Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Qwen3.7-Max Ran for 35 Hours on Unknown Hardware and Achieved a 10× Speedup

firethering.com

Post reply on HN
Print view View original

11–20 of 33 posts

Re: Qwen3.7-Max Ran for 35 Hours on Unknown Hardware and Achieved a 10× Speedup

#11

Post by userbinator »

keyle wrote: post #7

I don't doubt that it did it but I wouldn't want to maintain whatever it ended up spewing after 35 hrs. In my experience, AI fixes problems by mostly adding more code. It's a short term gain for a long term hurt.

In my experience, AI fixes problems by mostly adding more code.

In my experience, humans unfortunately tend to do the same.

Re: Qwen3.7-Max Ran for 35 Hours on Unknown Hardware and Achieved a 10× Speedup

#12

Post by Mashimo »

l23k4 wrote: post #6

LLM written. See the authors twitter, he speaks english at a rather basic level and certainly did not write this https://x.com/mohitgeryani/with_replies

Also I'm pretty sure the original source was linked here on HN before.

Re: Qwen3.7-Max Ran for 35 Hours on Unknown Hardware and Achieved a 10× Speedup

#13

Post by bahmboo »

keyle wrote: post #7

I don't doubt that it did it but I wouldn't want to maintain whatever it ended up spewing after 35 hrs. In my experience, AI fixes problems by mostly adding more code. It's a short term gain for a long term hurt.

How do you know that? What information do you have that would explain your position? We are talking about a specific circumstance and you have brought unsupported generalities to the discussion.

Re: Qwen3.7-Max Ran for 35 Hours on Unknown Hardware and Achieved a 10× Speedup

#15

Post by dist-epoch »

yjftsjthsd-h wrote: post #2

Obligatory: Either written by AI or by a human who has spent so much time with AI that they adopted its writing style. Anyways. > Over 35 hours it performed 432 kernel evaluations. Each cycle meant writing code, compiling it, running it, reading the profiling output, deciding what to change, and trying again. The model diagnosed compilation failures it hadn’t seen before, identified performance bottlenecks through ru…

Genetic algorithm is random. This is intelligent evolution. Big difference.

Re: Qwen3.7-Max Ran for 35 Hours on Unknown Hardware and Achieved a 10× Speedup

#16

Post by Iolaum »

userbinator wrote: post #11
keyle wrote: post #7

I don't doubt that it did it but I wouldn't want to maintain whatever it ended up spewing after 35 hrs. In my experience, AI fixes problems by mostly adding more code. It's a short term gain for a long term hurt.

In my experience, AI fixes problems by mostly adding more code. In my experience, humans unfortunately tend to do the same.

The LLM's had to learn that from somewhere :p

Re: Qwen3.7-Max Ran for 35 Hours on Unknown Hardware and Achieved a 10× Speedup

#17

Post by mosselman »

what a nonsense, generated, article.

> For context: GLM 5.1 ran the same task and reached 7.3x. Kimi K2.6 reached 5x. DeepSeek V4 Pro reached 3.3x. The models that stopped early did so because they issued no tool calls for five consecutive rounds, they concluded they couldn’t make further progress and stopped. Qwen3.7-Max didn’t stop.

By this reasoning I could release a model that lacks all the basic optimisations. Have it optimise itself for hours to reach 20x the throughput and then claim that the model is superior to the others?

I am not saying that is what happened here, but the reporting is abysmal.

Re: Qwen3.7-Max Ran for 35 Hours on Unknown Hardware and Achieved a 10× Speedup

#18

Post by skew-aberration »

I've had a very similar experience optimising a hidden markov model prediction tool I work on. I wanted to experiment with an alternative architecture and data structures. Opus 4.7 did the refactor, and eventually the only hot spot became the maths kernel. Over the course of an hour or two it iteratively rewrote that code using all the usual optimisations to improve branching, cache usage, vectorisation, etc. It reviewed the disassembly and the hardware counters with perf to verify that the changes were working as intended. It could have taken me several days to cover that much ground doing low level optimisations - and I would have spent most of it grappling with gcc, perf, searching for information about particular SIMD instructions, etc.

Re: Qwen3.7-Max Ran for 35 Hours on Unknown Hardware and Achieved a 10× Speedup

#19

Post by hobofan »

mosselman wrote: post #17

what a nonsense, generated, article. > For context: GLM 5.1 ran the same task and reached 7.3x. Kimi K2.6 reached 5x. DeepSeek V4 Pro reached 3.3x. The models that stopped early did so because they issued no tool calls for five consecutive rounds, they concluded they couldn’t make further progress and stopped. Qwen3.7-Max didn’t stop. By this reasoning I could release a model that lacks all the basic optimisations. H…

They didn't optimize their own kernels and optimize their own runtime, which I think is what you are implying.

Re: Qwen3.7-Max Ran for 35 Hours on Unknown Hardware and Achieved a 10× Speedup

#20

Post by big-chungus4 »

mosselman wrote: post #17

what a nonsense, generated, article. > For context: GLM 5.1 ran the same task and reached 7.3x. Kimi K2.6 reached 5x. DeepSeek V4 Pro reached 3.3x. The models that stopped early did so because they issued no tool calls for five consecutive rounds, they concluded they couldn’t make further progress and stopped. Qwen3.7-Max didn’t stop. By this reasoning I could release a model that lacks all the basic optimisations. H…

It optimized the Extend Attention operator in triton. All models were optimizing the same operator
Post reply on HN