Prisma MongoDB benchmarks against mongoose #14811

paulrostorp · 2022-08-15T12:36:19Z

paulrostorp
Aug 15, 2022

In #12793 I discovered some operations failed to use indexes, which has severe performance impacts.
I've created a simple repo to benchmarks prisma operations: https://github.com/paulrostorp/prisma-mongodb-mongoose-benchmark, currently only 4 operations are being tested, but hopefully we can add some more.

I've run the benchmark using the latest prisma version (4.2.1) and these are the result I get.
Performance issue is most significative on the in operator, which always fails to use the index.

                Fetch a single user by ID if exists (user_by_id)
┌─────────────────┬─────────────────────┬───────────────────────┬──────────────┐
│ Collection size │ Prisma average (ms) │ Mongoose average (ms) │ % difference │
├─────────────────┼─────────────────────┼───────────────────────┼──────────────┤
│          100000 │                 3.7 │                   1.6 │       131.3% │
│          200000 │                 3.0 │                   1.6 │        87.5% │
│          300000 │                 2.5 │                   1.4 │        78.6% │
│          400000 │                 2.8 │                   1.4 │       100.0% │
│          500000 │                 3.2 │                   1.3 │       146.2% │
│          600000 │                 3.3 │                   1.7 │        94.1% │
│          700000 │                 2.4 │                   1.3 │        84.6% │
│          800000 │                 3.7 │                   1.1 │       236.4% │
│          900000 │                 1.8 │                   1.4 │        28.6% │
│         1000000 │                 2.4 │                   1.9 │        26.3% │
└─────────────────┴─────────────────────┴───────────────────────┴──────────────┘

                Fetch a list user from an array of IDs (user_in)
┌─────────────────┬─────────────────────┬───────────────────────┬──────────────┐
│ Collection size │ Prisma average (ms) │ Mongoose average (ms) │ % difference │
├─────────────────┼─────────────────────┼───────────────────────┼──────────────┤
│          100000 │                31.3 │                   2.5 │      1152.0% │
│          200000 │                54.7 │                   2.7 │      1925.9% │
│          300000 │                83.1 │                   2.4 │      3362.5% │
│          400000 │               103.5 │                   2.6 │      3880.8% │
│          500000 │               129.8 │                   2.5 │      5092.0% │
│          600000 │               155.5 │                   2.0 │      7675.0% │
│          700000 │               177.6 │                   2.1 │      8357.1% │
│          800000 │               200.5 │                   2.6 │      7611.5% │
│          900000 │               225.8 │                   1.8 │     12444.4% │
│         1000000 │               251.6 │                   1.7 │     14700.0% │
└─────────────────┴─────────────────────┴───────────────────────┴──────────────┘

   Fetch user(s) by filtering on a non-indexed field (user_find_non_indexed)
┌─────────────────┬─────────────────────┬───────────────────────┬──────────────┐
│ Collection size │ Prisma average (ms) │ Mongoose average (ms) │ % difference │
├─────────────────┼─────────────────────┼───────────────────────┼──────────────┤
│          100000 │                34.3 │                  22.7 │        51.1% │
│          200000 │                63.3 │                  43.0 │        47.2% │
│          300000 │                92.1 │                  60.0 │        53.5% │
│          400000 │               118.6 │                  78.6 │        50.9% │
│          500000 │               150.2 │                  98.3 │        52.8% │
│          600000 │               172.2 │                 117.2 │        46.9% │
│          700000 │               201.0 │                 140.9 │        42.7% │
│          800000 │               231.2 │                 153.4 │        50.7% │
│          900000 │               257.6 │                 173.9 │        48.1% │
│         1000000 │               287.1 │                 194.8 │        47.4% │
└─────────────────┴─────────────────────┴───────────────────────┴──────────────┘

                       Upsert a single user (upsert_user)
┌─────────────────┬─────────────────────┬───────────────────────┬──────────────┐
│ Collection size │ Prisma average (ms) │ Mongoose average (ms) │ % difference │
├─────────────────┼─────────────────────┼───────────────────────┼──────────────┤
│          100000 │                 9.4 │                   2.4 │       291.7% │
│          200000 │                 6.5 │                   3.8 │        71.1% │
│          300000 │                 9.3 │                   3.0 │       210.0% │
│          400000 │                 4.7 │                   1.9 │       147.4% │
│          500000 │                 5.8 │                   2.8 │       107.1% │
│          600000 │                 7.8 │                   2.5 │       212.0% │
│          700000 │                 6.8 │                   3.4 │       100.0% │
│          800000 │                 6.0 │                   2.9 │       106.9% │
│          900000 │                 6.6 │                   1.9 │       247.4% │
│         1000000 │                 5.9 │                   3.4 │        73.5% │
└─────────────────┴─────────────────────┴───────────────────────┴──────────────┘

Answered by garrensmith

Aug 16, 2022

@paulrostorp this is really useful for us. Thank you for taking the time to generate these results for us. Could you be able to open an issue for this so that we can track it and look to fix it.

View full answer

garrensmith · 2022-08-16T13:27:33Z

garrensmith
Aug 16, 2022

@paulrostorp this is really useful for us. Thank you for taking the time to generate these results for us. Could you be able to open an issue for this so that we can track it and look to fix it.

1 reply

paulrostorp Aug 18, 2022
Author

#14871

FrancescoSaverioZuppichini · 2023-07-25T09:44:18Z

FrancescoSaverioZuppichini
Jul 25, 2023

what about prisma 5?

0 replies

ludralph · 2025-01-21T18:56:16Z

ludralph
Jan 21, 2025

Hi there,

To keep our discussions organized and focused on the most relevant topics, we’re reviewing and tidying up our backlog. As part of this process, we’re closing discussions that have already been marked as answered but remain open.

If this discussion still requires further input or clarification, feel free to reopen it or start a new one with updated details. Your contributions are invaluable to the community, and we’re here to help!

For more details about our priorities and vision for the future of Prisma ORM, check out our latest blog post: https://www.prisma.io/blog/prisma-orm-manifesto.

Thank you for your understanding and ongoing support of the Prisma community!

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Prisma MongoDB benchmarks against mongoose #14811

Uh oh!

{{title}}

Uh oh!

Replies: 3 comments 1 reply

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Select a reply

Uh oh!

Prisma MongoDB benchmarks against mongoose #14811

Uh oh!

paulrostorp Aug 15, 2022

Replies: 3 comments · 1 reply

Uh oh!

garrensmith Aug 16, 2022

Uh oh!

paulrostorp Aug 18, 2022 Author

Uh oh!

FrancescoSaverioZuppichini Jul 25, 2023

Uh oh!

ludralph Jan 21, 2025

paulrostorp
Aug 15, 2022

Replies: 3 comments 1 reply

garrensmith
Aug 16, 2022

paulrostorp Aug 18, 2022
Author

FrancescoSaverioZuppichini
Jul 25, 2023

ludralph
Jan 21, 2025