• I want to thank all the members that have upgraded your accounts. I truly appreciate your support of the site monetarily. Supporting the site keeps this site up and running as a lot of work daily goes on behind the scenes. Click to Support Signs101 ...

best computer for versaworks7?

Bryce I

I'm Brie
We've had issues with computers moving real slow when using versaworks. We're about to replace the computer that is mainly used for ripping and printing. What to get? Versaworks is now supported by macs. Anyone using that? When buying, what are the most important factors to consider, so that the software will run fast and smooth?
Any insights welcome, thanks!
 

Solventinkjet

DIY Printer Fixing Guide
what are the most important factors to consider
CPU:
Higher clock speed (4 - 5 GHz)
Large L3 cache (32-64mb)
Most CPUs come with at least 4 cores these days which is plenty for a RIP only station. If you plan on using it to RIP and print to multiple machines at the same time or the computer is being used to design and run other equipment, more cores helps
A decent CPU cooler or water cooled helps

Memory:
16 - 32+GB, Higher is better if you do large banners, wraps, or run multiple machines
Memory speed doesn't matter as much with RIPs. If you can get DDR4 cheaper than DDR5 and add more GBs, that's preferable

Storage:
NVMe Drive of whatever size you think you need for your volume of output. Some people get an old HDD or SSD for longer term storage but go with an NVMe for the main drive that will be doing the RIPing

Video cards don't help RIPs. So don't go out and get a 5090 or anything.
 

netsol

Premium Subscriber
there is a plateau we reach around 3.6ghz where a silicon chip is reaching maximum
beyond that speed it pretty much just generates a lot more heat and not a lot of useful COMPUTE speed
with hyperthreading we can now have 20 or more cores in a cpu (i have a fairly inexpensive Dell 7000 series micro on my desk with 20 cores. <not really
the speed demon it sounds like>)

not sure why we can not have 4 cores doing the rip functions, but we don't

someone who is more familiar with this issue than i am can explain, is it just that you would need 4 licenses if we had 4 cores performing RIP functions?
 
  • Agree
Reactions: 1 user

Solventinkjet

DIY Printer Fixing Guide
not sure why we can not have 4 cores doing the rip functions, but we don't
The RIP process is actually multiple serial and parallel processes in one. The large majority of the process is serial and can only be run on one thread. So if you have a 4 core RIP station and look at CPU usage during RIPing, you'll see one core pegged to near 100% running the serial process and the other 3 will be much lower in the 10-30% range running the parallel processes. So the limit is more a computer science issue than a license issue. That being said, a lot of RIPs will require you to buy another license to run multiple printers at the same time and RIP multiple files at the same time. That's more of a business decision than a computation limitation though.
 
  • Agree
Reactions: 1 user

netsol

Premium Subscriber
going back to the late 1990's you needed an additional license for each core (both on windows NT and Unix)
on unix you started each task manually, with <alt F1, alt F2, etc> in the days before hyperthreading

they were in court for years because microsoft wanted an additional full price license for each additional core (whether it was used or not)
 
  • Agree
Reactions: 1 user

WildWestDesigns

Active Member
The RIP process is actually multiple serial and parallel processes in one. The large majority of the process is serial and can only be run on one thread. So if you have a 4 core RIP station and look at CPU usage during RIPing, you'll see one core pegged to near 100% running the serial process and the other 3 will be much lower in the 10-30% range running the parallel processes. .
This is also going to depend on how interconnected the tasks are. Start trying to do things concurrently, have sharing sync issues, data races etc. Especially when talking about trying to implement things in C/C++ that don't tend to have the same training wheels that languages like Rust have. It is a CS issue and I don't see it actually improving really as there is a more modern dependence on "training wheels", which usually has other issues as well. Even going the Rust re-write for certain areas is going to have complications as well.
 
Top