the concept of Turbo Boost and the “energy bucket”.

Short spikes *above* TDP are allowed, as long as the 28 second moving average stays <TDP.

Okay, nothing too crazy yet.

Read 5 tweets

Share this page!

Enter URL or ID to Unroll

LaurieWired

Try unrolling a thread yourself!

More from @lauriewired

LaurieWired

LaurieWired

LaurieWired

LaurieWired

LaurieWired

LaurieWired

Did Thread Reader help you today?

Don't want to be a Premium member but still want to support us?

Send Email!