Is it normal to spend 10minutes on tuning nowadays? Do we need to spend another 10 minutes upon changing the code?

You mean autotune? I think 10 minutes is pretty normal, torch.compile('max-autotune') can be much slower than that for large models.

Add to that it can be done only once by developers before distribution for major hardware. Configs saved. Then on client side selected.

It's likely that Swift compiler is using LLVM LIT (https://llvm.org/docs/CommandGuide/lit.html), which is implemented in python, as the test driver

Python and LIT are used heavily to build and test the compiler, but that is only for building it, you do not need it to download and use the built toolchain. The python dependency is more about its use in LLDB.

> In the end, programs will want probably to stay conservative and will implement only the core ISA

Unlikely, as pointed out in sibling comments the core ISA is too limited. What might prevail is profiles, specifically profiles for application processors like RVA22U64 and RVA23U64, which the latter one makes a lot more sense IMHO.

Come on, what was to be understood is to 'stick to the core ISA' as much as possible.

I had to clarify the obvious: if a program does not need more than a conservative usage of the ISA to run at reasonable speed, no hardcore change to the hardware should be investigated.

Additionnally, the 'adding new machine instructions' fan boys tend to forget about machine instruction fusion (they probably want they names in the extension specifications) which has to be investigated first, and often in such niche cases, it may be not the CPU to think about, but specialized ASIC blocks and/or FPGA.

yes, it has been done for at least a decade if not more

> Even more of a wild idea is to pair up two cores and have them work together this way

I don't think that'll be profitable, because...

> When you have a core that would have been idle anyway

...you'll just schedule in another process. Modern OS rarely runs short on available tasks to run

The article is easy to follow but I think the author missed the e point: branchless programming (a subset of the more known constant time programming) is almost exclusively used in cryptography only nowadays. As shown by the benchmarks in the article, modern branch predictors can easily achieve over 95% if not 99% precision since like a decade ago

yes, the short answer is LLVM uses RegPressureTracker (https://llvm.org/doxygen/classllvm_1_1RegPressureTracker.htm...) to do all those calculations. Slightly longer answer: I should probably be a little more specific that in most cases, Machine Scheduler cares more about register pressure _delta_ caused by a single instruction, either traverses from bottom-up or top-down. In which case it's easier to make an estimation when some of other instructions are not scheduled yet.

The scenario suitable for TBM is surprisingly limited so even nowadays many tunnels are still dig using the good’o way, mostly with explosives

??? explosives? in mud? (well, clay)?!

tunnels are bored depending on the soil, length, diameter. Most projects I have seen use TBMs and the New Austrian Tunneling Method. Explosives are quite the minority (not many tunnels are in solid rock), even the Gotthard tunnels were dug with TBMs

Came to say Apple also did a great job on tagging my bois who are both grey-ish cats, even in pictures they faced backward, no idea how they did that

What I found impressive was that Apple Photos, given pictures of my cousins when they were 50 or more years old, was able to identify pictures of them as kids. On the other hand, it could never consistently distinguish between my two older brothers (although to be fair, they were identical twins). It also insists that a beagle I once owned was a cat. I mean, sure, he sometimes slept on his back with his paws in the air like a cat, but he was all dog.

On the other hand, it has no understanding of time. I have thousands of photos of me from the 1970s up through today, and Apple Photos is remarkably good at identifying me in all of them. And yet when my daughter was born it started identifying her, as a baby, as me. You'd think you could build a model to grasp the idea that a photo of a baby taken in 2015 is probably not of me.

I raise you the photo of a photo edge case.

Metadata is 2015, photo is 1960

But you shouldn't optimize for the edge case! (All of my 20k+ photos, dating back to 1905, have correct metadata + GPS).

I would argue that AI is exactly how you should handlr these edge cases, but likely a fine tuned model.

There would be hints that a photo is from the 60s vs 15s, a human would be able to tell in many cases even without other context.

That is exactly the use case AI is meant to excel at, something that is arguably hard to do algorithmically but is possible for an ML model

I hate to say it, but you are the edge case. Most users are not fixing dates of photos (especially pre-digital scans) or adding GPS data to photos which didn’t originally have it.

I'd suspect there are more people removing metadata from photos that have it than are adding metadata to photos that don't have it.

A photo from 1960 was scanned at <some later date>.

I thought the original comment meant “_for_ who doesn’t use X or Discord, here is the github mirror link”. There’s a “for” missing, and thus I think they agree with you

second this, it didn't show anything when I hovered over North America