rluvaton opened a new issue, #9083:
URL: https://github.com/apache/arrow-rs/issues/9083
**Is your feature request related to a problem or challenge? Please describe
what you are trying to do.**
Yes, I'm using RowConverter and it start to become a bottleneck,
some of the use cases of row conversions do not need ordering
for example:
1. grouping - grouping by a lot of columns does not need ordering
requirements
2. aggregation - using `count(distinct <struct>)` or array_agg with
distinct, both of which do not need ordering requirements
3. shuffling
and having ordering requirements limit optimizations.
Bottlenecks that I found and how the ordering limit optimizations described
below.
**Describe the solution you'd like**
<!--
A clear and concise description of what you want to happen.
-->
**Describe alternatives you've considered**
<!--
A clear and concise description of any alternative solutions or features
you've considered.
-->
**Additional context**
<!--
Add any other context or screenshots about the feature request here.
-->
-----------
All my benchmarks and profiling were done on the most reliable machine I
could get:
```
c5.metal
```
<details><summary>Env</summary>
<p>
```bash
[ec2-user@ip-172-31-21-167 build]$ ./fastfetch -c ../presets/all.jsonc
ec2-user@ip-
-------------------------
, #_ OS: Amazon Linux 2023.9.20251208 x86_64
~\_ ####_ Host: c5.metal (00001)
~~ \_#####\ BIOS (Legacy): 1.0 (4.15)
~~ \###| Chassis: Rack Mount Chassis
~~ \#/ ___ Kernel: Linux 6.1.158-180.294.amzn2023.x86_64
~~ V~' '-> Init System: systemd 252.23-10.amzn2023
~~~ / Uptime: 3 hours, 52 mins
~~._. _/ Loadavg: 0.04, 0.08, 0.05
_/ _/ Processes: 825
_/m/' Shell: bash 5.2.15
LM: sshd 8.7p1 (TTY)
Terminal: /dev/pts/0
Terminal Size: 160 columns x 57 rows (3200px x 2508px)
Terminal Theme: #ACB2BE (FG) - #21252A (BG) [Dark]
CPU: 2 x Intel(R) Xeon(R) Platinum 8275CL (96) @ 3.90
GHz
CPU Cache (L1): 48x32.00 KiB (D), 48x32.00 KiB (I)
CPU Cache (L2): 48x1.00 MiB (U)
CPU Cache (L3): 2x35.75 MiB (U)
CPU Usage: 0%
Memory: 2.00 GiB / 188.52 GiB (1%)
Disk (/): 15.36 GiB / 255.93 GiB (6%) - xfs
Date & Time: 2026-01-01 11:53:52
Locale: C.UTF-8
Network IO (enp125s0): 1.75 KiB/s (IN) - 4.69 KiB/s
(OUT)
Disk IO (Amazon Elastic Block Store): 0 B/s (R) - 4.00
KiB/s (W)
Physical Disk (Amazon Elastic Block Store): 256.00 GiB
[SSD, Fixed]
Version: fastfetch 2.56.1-16 (x86_64)
```
```bash
$ ./cpufetch --verbose
Name: Intel Xeon Platinum 8275CL
Microarchitecture: Cascade Lake
Technology: 14nm
Max Frequency: 3.900 GHz
Sockets: 2
Cores: 24 cores (48 threads)
Cores (Total): 48 cores (96 threads)
AVX: AVX,AVX2,AVX512
FMA: FMA3
L1i Size: 32KB (1.5MB Total)
L1d Size: 32KB (1.5MB Total)
L2 Size: 1MB (48MB Total)
L3 Size: 35.75MB (71.5MB Total)
Peak Performance: 11.98 TFLOP/s
```
</p>
</details>
------
When I tried to improve performance of row conversion I first started with
having the easiest case but with multiple columns
also all benchmarks run with:
`.cargo/config`:
```toml
[build]
rustflags = ["-C", "force-frame-pointers=yes"]
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
