Dell Precision 5540 one-month review

Given that you have double the number of cores of my old i7-3840QM, you're probably still faster for the parallel workloads I'm interested in, despite power saving mode.

Every benchmark I've seen indicates that the 5540 + i9 H/HK is thermally constrained (i.e. will never clock above 4ghz on all cores) so the 20-30% performance hit may in fact result in a quieter machine with less of a performance hit than expected.

Curious to see if this is the case, what's the max clock you get in "quiet" mode?

 

Ran a few Geekbench scores.

Dell Power Scheme:

Thermal Management to Optimized

-> Battery slider to Best Performance:
Geekbench 5262 / 27256. temperature stays most of the time around 60C, but spikes close to 100C, Fans work hard inbetween.

-> Batter slider to Better Performance
Geekbench 4993 / 26157, don't see a big diff in the temperature curve, towards the end of the benchmark fans are on high alert.

-> Batter slider to Better Battery
Geekbench 4984 / 26046, here the noise is much better. Temperature is getting controlled (from what i can see i Speccy) much more with the core speed, speed rarely goes above 4Ghz. In the 2 previous tests, speed was often at around 4.3 to 4.5

Thermal Management to Quiet.

-> Battery slider to Best Performance:
Geekbench 4761/ 14023. temperature never goes above 60C, and i can barely hear the Fans.

-> Batter slider to Better Performance
Geekbench 4719/ 14803, might be a fluke, but i did notice higher temps around the LLVM part and the fans turned on. Not really that much a difference though.

-> Batter slider to Better Battery
Geekbench 4501/ 14526, silent.

Did one test with Thermal management set to Ultimate Performance, slider to the best performance
Geekbench 5191 / 26655 - not really different from the "Optimized" scheme. Not louder, not hotter, not faster.

This is obviously not scietific, i had stuff running in the background (not a naked windows), i did only one run, the CPU might have been hotter to start one test than the other. But as a ballpark this is useful to me at least.

Frank

 

Well, that's wonderful, would certainly take the performance hit in exchange for a quiter machine (all the more so that I'll have 2X number of cores compared to current machine).

Not sure if the same level of thermal control can be had on Linux, however (best one can do is undervolt and hope OS power management is effective).

One last thing to check out would be the effect of hardware/BIOS changes. Think with Precision machines we can choose power saving mode in the BIOS.

At any rate, great to hear that if one wants it a cool & quiet machine is possible despite high power components in the 5540 (with expected performance hit).

 

One more performance update. I am currently building the swift compiler/runtime on the 5540. That is a long running, cpu and i/o heavy task, and my cpu is at around 100% for quiet a bit now (> 20min). I have the power setting on max, and i am getting all cores at 100%, but with just 2.3 - 2.5 ghz per TaskManager. Speccy tells me around the same, the core temp is at 70C, the fans are howling.

Disappointing to be honest. For a machine that loud i would have assumed to get a tad more power out of it. Dell could improve their termal handling a bit.

Frank

 

Looks like the thin chassis doesn't provide same cooling capacity as the thicker 7540. 2.5ghz is a pretty unfortunate number for sustained CPU-intensive tasks, that's exactly the work I do

At least there are 8 cores so if the task can be run in parallel then it should be completed faster than CPUs that run at higher clock speed and fewer cores.

No choice here, come November the 5540 will be the new work machine. Shedding 3+ pounds of weight will be nice, but loud fans and slow CPU "boost" speeds , not so much...

 

Anyone know why the 5540 is offered with a Xeon, but no ECC RAM?

I have the 5530 with the i9 (no ECC), but recall that I could have opted for a Xeon+ECC ram - but sadly I didn't configure it with it - as I intended to resell it.

However, I'm finding that the 5530 is so nice in comparison to the HP Zbook Studio x360 G5 that I have been trying so hard to make work!

Even the TB16 worked perfectly today - no crazy warnings like the HP with their TB3 dock!

 

5530 didn't support ECC RAM either, and was not sold with this as an option. Same story with the 5520 and 5510. People have tried to install ECC RAM in these systems with Xeon CPUs and they will not boot. It is likely that Dell didn't include the extra traces needed to hook the ECC pins up to the memory controller in the CPU.

This is sort of silly all around, because the only benefit to getting the Xeon over the same-clocked Core CPU is ECC memory support.

 

Understood - and thanks!

So, not a bad decision to get the i9 - other than being thermally limited by the cooling config....

Now to decide whether to stick with the Studio x360, the 5530 or go for the 5540! (i9 is supposed to have the better cooling solution, I read...)

 

ha, ha, good one, read above 5540 owner's latest performance review

Not only will the i9 throttle under load, it will throttle badly (all the way down to base frequency), and the fans will make a great deal of noise in the process. If the 7540 weren't such a heavy tank I'd pick up one of those instead of the 5540.

 

Can some please tell me about the brightness of the FHD panel? It is in the order of 300 or 500 nits?

If it's similar to the XPS 7590 than I would think that having the i5 + FHD reduces heat and increases battery life.