# MacOS RX580 Scissors Segment very slow
**URL:** https://discourse.slicer.org/t/macos-rx580-scissors-segment-very-slow/27852
**Category:** Support
**Created:** [February 16, 2023, 7:10am UTC](https://discourse.slicer.org/t/macos-rx580-scissors-segment-very-slow/27852 "2023-02-16T07:10:12Z")
**Posts on this page:** 6
**Page:** 1
### Author:  [@zyy](https://discourse.slicer.org/u/zyy)
#### Post date: [February 16, 2023, 7:10am UTC](https://discourse.slicer.org/t/macos-rx580-scissors-segment-very-slow/27852/1 "2023-02-16T07:10:12Z")
my computer
I want segment the vessel from cta film,but scissor tool so slow compare by in windows.
windows just need 1 second will give response.mac need 3 second to 1 minute or more，It is unormal.
[cta image file](https://drive.google.com/file/d/17s8dvz2-1PI8mKNaV9Jx3OOXI5ZBImSv/view?usp=sharing)
---
### Author:  [@zyy](https://discourse.slicer.org/u/zyy)
#### Post date: [February 16, 2023, 8:29am UTC](https://discourse.slicer.org/t/macos-rx580-scissors-segment-very-slow/27852/2 "2023-02-16T08:29:12Z")
3DSicer Version is 5.2.1
5.2.1 r31317 / 77da381
---
### Author:  [@pieper](https://discourse.slicer.org/u/pieper)
#### Post date: [February 16, 2023, 3:05pm UTC](https://discourse.slicer.org/t/macos-rx580-scissors-segment-very-slow/27852/3 "2023-02-16T15:05:26Z")
I have a similar mac pro and it took only a few seconds to perform the scissors operation. If it takes a very long time you could use the ActivityMonitor to sample the process to see what operation is taking time. For me it was too quick to even measure. Maybe there’s something else going on with your machine?
---
### Author:  [@zyy](https://discourse.slicer.org/u/zyy)
#### Post date: [February 17, 2023, 1:53am UTC](https://discourse.slicer.org/t/macos-rx580-scissors-segment-very-slow/27852/4 "2023-02-17T01:53:26Z")
seem e-core in cpu not work
maybe only Intel 12gen、13gen CPU in hackintosh have this problem,
becuse another computer have amd 5900 in hackintosh work well.
macos no official support Intel big.LITTLE cpu framework,
current solution is opencore,it closed CPU HT,and set logic core number to thread number.
13600kf have 6 p-core and 8 e-core 20 thead,in OpenCore 8.8 set it 20 logic core
---
### Author:  [@zyy](https://discourse.slicer.org/u/zyy)
#### Post date: [February 17, 2023, 1:39pm UTC](https://discourse.slicer.org/t/macos-rx580-scissors-segment-very-slow/27852/5 "2023-02-17T13:39:40Z")
I saw this issues
> [@Performance issues on Mac M1?](https://discourse.slicer.org/t/performance-issues-on-mac-m1/22564/11):
>
> Thanks for reporting. There does seem to be some instability on macs with Apple chips. I have run some simple tests and things run fine, but we have seen [previous issues](https://github.com/pieper/SlicerParallelProcessing/commit/2cf7d7d61e7b17cc949cb8fd31acefe9e4de25a1) indicating that the on-the-fly run time code conversion is not the same as running on native hardware. So until Apple fixes their emulation layer or Slicer is ported to the new architecture there are likely to be edge cases that fail. I tested the basic flow described by @semredogan and got the crash report below basically i…
and I test this film in macbook pro 16 m1pro(Slicer5.2.1)
scissors operation is twice as fast as 13600kf in hackintosh.
but slow too much than 13600kf in windows11(too quick to even measure),
and do 3-6 times big scissors will crashed.
m1pro in macos problem looked so same with 13600kf in hackintosh.
m1pro also is big.LITTLE design,have 8 p-core,2 e-core.
So I suspect big.LITTLE design is the reason
If m1pro could fix this problem,please take intel 13gen cpu in hackintosh,thanks a lot.
white mac is too expensive in memory and disk for medical image developer
---
### Author:  [@pieper](https://discourse.slicer.org/u/pieper)
#### Post date: [February 17, 2023, 3:12pm UTC](https://discourse.slicer.org/t/macos-rx580-scissors-segment-very-slow/27852/6 "2023-02-17T15:12:26Z")
Interesting - thanks for reporting. I don’t have experience with hackintosh but I can believe there would be CPU related performance issues. I agree mac computers are not cost effective for development, and I typically use linux for this kind of work.