Re: [PATCH v3 5/5] unpack-objects: unpack_non_delta_entry() read data in a stream

On 11/21/2021 10:32 PM, Han Xin wrote:

From: Han Xin <redacted>

We used to call "get_data()" in "unpack_non_delta_entry()" to read the
entire contents of a blob object, no matter how big it is. This
implementation may consume all the memory and cause OOM.

By implementing a zstream version of input_stream interface, we can use
a small fixed buffer for "unpack_non_delta_entry()".

However, unpack non-delta objects from a stream instead of from an entrie
buffer will have 10% performance penalty. Therefore, only unpack object
larger than the "big_file_threshold" in zstream. See the following
benchmarks:

    $ hyperfine \
    --prepare 'rm -rf dest.git && git init --bare dest.git' \
    'git -C dest.git unpack-objects <binary_320M.pack'
    Benchmark 1: git -C dest.git unpack-objects <binary_320M.pack
      Time (mean ± σ):     10.029 s ±  0.270 s    [User: 8.265 s, System: 1.522 s]
      Range (min … max):    9.786 s … 10.603 s    10 runs

    $ hyperfine \
    --prepare 'rm -rf dest.git && git init --bare dest.git' \
    'git -c core.bigFileThreshold=2m -C dest.git unpack-objects <binary_320M.pack'
    Benchmark 1: git -c core.bigFileThreshold=2m -C dest.git unpack-objects <binary_320M.pack
      Time (mean ± σ):     10.859 s ±  0.774 s    [User: 8.813 s, System: 1.898 s]
      Range (min … max):    9.884 s … 12.192 s    10 runs

It seems that you want us to compare this pair of results, and
hyperfine can assist with that by including multiple benchmarks
(with labels, using '-n') as follows:

$ hyperfine \
        --prepare 'rm -rf dest.git && git init --bare dest.git' \
        -n 'old' '~/_git/git-upstream/git -C dest.git unpack-objects <big.pack' \
        -n 'new' '~/_git/git/git -C dest.git unpack-objects <big.pack' \
        -n 'new (small threshold)' '~/_git/git/git -c core.bigfilethreshold=64k -C dest.git unpack-objects <big.pack'

Benchmark 1: old
  Time (mean ± σ):     20.835 s ±  0.058 s    [User: 14.510 s, System: 6.284 s]
  Range (min … max):   20.741 s … 20.909 s    10 runs
 
Benchmark 2: new
  Time (mean ± σ):     26.515 s ±  0.072 s    [User: 19.783 s, System: 6.696 s]
  Range (min … max):   26.419 s … 26.611 s    10 runs
 
Benchmark 3: new (small threshold)
  Time (mean ± σ):     26.523 s ±  0.101 s    [User: 19.805 s, System: 6.680 s]
  Range (min … max):   26.416 s … 26.739 s    10 runs
 
Summary
  'old' ran
    1.27 ± 0.00 times faster than 'new'
    1.27 ± 0.01 times faster than 'new (small threshold)'

(Here, 'old' is testing a compiled version of the latest 'master'
branch, while 'new' has your patches applied on top.)

Notice from this example I had a pack with many small objects (mostly
commits and trees) and I see that this change introduces significant
overhead to this case.

It would be nice to understand this overhead and fix it before taking
this change any further.

Thanks,
-Stolee