Not plain-text I hope. Was this extracted directly from providers? Given the fact that people reuse passwords, it should be easy to construct a larger db theoretically. But this number is mind-boggling.
When you call a cuda method, it is launched asynchronously. That is the function queues it up for execution on gpu and returns.
So if you need to wait for an op to finish, you need to `synchronize` as shown above.
`get_current_stream` because the queue mentioned above is actually called stream in cuda.
If you want to run many independent ops concurrently, you can use several streams.
Benchmarking is one use case for synchronize. Another would be if you let's say run two independent ops in different streams and need to combine their results.
Btw, if you work with pytorch, when ops are run on gpu, they are launched in background. If you want to bench torch models on gpu, they also provide a sync api.
I’ve always thought it was weird GPU stuff in python doesn’t use asyncio, and mostly assumed it was because python-on-GPU predates asyncio. But I was hoping a new lib like this might right that wrong, but it doesn’t. Maybe for interop reasons?
Do other languages surface the asynchronous nature of GPUs in language-level async, avoiding silly stuff like synchronize?
The reason is that the usage is completely different from coroutine based async. With GPUs you want to queue _as many async operations as possible_ and only then synchronize. That is, you would have a program like this (pseudocode):
b = foo(a)
c = bar(b)
d = baz(c)
synchronize()
With coroutines/async await, something like this
b = await foo(a)
c = await bar(b)
d = await baz(c)
would synchronize after every step, being much more inefficient.
It really depends on if you're dealing with an async stream or a single async result as the input to the next function. If a is an access token needed to access resource b, you cannot access a and b at the same time. You have to serialize your operations.
Well you can and should create multiple coroutine/tasks and then gather them. If you replace cuda with network calls, it’s exactly the same problem. Nothing to do with asyncio.
No, that's a different scenario. In the one I gave there's explicitly a dependency between requests. If you use gather, the network requests would be executed in parallel. If you have dependencies they're sequential by nature because later ones depend on values of former ones.
The 'trick' for CUDA is that you declare all this using buffers as inputs/outputs rather than values and that there's automatic ordering enforcement through CUDA's stream mechanism. Marrying that with the coroutine mechanism just doesn't really make sense.
Might have to look at specific lib implementations, but I'd guess that mostly gpu calls from python are actually happening in c++ land. And internally a lib might be using synchronize calls where needed.
I'd imagine that Meta is trying to avoid an AI / LLM monopoly from happening as the primary goal.
They suffered when Apple was a gateway to their service for ios users and decided to shut their access to user data off.
And AI is clearly going to be used widely to aid in content generation if not doing entirely. Also they tried jumping on the chatbot hype with M or something. That didn't pan out as well.
By opening up AI, they would enable tools that allow users to pump out more content easily and spend more time in app.
Getting the goodwill of dev community is a bonus.
A brilliant move from Meta.
Hang in there! And to add to what others have said, I found it extremely helpful to stay away from sad music, romantic shows, etc. until you feel strong again. And keep your body moving!
Glad to see Revolution OS in the list! Helped me get through times in grad school when I'd get stuck on something working on my thesis. Grab a bowl of ramen and put this movie on!
It checks for specific return code. It is useful for testing specific type of failure . Since in shell all non zero return codes are considered failure, we need a way to know that it failed because of the expected reason and not some other bug. So if testing for failure with err code 2, the function will "pass" the test only if the process actually fails with err code 2.
