In the last couple of days I wanted to try out the new definitive DeepSeek v4 releases. I gave it the repository of a semi-abandoned video compression codec and I told it to perform the usual benchmark -> profile -> verify -> research -> improve loop. I specifically chose this codec because the authors include a verifier for the bitstream to make sure you don't break stuff if you want to try your own implementation. I gave the agents access to the compiler's profiler and also Intel's VTune, which has fantastic output. In a couple of hours the LLM generated SSE and AVX implementations of the compression and decompression algorithms that almost doubled performance with a single core. Then I asked it to create a CUDA implementation using NVIDIA's NSIGHT profiler as a guide and it also started doing some good work.
Personally, I believe that LLMs should be treated like an advanced version of Prolog or linear programming: you give the constraints, you have a way of verifying correctness, and you give it a clear goal. If the LLM can verify itself and course-correct you can basically leave it on autopilot
I tried kernel autoreasearch using DeepSeek-V4-Flash as well. It spent about 1-2 hours to complete the FlashAttention optimization job (https://github.com/fengwang/FA5090/tree/main/v7) and cost me only $0.2. I believe we are ready to offload a lot of this kind well-defined constrained optimization problems to AI Agent autoresearch.
I've had a lot of success decompiling old video game ROMs in exactly this way. Like you say - give it a way of verifying correctness - put it in a loop - and they are quite surprising.
Even the cheap LLMs are great in doing the awful crud work in the beginning: finding offsets, firmware update file structures, brute forcing checksums, etc.
It still produces a lot of crap in the later steps (understanding the implementation itself) but I'm happy doing this stuff myself then.
> It still produces a lot of crap in the later steps
I've had success here by adding a phase called "grounding" that attempts to verify its "understanding" by creating tests that modify the running executable to ensure its made the right inference.
Is this variable really MARIO_X? Change it and see if Mario moves. Etc.
As an example in Donkey Kong - the system had trouble deciding if an array controlled barrels or fireballs. There was conflicting evidence.
After many trips through the loop - it realized it does BOTH, depending on which level you're on.
I did something similar recently with Google's C# protobuf library. I had spotted I was getting CPU bound rather than memory bandwidth bound when doing streaming of uint32 buffers in dotnet gRPC.
I then asked claude to compare the C#/.NET implementation in the library with the C++ version, and it quickly identified that the C# library was missing a couple of fairly cheap optimisations that were present in the C++ version.
If I can help get a PR merged, then it'll be by far the biggest impact of any work I've ever done.
I also compared the Rust version, it had this specific optimisation. The far more popular Tokio/Prost library did not.
Given appropriate guardrails, LLMs are impossibly fast at iterating to find root causes and specific performance bottlenecks.
You presented another thing LLMs excel at: integrating something from a project that is not present in another one. I think they work so well at this because both the starting and ending points have an already existing structure, so the LLM can guide itself effectively. In your case it's even more egregious because we are talking about the same exact algorithm/functionality implemented in two different, but rather similar, programming languages.
Could you have manually profiled and compared the execution paths? Sure. Could you have translated the C++ optimizations to C#? Sure. But in such an obvious case, the LLM managed itself.
Training material seems to be especially rich re GPU kernels and SIMD.
I wonder if there is extra effort put into this because they are useful for the researchers working on the models or just a sub-domain that language models are a great fit for and humans have trouble with?
Well, GPU kernels are co-designed really hard. A lot of it is, async tile pipelines + spam my MMA primtives.
Obviously it's still hard, but the point is that, by construction (cause like, NVIDIA literally releases primitives like this, and/or people like TK build slightly higher-level primitives over the base hardware primitives), if you learn the complicated language, you can get really good results, and on some level you "know" you're right by construction even before you go to the actual empirical tests (since you're operating over a higher-level "language", and not arbitrary byte accesses).
Honestly a lot of interfaces and frameworks you could argue are like that, so it's not really a point for GPU kernels relative to other things. But maybe a hint as to what I personally think is important in the AI era - finding the right cuts, the right high leverage abstractions, as otherwise AI is going to produce spaghetti nonsense.
Anecdotally, I saw Opus 5 come up with a complicated loop unrolling technique when I asked it to implement a simple biquad in SIMD, missing a simpler solution. Maybe it was a downgraded session, who knows. That SIMD instruction set, the one for the ESP32-P4, is not very popular and all the documentation it has is a couple of blog posts. So I'm pretty sure it has at most seen some code for a predecessor during its training. However, the LLM was able to derive a full listing of the operations and their arguments from gcc to get us started, and that's why I was able to come up with my own implementation. Along the way, it also came up with insights about possible gotchas. Then, when implementing algorithms, it has been able to reason things out and get things working, despite the ISA not being extremely well known.
Because pre-LLMs humans partially "autogenerated" kernels through hyperparameter search and in some sense eating the code complexity in return for performance, and thus built tools for the same automatic verifiability that is useful for LLMs.
In some other tasks, we never built the same level of automatic verifiability since the level of automation in creation being much lower meant it's not giving you as much of a marginal benefit. We prefer code readability and simplicity and such in say, web services, because, say, the database IO time is going to dominate. Here getting an LLM to write a cromulent C# web service is more difficult since it's not easy to automatically verify whether code is cromulent or not. So if you put up LLMs to it, you end up with slop (which works).
OTOH, in kernel design, you give it access to every perf counter, every observable possible and have it optimise all of them. And all are verifiable/hill-climbable
Isn't cholesky - used to substitute householder at a point - faster but less stable in some cases? I'm just recalling from memory since I had done a small project on qr decomposition with householder for an exam this year. I mean, if it is faster than the standard torch operation probably there are good reasons for which it is not the default standard torch operation. Might as well be wrong, I'm not sure
this is the first time ive heard of beam search. i would have reached for a genetic algorithm of some sort, although it seems like some stochastic versions of beam search exist to avoid local minima. i wonder if there are any good frameworks for building these that agents can construct and use.
I think the question is how do you keep track of the ideas that the agent is pursuing - like I was working on this for implementing a fft, and doing the optimizations, but I held its hand and was like - hey let's go back and retry this older thing you discounted because of a 3 % slowdown.
This was nowhere near the top submission. But even if a solo engineer could get a top kernel, you don't think that having thousands of engineers, infinite tokens, and stronger models than are available to the public would give the labs a significant edge?
Does the edge matter? I know you added significant as your hedge, but once you have feedback, your gain is largely irrelevant. Gain buys you bandwidth, so we are constructing systems run by the most powerful corporations where they are now optimizing for latency, as Archer says, do you want to flash crash civilization? This is how you do it.
People are always going to hate auto-research and "loop engineering". Because it's got 2 properties:
1) it's the only way to get something out of models (or people for that matter) that they don't know yet.
2) it's harder to do with an LLM than without. Not easier.
3) and when you fuck it up, half the time the LLM (or other ML technique) makes a fool out of you and you spent $1000 to find the quickest way to get a robot leg on the ground is just to crash it into the ground.
you specify the goal. if the goal is achieved, it's achieved.
the code the LLM writes will be read and maintained and developed further by LLMs. so it doesn't really matter what it produces as long as all the tests are green and it achieves exactly what you want it to achieve.
Every step here has an oracle: wall-clock, the profile, pass or fail from the verifier. I had an agent-built app audited task by task, 10 came back done and 7 worked, and the three misses were the ones needing a credential or a setting on someone else's dashboard. Nothing in the loop could tell the agent it had failed, so it said done and moved on.
In the last couple of days I wanted to try out the new definitive DeepSeek v4 releases. I gave it the repository of a semi-abandoned video compression codec and I told it to perform the usual benchmark -> profile -> verify -> research -> improve loop. I specifically chose this codec because the authors include a verifier for the bitstream to make sure you don't break stuff if you want to try your own implementation. I gave the agents access to the compiler's profiler and also Intel's VTune, which has fantastic output. In a couple of hours the LLM generated SSE and AVX implementations of the compression and decompression algorithms that almost doubled performance with a single core. Then I asked it to create a CUDA implementation using NVIDIA's NSIGHT profiler as a guide and it also started doing some good work.
Personally, I believe that LLMs should be treated like an advanced version of Prolog or linear programming: you give the constraints, you have a way of verifying correctness, and you give it a clear goal. If the LLM can verify itself and course-correct you can basically leave it on autopilot
I tried kernel autoreasearch using DeepSeek-V4-Flash as well. It spent about 1-2 hours to complete the FlashAttention optimization job (https://github.com/fengwang/FA5090/tree/main/v7) and cost me only $0.2. I believe we are ready to offload a lot of this kind well-defined constrained optimization problems to AI Agent autoresearch.
I've had a lot of success decompiling old video game ROMs in exactly this way. Like you say - give it a way of verifying correctness - put it in a loop - and they are quite surprising.
https://github.com/qarl/arcade-js
Same. I love reverse engineering embedded stuff.
Even the cheap LLMs are great in doing the awful crud work in the beginning: finding offsets, firmware update file structures, brute forcing checksums, etc.
It still produces a lot of crap in the later steps (understanding the implementation itself) but I'm happy doing this stuff myself then.
> It still produces a lot of crap in the later steps
I've had success here by adding a phase called "grounding" that attempts to verify its "understanding" by creating tests that modify the running executable to ensure its made the right inference.
Is this variable really MARIO_X? Change it and see if Mario moves. Etc.
As an example in Donkey Kong - the system had trouble deciding if an array controlled barrels or fireballs. There was conflicting evidence.
After many trips through the loop - it realized it does BOTH, depending on which level you're on.
So the "understanding" grows with each iteration.
I did something similar recently with Google's C# protobuf library. I had spotted I was getting CPU bound rather than memory bandwidth bound when doing streaming of uint32 buffers in dotnet gRPC.
I then asked claude to compare the C#/.NET implementation in the library with the C++ version, and it quickly identified that the C# library was missing a couple of fairly cheap optimisations that were present in the C++ version.
If I can help get a PR merged, then it'll be by far the biggest impact of any work I've ever done.
I also compared the Rust version, it had this specific optimisation. The far more popular Tokio/Prost library did not.
Given appropriate guardrails, LLMs are impossibly fast at iterating to find root causes and specific performance bottlenecks.
You presented another thing LLMs excel at: integrating something from a project that is not present in another one. I think they work so well at this because both the starting and ending points have an already existing structure, so the LLM can guide itself effectively. In your case it's even more egregious because we are talking about the same exact algorithm/functionality implemented in two different, but rather similar, programming languages.
Could you have manually profiled and compared the execution paths? Sure. Could you have translated the C++ optimizations to C#? Sure. But in such an obvious case, the LLM managed itself.
This is exactly how I use it. I mean not on abandoned repos, but in a benchmark - profile - verify - research - improve loop.
I'm hoping to release a native Wine for MacOS w/ Win32 FEX support in the next few days.
Watching claude and codex play winquake and age of empires, and debug support for Firefox 52 has been wild.
This is the way. Checkout the technique mentioned in the alphaevolve paper and see if it works well for your setting.
Thank you for the reference, I hope to be smart enough to try it out!
Meta commentary but it felt fresh to read a long wall of text that didn't seem to be AI generated. Thanks.
Training material seems to be especially rich re GPU kernels and SIMD.
I wonder if there is extra effort put into this because they are useful for the researchers working on the models or just a sub-domain that language models are a great fit for and humans have trouble with?
Well, GPU kernels are co-designed really hard. A lot of it is, async tile pipelines + spam my MMA primtives.
Obviously it's still hard, but the point is that, by construction (cause like, NVIDIA literally releases primitives like this, and/or people like TK build slightly higher-level primitives over the base hardware primitives), if you learn the complicated language, you can get really good results, and on some level you "know" you're right by construction even before you go to the actual empirical tests (since you're operating over a higher-level "language", and not arbitrary byte accesses).
Honestly a lot of interfaces and frameworks you could argue are like that, so it's not really a point for GPU kernels relative to other things. But maybe a hint as to what I personally think is important in the AI era - finding the right cuts, the right high leverage abstractions, as otherwise AI is going to produce spaghetti nonsense.
Anecdotally, I saw Opus 5 come up with a complicated loop unrolling technique when I asked it to implement a simple biquad in SIMD, missing a simpler solution. Maybe it was a downgraded session, who knows. That SIMD instruction set, the one for the ESP32-P4, is not very popular and all the documentation it has is a couple of blog posts. So I'm pretty sure it has at most seen some code for a predecessor during its training. However, the LLM was able to derive a full listing of the operations and their arguments from gcc to get us started, and that's why I was able to come up with my own implementation. Along the way, it also came up with insights about possible gotchas. Then, when implementing algorithms, it has been able to reason things out and get things working, despite the ISA not being extremely well known.
They are easily verifiable and hill-climbable.
Because pre-LLMs humans partially "autogenerated" kernels through hyperparameter search and in some sense eating the code complexity in return for performance, and thus built tools for the same automatic verifiability that is useful for LLMs.
In some other tasks, we never built the same level of automatic verifiability since the level of automation in creation being much lower meant it's not giving you as much of a marginal benefit. We prefer code readability and simplicity and such in say, web services, because, say, the database IO time is going to dominate. Here getting an LLM to write a cromulent C# web service is more difficult since it's not easy to automatically verify whether code is cromulent or not. So if you put up LLMs to it, you end up with slop (which works).
OTOH, in kernel design, you give it access to every perf counter, every observable possible and have it optimise all of them. And all are verifiable/hill-climbable
Isn't cholesky - used to substitute householder at a point - faster but less stable in some cases? I'm just recalling from memory since I had done a small project on qr decomposition with householder for an exam this year. I mean, if it is faster than the standard torch operation probably there are good reasons for which it is not the default standard torch operation. Might as well be wrong, I'm not sure
some of these submissions seem to be omitting the actual rules. the #1 on edinh has a line that says "bypass ban check"
This is really cool - I really like the beam search idea,
this is the first time ive heard of beam search. i would have reached for a genetic algorithm of some sort, although it seems like some stochastic versions of beam search exist to avoid local minima. i wonder if there are any good frameworks for building these that agents can construct and use.
I think the question is how do you keep track of the ideas that the agent is pursuing - like I was working on this for implementing a fft, and doing the optimizations, but I held its hand and was like - hey let's go back and retry this older thing you discounted because of a 3 % slowdown.
Damn! If a solo engineer can do this, it makes the most around OAI/Anthropic start to look pretty weak.
This was nowhere near the top submission. But even if a solo engineer could get a top kernel, you don't think that having thousands of engineers, infinite tokens, and stronger models than are available to the public would give the labs a significant edge?
That’s not what a moat is =]
Does the edge matter? I know you added significant as your hedge, but once you have feedback, your gain is largely irrelevant. Gain buys you bandwidth, so we are constructing systems run by the most powerful corporations where they are now optimizing for latency, as Archer says, do you want to flash crash civilization? This is how you do it.
I don't know. That just sounds like throwing money at a problem until it goes away. I'm not convinced that is the correct path forward.
People are always going to hate auto-research and "loop engineering". Because it's got 2 properties:
1) it's the only way to get something out of models (or people for that matter) that they don't know yet.
2) it's harder to do with an LLM than without. Not easier.
3) and when you fuck it up, half the time the LLM (or other ML technique) makes a fool out of you and you spent $1000 to find the quickest way to get a robot leg on the ground is just to crash it into the ground.
What do you mean by 2?
LLMs will take shortcuts and cheat in ways that a human would realize are counterproductive.
you specify the goal. if the goal is achieved, it's achieved.
the code the LLM writes will be read and maintained and developed further by LLMs. so it doesn't really matter what it produces as long as all the tests are green and it achieves exactly what you want it to achieve.
I've seen this several times now. Disturbing.
IMO, LLMs will be a dead end to anything close to AGI because of this and hallucinations.
We're missing something in the mix, which I suspect is some kind of advanced JEPA model.
LLMs take shortcuts if allowed.
Humans do it ignorantly.
The LLMs will improve while average human IQ in the west dips closer and closer to the 80s on the global scale.
Every step here has an oracle: wall-clock, the profile, pass or fail from the verifier. I had an agent-built app audited task by task, 10 came back done and 7 worked, and the three misses were the ones needing a credential or a setting on someone else's dashboard. Nothing in the loop could tell the agent it had failed, so it said done and moved on.
I’ve tried it on simple UI tasks. Give it screenshot to work towards (or figma MCP), let it get screenshots from chrome to check its work.
The existing models are surprisingly bad at it.
It's really difficult to understand what your definition of surprisingly bad is. What was it continuously having problems with?
Achieving a 232x speedup on a kernel via automated tooling is an incredible engineering feat. Fascinating read on optimization.