Hacker News: cec

New comment by cec in "Ask HN: What Are You Working On? (July 2026)"

cec — Mon, 13 Jul 2026 01:45:06 +0000

I'm not a big fan of the encroachment of AI into Adobe's apps, so I'm using AI to build a replacement for those apps (a small web-based photo organizer and editor, just the tiny subset of the Lightroom features I need for my workflow)

New comment by cec in "Meta LLM Compiler: neural optimizer and disassembler"

cec — Fri, 28 Jun 2024 20:17:07 +0000

Hey! The idea isn't to replace the compiler with an LLM, the tech is not there yet. Where we see value is in using these models to guide an existing compiler. E.g. orchestrating optimization passes. That way the LLM won't break your code, nor will the compiler (to the extent that your compiler is free from bugs, which can tricky to detect - cf Sec 3.1 of our paper).

New comment by cec in "Show HN: Paisa – Open-Source Personal Finance Manager"

cec — Sat, 23 Sep 2023 02:15:57 +0000

Is a MacOS build of YNAB 4 still available?

New comment by cec in "Large Language Models for Compiler Optimization"

cec — Mon, 18 Sep 2023 21:09:09 +0000

We use the same architecture as other LLMs, but we include no natural language in our pretraining. We figured a single-domain training corpus would make evaluation easier. We’ll be looking at layering this on top of something like Code Llama next

New comment by cec in "Large Language Models for Compiler Optimization"

cec — Mon, 18 Sep 2023 21:04:27 +0000

Hey, we’re targeting code size in this work, not runtime performance. You would use an option like -O3 to optimize for runtime and -Oz to optimize for code size. The pass pipelines are different for both

New comment by cec in "Large Language Models for Compiler Optimization"

cec — Mon, 18 Sep 2023 21:00:27 +0000

Program synthesis is part of the loss function, which is what makes it a auxiliary learning task.

We haven’t experimented with model size yet, we just used the same configuration as the smallest Code Llama. We did play with dataset size and found thah performance tracks the usual scaling laws. Details in the paper

New comment by cec in "Large Language Models for Compiler Optimization"

cec — Mon, 18 Sep 2023 20:56:51 +0000

There’s a researcher at Edinburgh who is looking into this. You may want to reach out. Check the related work section in this paper

New comment by cec in "Large Language Models for Compiler Optimization"

cec — Mon, 18 Sep 2023 17:33:15 +0000

Hey Jon, we don't change semantics. We just choose optimization passes to run to match a particular input code. Agreed that code size wins could regress performance. We don't measure that yet, but will be looking into it next. There are some applications where codesize above-all-else is what to optimize for.

New comment by cec in "Large Language Models for Compiler Optimization"

cec — Mon, 18 Sep 2023 17:31:01 +0000

Hey, yes that's right, and good callout on GEP instructions. We admit that instruction count is a bit handwavy, but we use it as a starting point as that's what the prior works we compare against optimize for. We'll be looking at true binary size next, and code runtime after.

New comment by cec in "Large Language Models for Compiler Optimization"

cec — Mon, 18 Sep 2023 15:03:03 +0000

Hey, author here. We use machine learning to generate a list of _optimization passes_ for the compiler to run. These optimization passes give us a 3.0% improvement over the default (-Oz), and are what generates correct code. We don't do anything to the code that breaks semantics (assuming no bugs in the compiler passes; some nuance is needed here ;) ).

We also train the model to generate what it thinks the optimized code will look like. We find that this helps the model choose better pass lists, but obviously the code cannot be trusted and semantics are not guaranteed. It only compiles in 91% of cases. "Perfectly emulating the output of the compiler" means the model spat out code that is character-for-character identical to what the compiler generates with the given pass list (even choosing the same variable names etc). IMO this is no mean feat, but is still a long way to go before using LLMs for codegen. We provide a bunch of examples in the paper of things that LLMs can and cannot do.

New comment by cec in "CompilerGym: A toolkit for reinforcement learning for compiler optimization"

cec — Wed, 03 Feb 2021 00:25:06 +0000

This. For now we rely on differential testing against a gold-standard implementation (e.g. unoptimized). For the action space we expose, any semantics-breaking change induced by our tool is a compiler bug.

New comment by cec in "Deep Learning for Compilers"

cec — Tue, 01 Sep 2020 21:44:16 +0000

Author here. In the sense that a trained neural net produces the same output from the same input, this is deterministic. But I don’t think that’s what you’re getting at. Where it gets interesting is if we inserted a feedback loop (such as FDO) so that the compiler could fine tune itself on past decisions. In that case, the build would still be deterministic, but the compiler would keep changing, invalidating the build cache.

New comment by cec in "Deep Learning for Compilers"

cec — Tue, 01 Sep 2020 16:33:32 +0000

https://github.com/chriscummins/phd

New comment by cec in "Pik – a new lossy image format for the internet"

cec — Mon, 24 Jul 2017 19:41:02 +0000

I'd be careful about asking your co-workers for "dick pics".

New comment by cec in "Machine Learning"

cec — Wed, 17 May 2017 22:47:49 +0000

No, HI is Brady Haran and CGP Grey, though I suspect Brady may have collaborated with Tom.

New comment by cec in "An In-Depth Look at Google's Tensor Processing Unit Architecture"

cec — Thu, 06 Apr 2017 09:06:04 +0000

The paper says "we started a high-priority project to quickly produce a custom ASIC for inference (and bought off-the-shelf GPUs for training)."

New comment by cec in "Ask HN: What LaTeX editor do you use?"

cec — Sun, 25 Sep 2016 00:45:07 +0000

Give texstudio a try. I find it offers exactly the kind of IDE-style features that I was missing from Emacs and other editors.

New comment by cec in "A linguistics tool to help you simplify your writing"

cec — Thu, 02 Jun 2016 15:51:36 +0000

For offline writing (e.g. emacs/vim), I've found textlint[1] to work really nicely, and fits well into a LaTeX+Makefile workflow.

[1] https://github.com/textlint/textlint

New comment by cec in "Universal Install Script by XKCD – Combat Test"

cec — Tue, 19 Apr 2016 11:58:59 +0000

No. && would cause the next command to begin once the previous had successfully completed. A successful install shoudn't start the next install attempt, hence ||.

On the other hand, OP and this comment miss the beauty of Randall's implementation, which is each operation is launched asynchronously ('&'), so it isn't the case that you try a few approaches one at a time, i's that you try all approaches at the same time.

New comment by cec in "Elon Musk needs a vacation"

cec — Mon, 05 Oct 2015 20:05:46 +0000

I'm not sure if I can agree with that reasoning. Sure, no-one is impressed with an Apple II now, but that doesn't degrade how significant it was then. By that measure, did the Wright bothers really need to put in all of that time since today we have 747s?