FAST | FAST App Search Tool | Runtime Evironment library

(fastest methods, 3 and 4, are at the end)

In a fast NumPy method you need to specify dtype=np.object so that NumPy does not convert Python int to its own dtypes (np.int64 or others). It will now give you correct results (checked it up to N=100000).

Since you mention executing the same files (with proper performance) from within Git Bash, I'm going to make an assumption here. Correct me if I'm wrong on this, and I'll delete the answer and look for another possibility.

This would be explained (and expected) if your files are stored on /mnt/c (a.k.a. C:, or /C under Git Bash) or any other Windows drive, as they would likely need to be to be accessed by Git Bash.

WSL2 uses the 9P protocol to access Windows drives, and it is currently known to be very slow when compared to:

• Native NTFS (obviously)
• The ext4 filesystem on the virtual disk used by WSL2
• And even the performance of WSL1 with Windows drives

I've seen a git clone of a large repo (the WSL2 Linux kernel Github) take 8 minutes on WSL2 on a Windows drive, but only seconds on the root filesystem.

Two possibilities:

• If possible (and it is for most Node projects), convert your WSL to version 1 with wsl --set-version 1. I always recommend making a backup with wsl --export first.

  And since you are making a backup anyway, you may as well just create a copy of the instance by wsl --importing your backup as --version 1 (as the last argument). WSL1 and WSL2 both have their uses, and you may find it helpful to keep both around.

  See this answer for more details on the exact syntax..

• Or just move the project over to somewhere under the WSL root, such as /home/username/src/.

There's also no meaningful latency benefit to stronger memory orders, even if latency of seeing a change to a keep_running or exit_now flag was important.

IDK why Herb thinks stop.store shouldn't be relaxed; in his talk, his slides have a comment that says // not relaxed on the assignment, but he doesn't say anything about the store side before moving on to "is it worth it".

Of course, the load runs inside the worker loop, but the store runs only once, and Herb really likes to recommend sticking with SC unless you have a performance reason that truly justifies using something else. I hope that wasn't his only reason; I find that unhelpful when trying to understand what memory order would actually be necessary and why. But anyway, I think either that or a mistake on his part.

The ISO C++ standard doesn't say anything about how soon stores become visible or what might influence that, just Section 6.9.2.3 Forward progress

18. An implementation should ensure that the last value (in modification order) assigned by an atomic or synchronization operation will become visible to all other threads in a finite period of time.

Another thread can loop arbitrarily many times before its load actually sees this store value, even if they're both seq_cst, assuming there's no other synchronization of any kind between them. Low inter-thread latency is a performance issue, not correctness / formal guarantee.

And non-infinite inter-thread latency is apparently only a "should" QOI (quality of implementation) issue. :P Nothing in the standard suggests that seq_cst would help on an implementation where store visibility could be delayed indefinitely, although one might guess that could be the case, e.g. on a hypothetical implementation with explicit cache flushes instead of cache coherency. (Although such an implementation is probably not practically usable in terms of performance with CPUs anything like what we have now; every release and/or acquire operation would have to flush the whole cache.)

On real hardware (which uses some form of MESI cache coherency), different memory orders for store or load don't make stores visible sooner in real time, they just control whether later operations can become globally visible while still waiting for the store to commit from the store buffer to L1d cache. (After invalidating any other copies of the line.)

Stronger orders, and barriers, don't make things happen sooner in an absolute sense, they just delay other things until they're allowed to happen relative to the store or load. (This is the case on all real-world CPUs AFAIK; they always try to make stores visible to other cores ASAP anyway, so the store buffer doesn't fill up, and

See also (my similar answers on):

• Does hardware memory barrier make visibility of atomic operations faster in addition to providing necessary guarantees?
• If I don't use fences, how long could it take a core to see another core's writes?
• memory_order_relaxed and visibility
• Thread synchronization: How to guarantee visibility of writes (it's a non-issue on current real hardware)

The second Q&A is about x86 where commit from the store buffer to L1d cache is in program order. That limits how far past a cache-miss store execution can get, and also any possible benefit of putting a release or seq_cst fence after the store to prevent later stores (and loads) from maybe competing for resources. (x86 microarchitectures will do RFO (read for ownership) before stores reach the head of the store buffer, and plain loads normally compete for resources to track RFOs we're waiting for a response to.) But these effects are extremely minor in terms of something like exiting another thread; only very small scale reordering.

because who cares if the thread stops with a slightly bigger delay.

More like, who cares if the thread gets more work done by not making loads/stores after the load wait for the check to complete. (Of course, this work will get discarded if it's in the shadow of a a mis-speculated branch on the load result when we eventually load true.) The cost of rolling back to a consistent state after a branch mispredict is more or less independent of how much already-executed work had happened beyond the mispredicted branch. And it's a stop flag so the total amount of wasted work costing cache/memory bandwidth for other CPUs is pretty minimal.

That phrasing makes it sound like an acquire load or release store would actually get the the store seen sooner in absolute real time, rather than just relative to other code in this thread. (Which is not the case).

The benefit is more instruction-level and memory-level parallelism across loop iterations when the load produces a false. And simply avoiding running extra instructions on ISAs where an acquire or especially an SC load needs extra instructions, especially expensive 2-way barrier instructions, not like ARM64 ldapr.

BTW, Herb is right that the dirty flag can also be relaxed, only because of the thread.join sync between the reader and any possible writer. Otherwise yeah, release / acquire.

But in this case, dirty only needs to be atomic<> at all because of possible simultaneous writers all storing the same value, which ISO C++ still deems data-race UB. e.g. because of the theoretical possibility of hardware race-detection that traps on conflicting non-atomic accesses.

At first blush, one would hope this would be enough:

Take a look at the recursive formula for combinations:

Suppose you have a combination space C(n,k). You can divide that space into two subspaces:

• C(n-1,k-1) all combinations, where the first element of the original set (of length n) is present
• C(n-1, k) where first element is not preset

If you have an index X that corresponds to a combination from C(n,k), you can identify whether the first element of your original set belongs to the subset (which corresponds to X), if you check whether X belongs to either subspace:

• X < C(n-1, k-1) : belongs
• X >= C(n-1, k-1): doesn't belong

Then you can recursively apply the same approach for C(n-1, ...) and so on, until you've found the answer for all n elements of the original set.

Python code to illustrate this approach:

I'm unsure as to the cause of that log message, but I do see that you are returning a response from your function before it completes all of its work. In a deployed function, as soon as the function returns, all further actions should be treated as if they will never be executed as documented here. An "inactive" function might be terminated at any time, is severely throttled and any network calls you make (like setting data in the RTDB) may never be executed.

I know you are new to this, but its a good habit to get into now: don't assume the person calling your function is you. Check for problems like missing query parameters and dodgy data before you blindly action something. The Admin SDK bypasses your database's security rules and if you are not careful a malicious user can cause some damage (e.g. a user that updates /users/$theirUid/roles/admin to true).

I asked around on GitHub and IRC. The solution was to download and use rakudo-pkg to get a newer version of zef. The one that can be installed via apt is too old.

The first method sends everything to any() whilst the second only sends to any() when there's an odd number, so any() has fewer elements to go through.

Here's the code in question, from the Python 3.10 branch (in ceval.c, and called from the same file's implementation of the BINARY_ADD opcode). As @jasonharper noted in a comment, it peeks ahead to see whether the result of the BINARY_ADD will next be bound to the same name from which the left-hand addend came. In fast(), it is (operand came from s and result stored into s), but in slow() it isn't (operand came from s but stored into t).

There's no guarantee this optimization will persist, though. For example, I noticed that your fast() is no faster than your slow() on the current development CPython main branch (which is the current work-in-progress toward an eventual 3.11 release).

As noted, there's no guarantee this optimization will persist. "Serious" Python programmers should know better than to rely on dodgy CPython-specific tricks, and, indeed, PEP 8 explicitly warns against relying on this specific one:

Code should be written in a way that does not disadvantage other implementations of Python (PyPy, Jython, IronPython, Cython, Psyco, and such).

For example, do not rely on CPython's efficient implementation of in-place string concatenation for statements in the form a += b or a = a + b ...