git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: 2.18.0 Regression: packing performance and effectiveness

From
Elijah Newren <newren@gmail.com>
Date
Jul 19, 2018, 16:42 UTC
Message-ID
<CABPp-BGE0yAgtOkSC0jUhq3M3ynX9pks5VnjVB_WMkYUVdirzA@mail.gmail.com>
In-Reply-To
<20180719151640.GA24997@duynguyen.home>
On Thu, Jul 19, 2018 at 8:16 AM, Duy Nguyen <pclouds@gmail.com> wrote:
Show 13 quoted lines
> On Thu, Jul 19, 2018 at 07:57:37AM +0200, Duy Nguyen wrote:
>> I thought 2M was generous but I was obviously wrong. I'd like to see
>> the biggest delta size in this pack and whether it's still reasonable
>> to increase OE_DELTA_SIZE_BITS. But if it's very close to 4GB limit
>> then we could just store 64-bit delta size in a separate array.
>
> I realize now that these two options don't have to be mutually
> exclusive and I should provide a good fallback in terms of performance
> anyway.
>
> So Elijah, could you try this patch and see if performance and pack
> size go back to 1.7.0? I can write proper commit message and test
> support later if it proves a good improvement.

Thanks for the quick turnaround. Unfortunately, I have some bad news. With this patch, I get the following:

$ /usr/bin/time -f 'MaxRSS:%M Time:%e' git gc --aggressive Enumerating objects: 4460703, done. Counting objects: 100% (4460703/4460703), done. Delta compression using up to 40 threads. Compressing objects: 100% (3807140/3807140), done. Writing objects: 100% (4460703/4460703), done. Total 4460703 (delta 2831383), reused 1587071 (delta 0) error: failed to unpack compressed delta at offset 183854150 from .git/objects/pack/pack-30d4f0b0e5a03dc91a658a0586f4e74cdf4a94d6.pack fatal: packed object 20ce811e53dabbb8ef9368c108cbbdfa65639c03 (stored in .git/objects/pack/pack-30d4f0b0e5a03dc91a658a0586f4e74cdf4a94d6.pack) is corrupt error: failed to run prune MaxRSS:40025196 Time:2531.52

It looks like this comes after it deleted the old packs, so running git fsck afterwards prints lots of errors. Just in case they're helpful at all, there are 26 "error: failed to unpack compressed delta at offset..." messages, 26 "error: cannot unpack ... from ... at offset ..." messages, 17 "error: failed to read delta base object..." messages, 7 "broken link from ... to ..." messages, four missing blobs, and three missing trees.

(This was run on a copy of the repo, so no harm done; I can still use the original.)

Show 115 quoted lines
> -- 8< --
> diff --git a/builtin/pack-objects.c b/builtin/pack-objects.c
> index ebc8cefb53..7f3546057d 100644
> --- a/builtin/pack-objects.c
> +++ b/builtin/pack-objects.c
> @@ -2023,10 +2023,6 @@ static int try_delta(struct unpacked *trg, struct unpacked *src,
>         delta_buf = create_delta(src->index, trg->data, trg_size, &delta_size, max_size);
>         if (!delta_buf)
>                 return 0;
> -       if (delta_size >= (1U << OE_DELTA_SIZE_BITS)) {
> -               free(delta_buf);
> -               return 0;
> -       }
>
>         if (DELTA(trg_entry)) {
>                 /* Prefer only shallower same-sized deltas. */
> diff --git a/pack-objects.c b/pack-objects.c
> index 92708522e7..191ed45faf 100644
> --- a/pack-objects.c
> +++ b/pack-objects.c
> @@ -160,6 +160,8 @@ struct object_entry *packlist_alloc(struct packing_data *pdata,
>
>                 if (!pdata->in_pack_by_idx)
>                         REALLOC_ARRAY(pdata->in_pack, pdata->nr_alloc);
> +               if (pdata->delta_size)
> +                       REALLOC_ARRAY(pdata->delta_size, pdata->nr_alloc);
>         }
>
>         new_entry = pdata->objects + pdata->nr_objects++;
> @@ -177,3 +179,17 @@ struct object_entry *packlist_alloc(struct packing_data *pdata,
>
>         return new_entry;
>  }
> +
> +void oe_prepare_delta_size_array(struct packing_data *pack)
> +{
> +       int i;
> +
> +       /*
> +        * nr_alloc, not nr_objects to align with realloc() strategy in
> +        * packlist_alloc()
> +        */
> +       ALLOC_ARRAY(pack->delta_size, pack->nr_alloc);
> +
> +       for (i = 0; i < pack->nr_objects; i++)
> +               pack->delta_size[i] = pack->objects[i].delta_size_;
> +}
> diff --git a/pack-objects.h b/pack-objects.h
> index edf74dabdd..978500e474 100644
> --- a/pack-objects.h
> +++ b/pack-objects.h
> @@ -14,7 +14,7 @@
>   * above this limit. Don't lower it too much.
>   */
>  #define OE_SIZE_BITS           31
> -#define OE_DELTA_SIZE_BITS     20
> +#define OE_DELTA_SIZE_BITS     21
>
>  /*
>   * State flags for depth-first search used for analyzing delta cycles.
> @@ -93,7 +93,6 @@ struct object_entry {
>                                      * uses the same base as me
>                                      */
>         unsigned delta_size_:OE_DELTA_SIZE_BITS; /* delta data size (uncompressed) */
> -       unsigned delta_size_valid:1;
>         unsigned in_pack_idx:OE_IN_PACK_BITS;   /* already in pack */
>         unsigned z_delta_size:OE_Z_DELTA_BITS;
>         unsigned type_valid:1;
> @@ -130,6 +129,7 @@ struct packing_data {
>         uint32_t index_size;
>
>         unsigned int *in_pack_pos;
> +       unsigned long *delta_size;
>
>         /*
>          * Only one of these can be non-NULL and they have different
> @@ -330,20 +330,27 @@ static inline void oe_set_size(struct packing_data *pack,
>  static inline unsigned long oe_delta_size(struct packing_data *pack,
>                                           const struct object_entry *e)
>  {
> -       if (e->delta_size_valid)
> +       if (!pack->delta_size)
>                 return e->delta_size_;
> -       return oe_size(pack, e);
> +       return pack->delta_size[e - pack->objects];
>  }
>
> +void oe_prepare_delta_size_array(struct packing_data *pack);
>  static inline void oe_set_delta_size(struct packing_data *pack,
>                                      struct object_entry *e,
>                                      unsigned long size)
>  {
>         e->delta_size_ = size;
> -       e->delta_size_valid = e->delta_size_ == size;
> -       if (!e->delta_size_valid && size != oe_size(pack, e))
> -               BUG("this can only happen in check_object() "
> -                   "where delta size is the same as entry size");
> +       if (!pack->delta_size && e->delta_size_ == size)
> +               return;
> +       /*
> +        * We have had at least one delta size exceeding OE_DELTA_SIZE_BITS
> +        * limit. delta_size_ will not be used anymore. All delta sizes are now
> +        * from the delta_size[] array.
> +        */
> +       if (!pack->delta_size)
> +               oe_prepare_delta_size_array(pack);
> +       pack->delta_size[e - pack->objects] = size;
>  }
>
>  #endif
> -- 8< --
>
>
> --
> Duy
Previous: Duy NguyenNext: Jeff King
Message 11 of 44 in “2.18.0 Regression: packing performance and effectiveness”
  1. Elijah NewrenJul 18, 2018
  2. fix-v1: revert "pack-objects: shrink delta_size field in struct object_entry"Elijah Newren, Jul 18, 2018
  3. fix-v2: make OE_DELTA_SIZE_BITS a bit biggerElijah Newren, Jul 18, 2018
  4. Duy NguyenJul 19, 2018
  5. Jeff KingJul 19, 2018
  6. Elijah NewrenJul 19, 2018
  7. Duy NguyenJul 19, 2018
  8. Jeff KingJul 19, 2018
  9. Duy NguyenJul 19, 2018
  10. Duy NguyenJul 19, 2018
  11. Elijah NewrenJul 19, 2018
  12. Jeff KingJul 19, 2018
  13. Duy NguyenJul 19, 2018
  14. Duy NguyenJul 19, 2018
  15. Jeff KingJul 19, 2018
  16. Elijah NewrenJul 19, 2018
  17. Jeff KingJul 20, 2018
  18. Jeff KingJul 20, 2018
  19. Duy NguyenJul 20, 2018
  20. Elijah NewrenJul 20, 2018
  21. Jeff KingJul 19, 2018
  22. Junio C HamanoJul 19, 2018
  23. Junio C HamanoJul 19, 2018
  24. pack-objects: fix performance issues on packing large deltasNguyễn Thái Ngọc Duy, Jul 20, 2018
  25. Jeff KingJul 20, 2018
  26. Duy NguyenJul 21, 2018
  27. Jeff KingJul 23, 2018
  28. Elijah NewrenJul 20, 2018
  29. Elijah NewrenJul 20, 2018
  30. Duy NguyenJul 21, 2018
  31. Duy NguyenJul 21, 2018
  32. Duy NguyenJul 21, 2018
  33. Elijah NewrenJul 21, 2018
  34. Duy NguyenJul 21, 2018
  35. Elijah NewrenJul 22, 2018
  36. Duy NguyenJul 22, 2018
  37. Elijah NewrenJul 23, 2018
  38. Duy NguyenJul 23, 2018
  39. pack-objects: fix performance issues on packing large deltasNguyễn Thái Ngọc Duy, Jul 22, 2018
  40. Junio C HamanoJul 23, 2018
  41. Duy NguyenJul 23, 2018
  42. Duy NguyenJul 23, 2018
  43. Jeff KingJul 23, 2018
  44. Johannes SixtJul 26, 2018

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.