Why is Slic3r PR so slow compared to S3D
Hello guys, I received a Prusa i3 MK3 few days ago and I'm doing some tests.
I never used Slic3r before, but I like it, even if it lacks of some important features, like the estimated print time without the need to export gcode (if I'm wrong, please tell me).
I like the fact that Slic3r has many profiles and material presets, but I do prints for clients, so print time is an important factor. I noticed a quite huge difference between Slic3r an S3D print times.
I used the standard profile for the i3 MK3 that S3D has and I increased many speeds in Slic3r and disabled the z-lift trying to matching S3D times, but with those changed I had:
Print of 10 small badges:
- Slic3r --> estimated 3:00 (don't know actual print time)
- S3D --> estimated 2:15; actual 2:40
Print of some quite tall but skinny objects:
- Slic3r --> estimated 7:00
- S3D --> estimated 4:00
S3D usually underestimate times, Slic3r I don't know.
First scenario could be acceptable if Slic3r time was exact. I printed both versions (but I didn't check the Slic3r actual print time) and quality is pretty much the same.
In the second scenario the difference is pretty huge.
Is there something I'm missing here?
Tomorrow at work I can do some other tests.
Thank you in advance
Re: Why is Slic3r PR so slow compared to S3D
From my experience if you set the same speeds/accelerations in Slic3r, you'll get very similar estimates. Feel free to use the free gcode analyzer to see the the difference.
Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram
Re: Why is Slic3r PR so slow compared to S3D
Any suggestion about which gcode analyzer to use?
Re: Why is Slic3r PR so slow compared to S3D
Any suggestion about which gcode analyzer to use?
Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram
Re: Why is Slic3r PR so slow compared to S3D
Maybe you should compare the actual print time for the same object.
Re: Why is Slic3r PR so slow compared to S3D
Just print the same part with Slic3r and S3D and note the actual print time (try to set the same parameters).
Slic3r is really close to reality for estimating (5% according to PRUSA).
S3D is really bad, I made a comparison on severals print, the difference vary from 30 to 100% (in the wrong way of course).
I know always use Sli3r for estimating.
I don't understand as well why you have to export Gcode to get the estimate on Slic3r.
Re: Why is Slic3r PR so slow compared to S3D
Maybe you should compare the actual print time for the same object.
I definitely should and I will.
Just print the same part with Slic3r and S3D and note the actual print time (try to set the same parameters).
Slic3r is really close to reality for estimating (5% according to PRUSA).
S3D is really bad, I made a comparison on severals print, the difference vary from 30 to 100% (in the wrong way of course).
That's good to hear.
Yes, I primarily use S3D and it sucks at estimates times.
I don't understand as well why you have to export Gcode to get the estimate on Slic3r.
I chose the preset, I change parameters I want to change, then I click "Slice now" and I can see the Preview. If i click "Export G-code..." and save the file, the "Sliced Info" box appears in the bottom right corner (with estimated time).
Maybe there is an option to change this. I will check.
Re: Why is Slic3r PR so slow compared to S3D
Did a test with Slic3r for the 10 small badges, after tweaking many speeds.
Estimated 3:04, actual 2:58.
Pretty close and precise.
I still don't understand why it's slower, since Slic3r have 200mm/s of infill and solid infill and S3D probably use the standard speed (80mm/s).
I increased speeds a lot compared to standard official values, but ok. I will do other tests.
Re: Why is Slic3r PR so slow compared to S3D
Gina Häußge, the maintainer of Octoprint, provides an explanation of why print estimations are difficult in this video at about the 44:40 mark. There are a lot of factors besides speed at play. I've noticed that Simplify 3D is consistently overly-optimistic (low). The only thing that counts is actual print time. Slic3r seems to get closer, although none are spot-on in my experience.
and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan
Re: Why is Slic3r PR so slow compared to S3D
It's very simple explained. If your estimated time vary more then 5-10% then your speed/jerk/acceleration settings in the slicer doesn't match what the printer (hardware) can handle.
For example: You can always say to the printer "print with 200mm/s and accelerate 5000mm/s ^2" but if you print with your MK2 in silent mode, the hardware will accelerate at 2500mm/s^2 and can reach only 100mm/s. So the real print will take twice as calculated at best.
My S3D estimates around 5% faster then the real print. It's all about correct speed settings matching your printer.
Often linked posts:
Going small with MMU2
Real Multi Material
My prints on Instagram
Re: Why is Slic3r PR so slow compared to S3D
I agree, they developed Slic3r PE especially for Prusa's printer, so that's why it's really precise with default settings.
The question of wladimiro is why S3D is faster than slic3r, I think it's not true.
Did a test with Slic3r for the 10 small badges, after tweaking many speeds.
Estimated 3:04, actual 2:58.
Pretty close and precise.
Did you tried the same parts with S3D ?
Maybe it will estimate 2 hours, but it will be 3 in reality, so it's not faster.
Re: Why is Slic3r PR so slow compared to S3D
Did you tried the same parts with S3D ?
Maybe it will estimate 2 hours, but it will be 3 in reality, so it's not faster.
Yes, it was in my opening post and I forgot to write it again alongside Slic3r results:
S3D --> estimated 2:15 hours; actual 2:40 hours
It's faster, but not by a wide margin. Same quality overall.
I will try to do other tests with longer prints (I suppose the difference would increase), but since we do prints for customers, I don't know if I will be able to.
What nikolai say sounds very true to me. I always struggle with those settings, especially for those less open machines. The best place to find speed/jerk/acceleration values is the firmware, right?
Re: Why is Slic3r PR so slow compared to S3D
As far as I remember (i might be wrong), the default S3D setting does not enable Z-hop while Slic3r does.
This may be part of the reason why Slic3r is slower.
Re: Why is Slic3r PR so slow compared to S3D
As far as I remember (i might be wrong), the default S3D setting does not enable Z-hop while Slic3r does.
This may be part of the reason why Slic3r is slower.
I can confirm it does. I already tryed to disable it, but didn't change that much in the prints I did. I will test more
Re: Why is Slic3r PR so slow compared to S3D
For me, I have been wondering where I could input the acceleration and jerk settings into S3D. The fact I can't change that makes the estimated print time dramatically incorrect.
As a reference, a 12 hour estimated print actually takes 19 hours to print.
Jonathan Kayne
Virginia Tech Class of 2021 - Electrical Engineering
Thingiverse Profile: https://www.thingiverse.com/jzkmath/about "I am always thinking about making. My future begins when I wake up and see the light." - Miles Davis
Re: Why is Slic3r PR so slow compared to S3D
For me, I have been wondering where I could input the acceleration and jerk settings into S3D. The fact I can't change that makes the estimated print time dramatically incorrect.
As with so many things, you can't in Simplify 3D. At least, not in a way that will impact the time estimates. You can put appropriate commands in the startup gcode, but estimates will be way optimistic. I just ignored them.
and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan
Re: Why is Slic3r PR so slow compared to S3D
I can confirm Slic3r is much slower than S3D - even when comparing the newest versions of each. A 20 hour S3D print took about 28 hours in Slic3r. This is real time, not estimated time.
For the test, I used the same object (the roof of a building, since I am printing several of them). I also used Prusa's S3D profile he posted on this forum and the default settings in Slic3r. I figured Prusa created all the profiles I am using, so any speed differences is completely caused by them (or the way the program).
I have not bothered to look much into it, but my guess is Prusa uses MUCH lower multipliers to adjust the speeds in Slic3r than they do in S3D.
In small prints, it is only a few minutes, so who really cares. But in large prints, it can be quite large.