processor speed vs turbo

[StephenW]

It seems as if our IT dept has decided independently that they can save a lot of money on getting computers with slower processors as long as the processor has a good Turbo speed.

I've never paid much attention to the Turbo speed when specifying a processor for Creo.

What are the thoughts on Creo using the processor Turbo?

This is a co-workers new workstation, I'm going to ask him to run the creosite benchmark on it and on his old computer too. 

I just can't imagine this being an upgrade, but maybe I just need an education...not sure.

The processor in our specific situation is a Xeon W3-2423 2.1 GHz base speed. My current workstation (~3 years old) uses a Xeon W-2223 @ 3.6 GHz base speed.

https://www.intel.com/content/www/us/en/products/sku/233484/intel-xeon-w32423-processor-15m-cache-2-10-ghz/specifications.html

[rreifsnyder · Accepted solution]

I'm sorry but both of those Xeon processors are not good for Creo performance. As has been addressed many times in the community and with PTC, Creo is MOSTLY single thread processing. I use the Passmark website www.cpubencmark.net to compare cpu's. I would tend to i9 or even i7 processors. There are other differences between computers designed for Xeon vs. the Core series that might favor the Xeon's but the difference in the single thread performance is so stark I would recommend the (by the way cheaper) Core series. The i9 in the comparison I'm posting is what is currently spec'd for our engineers. Even the newer Xeons are not as fast in single-thread operations.

If you can, I would also recommend trying to get loaners of the computers you are considering and try running the OCUS benchmark (www.creosite.com)

[StephenW]

I wouldn't doubt what you are saying. Many years ago, our setups compared reasonably vs. high end custom stations based on the creosite benchmark. I suspect that the last couple of rounds of replacements have degraded performance

When I look at the i9 core, I see 7 processor speed. An i7 processor speed "base" is faster but max turbo is slower. Comparisons are more difficult than they used to be. Does Creo actually utilize Turbo within the processor, or better question, how does the processor determine what runs at all the levels turbo that it offers?

[rreifsnyder]

I have never taken the "turbo" into account since we don't have actual single thread performance data for the turbo. It's a somewhat reasonable to talk about the clock speed because it used to be that there was a pretty good correlation but since clock speeds mostly have topped out I just don't know the impact anymore so if I can't run the actual OCUS I tend to follow the published single thread performance.

Several years ago I inserted myself into the workstation selection process at my company and showed that the workstations that they were considering were more expensive and slower but I actually had to fight because there was a concern that the workstation "number" that I was recommending was in a lower series (200 series vs. 300 series) and so the users would complain. It never happened because the users recognized that the new computers were so much faster than the old computers they had in the 300 series.

[StephenW]

Thanks for the input. Its been so long since I have had input, I didn't realize what was going on.

I've asked the user to run OCUS on his old and new computer. I'm not in the office now so I don't know if he will be able to accomplish the task.

I think we may be in a similar situation  with "series" issue you had. We've been "stuck" on HP Z4 (or it's equivalent) for years. It looks like I can customize the Z2 with better specs. Maybe it was always the case and we just didn't do the research. I haven't inserted myself in the selection process in years...I suppose its time again.

[ReidT]

I just got done running an OCUS benchmark v7.1.1 with Creo 10.0.4.0 on a new Dell XPS 8960 and was presently surprised by the results. Results show an 11% speed improvement compared to our current Gruman Creations 2209 machines.