I have ran cinebench r23 trying to see what kind of single core performance I would get and it’s absolutely terrible with a score of 945pts . This is a new system I got at work and my work laptop feels faster and performs better it has i7-11800H it has 2 less cores and out performs the 10core W-2255 in single core and multi core because it will actually reach 4.5ghz. Does anyone know why this Xeon processor will not boost at all ?
Which power plan is it in? I’ve had some wildly different performance between plans on some laptops.
The W-2255 is in a desktop and I have all power settings maxed out and also have turbo boost enabled in the Bios and it’s a Dell… Sorry I did not make that very clear. But my laptop is performing better than
my desktop…
No worries. I’d even looked at the ARK page and missed it was a desktop chip. Blaming eyeballing too many P-series thinkpads with Xeons lately, and not enough coffee. 
Which model of Dell? From poking around a bit the 5860’s seem to have had something like you’re describing that needed a bios update, but that’s a newer chip (reddit thread). Otherwise it sounds like you’ve covered all the stuff I would suggest, hopefully someone else has some hands-on with those chips.
More data is needed, firmware, powerplan, cooling, power delivery, it can be literally anything.
Run hwinfo portable sensor mode in background while you run your benchmarks.
Look for suspicious temperatures, how much power CPU reports being used, real frequency max and average values etc.
It’s a Precision 5820, I got it about a month ago , I have no clue on when it was purchased but it’s already discontinued. I may try to reach out to them but there scan say’s everything is ok . I believe I already did update the Bios , but I will have to double check just to make sure. I also think I have done everything I can but no luck yet.
I also don’t know why are IT ops for the XEON line, we really don’t need it , I feel like a 13900k would be much better suited for the work we do in CAD/CAM and would be more cost effective.
Just run the monitoring tool mentioned above, if cpu throttles under load, it will tell you and show you.
Since its second hand unit, did you open it up to check for dust and blockages? It might be that simple.
It’s max temp on the package was 78c and went down longer in the run settling down to 70c , the core temps never maxed over 76-77c while running R23 in multicore. However when using single core benchmark I noticed that not any of the 20cpu’s was at 99% or any above 30% so I actually could not tell what CPU R23 was using in single core.
There was only one Core out of 10 that went to 4.5ghz in the multicore run and I would of never noticed it if it wasn’t in the max value. I was watching the MHZ very closely during the run and all of them stayed around the base 3.7ghz. except the one that shows 4.5 in the max column and it must of been very very brief .
The PC is new I just don’t know how long ago work purchased it and the time it took for them to give it to me . It could of been a few months.
And total power and the throttling values? This can be expected platform behaviour.
Note down both and try some other workloads, like good old P95 and linpack for example.
OCCT is anoying with its constant nagging, bt useful in this case.
I almost never see any Xeon hit max frequency unless I shut the other cores off in the BIOS or over clock (e5-16** v1 to 3 and the latest unlocked multiplayer Xeons). If you need more single core performance try shutting down cores or turning off hyperthreading. Remember that the Xeons are designed for mission critical use cases…says the guy who runs his e5-1660 v3s at 4.4ghz…but not my platinum Xeons that are grinding v-100s for days.
This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.