We Keep Coding

What's an intelligent way of loading a compressed array completely from disk into memory - also (indentically) compressed?

I am experimenting with a 3-dimensional zarr-array, stored on disk:
Name: /data
Type: zarr.core.Array
Data type: int16
Shape: (102174, 1100, 900)
Chunk shape: (12, 220, 180)
Order: C
Read-only: True
Compressor: Blosc(cname='zstd', clevel=3, shuffle=BITSHUFFLE, blocksize=0)
Store type: zarr.storage.DirectoryStore
No. bytes: 202304520000 (188.4G)
No. bytes stored: 12224487305 (11.4G)
Storage ratio: 16.5
Chunks initialized: 212875/212875
As I understand it, zarr-arrays can also reside in memory - compressed, as if they were on disk. So I thought why not try to load the entire thing into RAM on a machine with 32 GByte memory. Compressed, the dataset would require approximately 50% of RAM. Uncompressed, it would require about 6 times more RAM than available.
Preparation:
import os
import zarr
from numcodecs import Blosc
import tqdm
zpath = '...' # path to zarr data folder
disk_array = zarr.open(zpath, mode = 'r')['data']
c = Blosc(cname = 'zstd', clevel=3, shuffle = Blosc.BITSHUFFLE)
memory_array = zarr.zeros(
disk_array.shape, chunks = disk_array.chunks,
dtype = disk_array.dtype, compressor = c
)
The following experiment fails almost immediately with an out of memory error:
memory_array[:, :, :] = disk_array[:, :, :]
As I understand it, disk_array[:, :, :] will try to create an uncompressed, full-size numpy array, which will obviously fail.
Second attempt, which works but is agonizingly slow:
chunk_lines = disk_array.chunks[0]
chunk_number = disk_array.shape[0] // disk_array.chunks[0]
chunk_remain = disk_array.shape[0] % disk_array.chunks[0] # unhandled ...
for chunk in tqdm.trange(chunk_number):
chunk_slice = slice(chunk * chunk_lines, (chunk + 1) * chunk_lines)
memory_array[chunk_slice, :, :] = disk_array[chunk_slice, :, :]
Here, I am trying to reads a certain number of chunks at a time and put them into my in-memory array. It works, but it is about 6 to 7 times slower than what it took to write this thing to disk in the first place. EDIT: Yes, it's still slow, but the 6 to 7 times happened due to a disk issue.
What's an intelligent and fast way of achieving this? I'd guess, besides not using the right approach, my chunks might also be too small - but I am not sure.
EDIT: Shape, chunk size and compression are supposed to be identical for the on-disk array and the in-memory array. It should therefore be possible to eliminate the decompress-compress procedure in my example above.
I found zarr.convenience.copy but it is marked as an experimental feature, subject to further change.
Related issue on GitHub

You could conceivably try with fsspec.implementations.memory.MemoryFileSystem, which has a .make_mapper() method, with which you can make the kind of object expected by zarr.
However, this is really just a dict of path:io.BytesIO, which you could make yourself, if you want.

There are a couple of ways one might solve this issue today.
Use LRUStoreCache to cache (some) compressed data in memory.
Coerce your underlying store into a dict and use that as your store.
The first option might be appropriate if you only want some frequently used data in-memory. Of course how much you load into memory is something you can configure. So this could be the whole array. This will only happen with data on-demand, which may be useful for you.
The second option just creates a new in-memory copy of the array by pulling all of the compressed data from disk. The one downside is if you intend to write back to disk this will be something you need to do manually, but it is not too difficult. The update method is pretty handy for facilitating this copying of data between different stores.

numpy memmap memory usage - want to iterate once

let say I have some big matrix saved on disk. storing it all in memory is not really feasible so I use memmap to access it
A = np.memmap(filename, dtype='float32', mode='r', shape=(3000000,162))
now let say I want to iterate over this matrix (not essentially in an ordered fashion) such that each row will be accessed exactly once.
p = some_permutation_of_0_to_2999999()
I would like to do something like that:
start = 0
end = 3000000
num_rows_to_load_at_once = some_size_that_will_fit_in_memory()
while start < end:
indices_to_access = p[start:start+num_rows_to_load_at_once]
do_stuff_with(A[indices_to_access, :])
start = min(end, start+num_rows_to_load_at_once)
as this process goes on my computer is becoming slower and slower and my RAM and virtual memory usage is exploding.
Is there some way to force np.memmap to use up to a certain amount of memory? (I know I won't need more than the amount of rows I'm planning to read at a time and that caching won't really help me since I'm accessing each row exactly once)
Maybe instead is there some other way to iterate (generator like) over a np array in a custom order? I could write it manually using file.seek but it happens to be much slower than np.memmap implementation
do_stuff_with() does not keep any reference to the array it receives so no "memory leaks" in that aspect
thanks

This has been an issue that I've been trying to deal with for a while. I work with large image datasets and numpy.memmap offers a convenient solution for working with these large sets.
However, as you've pointed out, if I need to access each frame (or row in your case) to perform some operation, RAM usage will max out eventually.
Fortunately, I recently found a solution that will allow you to iterate through the entire memmap array while capping the RAM usage.
Solution:
import numpy as np
# create a memmap array
input = np.memmap('input', dtype='uint16', shape=(10000,800,800), mode='w+')
# create a memmap array to store the output
output = np.memmap('output', dtype='uint16', shape=(10000,800,800), mode='w+')
def iterate_efficiently(input, output, chunk_size):
# create an empty array to hold each chunk
# the size of this array will determine the amount of RAM usage
holder = np.zeros([chunk_size,800,800], dtype='uint16')
# iterate through the input, replace with ones, and write to output
for i in range(input.shape[0]):
if i % chunk_size == 0:
holder[:] = input[i:i+chunk_size] # read in chunk from input
holder += 5 # perform some operation
output[i:i+chunk_size] = holder # write chunk to output
def iterate_inefficiently(input, output):
output[:] = input[:] + 5
Timing Results:
In [11]: %timeit iterate_efficiently(input,output,1000)
1 loop, best of 3: 1min 48s per loop
In [12]: %timeit iterate_inefficiently(input,output)
1 loop, best of 3: 2min 22s per loop
The size of the array on disk is ~12GB. Using the iterate_efficiently function keeps the memory usage to 1.28GB whereas the iterate_inefficiently function eventually reaches 12GB in RAM.
This was tested on Mac OS.

I've been experimenting with this problem for a couple days now and it appears there are two ways to control memory consumption using np.mmap. The first is reliable while the second would require some testing and will be OS dependent.
Option 1 - reconstruct the memory map with each read / write:
def MoveMMapNPArray(data, output_filename):
CHUNK_SIZE = 4096
for idx in range(0,x.shape[1],CHUNK_SIZE):
x = np.memmap(data.filename, dtype=data.dtype, mode='r', shape=data.shape, order='F')
y = np.memmap(output_filename, dtype=data.dtype, mode='r+', shape=data.shape, order='F')
end = min(idx+CHUNK_SIZE, data.shape[1])
y[:,idx:end] = x[:,idx:end]
Where data is of type np.memmap. This discarding of the memmap object with each read keeps the array from being collected into memory and will keep memory consumption very low if the chunk size is low. It likely introduces some CPU overhead but was found to be small on my setup (MacOS).
Option 2 - construct the mmap buffer yourself and provide memory advice
If you look at the np.memmap source code here, you can see that it is relatively simple to create your own memmapped numpy array relatively easily. Specifically, with the snippet:
mm = mmap.mmap(fid.fileno(), bytes, access=acc, offset=start)
mmap_np_array = ndarray.__new__(subtype, shape, dtype=descr, buffer=mm, offset=array_offset, order=order)
Note this python mmap instance is stored as the np.memmap's private _mmap attribute.
With access to the python mmap object, and python 3.8, you can use its madvise method, described here.
This allows you to advise the OS to free memory where available. The various madvise constants are described here for linux, with some generic cross platform options specified.
The MADV_DONTDUMP constant looks promising but I haven't tested memory consumption with it like I have for option 1.

python file io loop mystery (CPU goes disk sleep)

I'm looping histogram operation on HDF5 files of size ~800MB each (equal size).
The result of histogram is stored in text files each with ~5column x 30 lines.
t0 = time.time()
for f in filelist:
d = h5py.File(f,'r')
result = make_histogram(d['X'].value)
ascii_write(result)
print time.time()-t0
d.close()
One pass through the loop normally seems to take ~6-7 seconds for each file.
However, at some point it takes significantly longer to pass one loop.
And this point in time seems to start rather randomly if I try running multiple times with different files starting first.
I noticed that in my system monitor, at this point, CPU is in "disk sleep".
How can I fix this?
It seems to be related to this question, but I could not find a definitive answer.

Write ascii output from numpy arrays

I want to write some random numbers into an ascii output file.
I generate the numbers with numpy, so the numbers are stored in numpy.array
import numpy as np
random1=np.random.uniform(-1.2,1.2,7e6)
random2=...
random3=...
All three array are of the same size.
I used standard file output, but this is really slow. Just about 8000 lines per 30 min. This may because I loop over three large arrays though.
fout1 = open("output.dat","w")
for i in range(len(random1)):
fout1.write(str(random1[i])+"\t"+ str(random2[i])+"\t"+ str(random3[i])+"\n")
fout1.close()
I also just used print str(random1[i])+"\t"+ str(random2[i])+"\t"+ str(random3[i]) and dumped everything in a file usind shell ./myprog.py > output.dat which seems a bit faster but still I am not satisfied with the output speed.
Any recommendations are really welcome.

Have you tried
random = np.vstack((random1, random2, random3)).T
random.savetxt("output.dat", delimiter="\t")

Im guessing the disk io is the most expensive operation you are doing.. You could try to create your own buffer to deal with this, instead of writing every line every loop buffer up say 100 lines and write them in one big block. Then experiment with this and see what the most benficial buffer size is

OpenCL delete data from RAM

I'm coping data in python using OpenCL onto my graphic card. There I've a kernel processing the data with n threads.
After this step I copy the result back to python and in a new kernel. (The data is very big 900MB and the result is 100MB) With the result I need to calculate triangles which are about 200MB. All data exceed the memory on my graphic card.
I do not need the the first 900MB anymore after the first kernel finished it's work.
My question is, how can I delete the first dataset (stored in one array) from the graphic card?
Here some code:
#Write
self.gridBuf = cl.Buffer(self.context, cl.mem_flags.READ_ONLY | cl.mem_flags.COPY_HOST_PTR, hostbuf=self.grid)
#DO PART 1
...
#Read result
cl.enqueue_read_buffer(self.queue, self.indexBuf,index).wait()

You will need to call clReleaseMemObject with the mem object you created with the call to clCreateBuffer. If the reference count becomes zero with this call, the underlying device/shared memory is released by the implementation.