undefined | Better HN

0 pointsjiggawatts2y ago0 comments

Read the article you linked.

It literally says that the GPU is deterministic, the NVIDIA libraries on top are deterministic, but it is Tensorflow that introduces variability (errors!) for “performance”.

My argument is that it is the AI/ML code that is introducing non-determinism, usually by sacrificing repeatability to gain performance.

That's precisely what's happening here. Tensorflow introduced a "harmless"[1] data race to improve performance by not having to use a deterministic but slower algorithm.

The individual floating point computations are deterministic, it's the multi-threaded design on top that's introducing the variability in the output.

[1] Used to be harmless, but cutting corners like this will make it nigh impossible to repeatably validate the safety of future models like GPT5. That seems pretty dangerous...

0 comments

5 comments · 1 top-level

n2d42y ago· 4 in thread

As the article says, cuBLAS is deterministic, but other CUDA primitives (eg. some of those in cudNN) are not.

Yes, the non-determinism is being introduced somewhere, but that is splitting hairs. The point is that the primitives that you work with on GPUs are non-deterministic by design.

I mostly take issue with you phrasing it as a bug and using it to insult the authors.

xyzzy_plugh2y ago

How is that splitting hairs?

> The point is that the primitives that you work with on GPUs are non-deterministic by design.

This is just blatantly wrong. There are _some_ operations that can be non-deterministic in some scenarios but they are not necessary.

GPUs are deterministic. If you ask them to add a million floats in order, you get the same result every time. If you ask them to add a million floats in some arbitrary order, then you may get different results every time. The distinction is that someone had to ask the GPU to do that. It's a choice.

> I mostly take issue with you phrasing it as a bug and using it to insult the authors.

It's a bug, whether it insults the authors or not is irrelevant. It's most definitely a bug.

spott2y ago

Basically any parallel map-reduce operation using non-commutative reduce operators[0] is non-deterministic unless you specifically sort after/during the gather, or block on the gather (and gather to a thread-determined memory location). Sorting and blocking takes time. If you remove the sort/block, you will get a non-deterministic answer when operating on floats for a wide variety of reduce operations, but it will be faster. This is true of any parallel map-reduce, done anywhere (MPI, cuda kernels, openMP, spark, etc.), and is not unique to gpus/cuda.

> If you ask them to add a million floats in order, you get the same result every time.

There are a bunch of ways to add a million floats in order on a gpu, but they will all get you different results.:

* split the million floats into ‘n’ chunks, each chunk is summed, then you sum the ‘n’ results. * if you sum results as they are gathered (you don’t need to block) you will get a non-deterministic result, as the threads finishing (outside of a warp) is non-deterministic in order. * if you change ‘n’, your result will change. * if you sort after gathering , your result will change.

TLDR: parallel race-conditions are nondeterministic. Map-reduce has an underlying race-condition that you can prevent, but it costs time/performance. Sometimes you don’t care about the non-determinism enough to pay the performance penalty to fix it.

[0] https://www.microsoft.com/en-us/research/wp-content/uploads/...

2 more replies

thaw135792y ago

I think whether it’s a bug or not depends on the software requirements and expectations. If the code has some expected bounds on runtime, switching the GPU code to sequential processing (for the sake of exact reproducibility) would break that expectation and could be considered a bug as well. If we expect performant code and exact reproducibility, that just might not be possible…

ascar2y ago

It's hard to call it a bug given that any concurrent float sum or product will be different in regards to changing the amount of concurrency. Even if you order the final value per thread before reducing the result will differ if you use a different amount of threads to split the problem.

Because in floating point arithmetic 1 + 2 + 3 + 4 is different than (1+2) + (3+4).

j / k navigate · click thread line to collapse