3 ms·

Sounds like you're talking out of your butt instead of doing the math.

by lostmsu 4mo ago

zozbot234 4mo ago
What do you mean by doing the math? If you repeatedly sample n_active experts out of n_total, why wouldn't you expect to get some meaningful probability of reuse/overlap once your batch grows past size 5 or so (for the sparsest MoE models in common use)? And you only need enough reuse to fill the compute headroom which is quite small on consumer platforms (we won't have huge TOPS numbers for the typical integrated GPU in Strix Halo or even the upcoming RTX Spark). Plus if you're a single user running multiple streams in parallel the choice of experts will be highly biased leading to more reuse.
- lostmsu 4mo ago
  Yeah, that's what talking out of your butt is literally. "theoretical", no ballparks, ignorant assumptions about expert reuse.
  - zozbot234 4mo ago
    There's been some very rough experiments with batching on Apple Silicon (and that's not a highly suitable platform since the compute/thermals bottleneck hits sooner than elsewhere) that seem to be broadly consistent with what I argued, showing as much as 2x total decode throughput with an 8-wide batch. That's substantial in this context.
    - lostmsu 4mo ago
      Assuming you magically use all 128GiB of xRAM you need to read ~32GiB per token in batched mode. On a good SSD that would be 1/3 tokens per second. Cool, 2x that you can do 2/3 tokens per second. Let's assume you are lucky and can actually do 6/7 tokens per second. That's still an extremely far cry from 20+ tokens per second of 27B before any batching.
      - fc417fc802 4mo ago
        I don't understand where your numbers are coming from. Why is there a 20x (40x?) slowdown of tps after batching?
        lostmsu 4mo ago
        Before batching. The slowdown is because the model does not fit the xRAM so experts will have to be read from SSD on every forward pass. That's why it is impractically slow. Batching could allow you to generate 10 tokens for 10 different conversations at the time, but it also means that you need to load different experts for different tokens, so it does not help as much as it does for dense models.
        fc417fc802 4mo ago
        But IIUC the point is that each expert gets used for more than just the one token. So yes, the tps of a given thread takes a hit because now you're sometimes going to schedule in unrelated experts and it will have to pause. But overall you're utilizing the hardware much more efficiently and so in aggregate there's a speedup. On top of that (as previously pointed out by zoz) for a single user running a single overarching task the choice of experts is expected to be highly biased.
        lostmsu 4mo ago
        > the choice of experts is expected to be highly biased Why? Why do you think that's the case? Part of the training is balancing load between experts. > so in aggregate there's a speedup. Yes. 2x. Over theoretical under 1 tok/s
        fc417fc802 4mo ago
        > Why do you think that's the case? Part of the training is balancing load between experts. That is a fair point. That expectation may have been misplaced on my part. I'm not sufficiently familiar with the details of MoE training. > The slowdown is because the model does not fit the xRAM so experts will have to be read from SSD on every forward pass. > 20+ tokens per second of 27B before any batching. Does the model fit in RAM or not? What is your justification for your stated expectation that the unbatched model will perform 20x faster than the aggregate tps (note, not the single stream tps) of the batched model? My expectation is that if the unbatched model is 20 tps and batching provides a 2x speedup then each individual stream will be slower but the aggregate throughput should rise to 40 tps. What do you believe me to be missing here?

Sounds like you're talking out of your butt instead of doing the math.

by lostmsu 4mo ago

Sounds like you're talking out of your butt instead of doing the math.

What do you mean by doing the math? If you repeatedly sample n_active experts out of n_total, why wouldn't you expect to get some meaningful probability of reuse/overlap once your batch grows past size 5 or so (for the sparsest MoE models in common use)? And you only need enough reuse to fill the compute headroom which is quite small on consumer platforms (we won't have huge TOPS numbers for the typical integrated GPU in Strix Halo or even the upcoming RTX Spark). Plus if you're a single user running multiple streams in parallel the choice of experts will be highly biased leading to more reuse.

lostmsu 4mo ago
Yeah, that's what talking out of your butt is literally. "theoretical", no ballparks, ignorant assumptions about expert reuse.
- zozbot234 4mo ago
  There's been some very rough experiments with batching on Apple Silicon (and that's not a highly suitable platform since the compute/thermals bottleneck hits sooner than elsewhere) that seem to be broadly consistent with what I argued, showing as much as 2x total decode throughput with an 8-wide batch. That's substantial in this context.
  - lostmsu 4mo ago
    Assuming you magically use all 128GiB of xRAM you need to read ~32GiB per token in batched mode. On a good SSD that would be 1/3 tokens per second. Cool, 2x that you can do 2/3 tokens per second. Let's assume you are lucky and can actually do 6/7 tokens per second. That's still an extremely far cry from 20+ tokens per second of 27B before any batching.
    - fc417fc802 4mo ago
      I don't understand where your numbers are coming from. Why is there a 20x (40x?) slowdown of tps after batching?
      - lostmsu 4mo ago
        Before batching. The slowdown is because the model does not fit the xRAM so experts will have to be read from SSD on every forward pass. That's why it is impractically slow. Batching could allow you to generate 10 tokens for 10 different conversations at the time, but it also means that you need to load different experts for different tokens, so it does not help as much as it does for dense models.
        fc417fc802 4mo ago
        But IIUC the point is that each expert gets used for more than just the one token. So yes, the tps of a given thread takes a hit because now you're sometimes going to schedule in unrelated experts and it will have to pause. But overall you're utilizing the hardware much more efficiently and so in aggregate there's a speedup. On top of that (as previously pointed out by zoz) for a single user running a single overarching task the choice of experts is expected to be highly biased.
        lostmsu 4mo ago
        > the choice of experts is expected to be highly biased Why? Why do you think that's the case? Part of the training is balancing load between experts. > so in aggregate there's a speedup. Yes. 2x. Over theoretical under 1 tok/s
        fc417fc802 4mo ago
        > Why do you think that's the case? Part of the training is balancing load between experts. That is a fair point. That expectation may have been misplaced on my part. I'm not sufficiently familiar with the details of MoE training. > The slowdown is because the model does not fit the xRAM so experts will have to be read from SSD on every forward pass. > 20+ tokens per second of 27B before any batching. Does the model fit in RAM or not? What is your justification for your stated expectation that the unbatched model will perform 20x faster than the aggregate tps (note, not the single stream tps) of the batched model? My expectation is that if the unbatched model is 20 tps and batching provides a 2x speedup then each individual stream will be slower but the aggregate throughput should rise to 40 tps. What do you believe me to be missing here?