Dell XPS 15 9570 benchmarks + temps

Put some responses in red above.

 

Must have been @Eason

 

I did some iGPU undervolting but that caused some latency issues as documented on the "ThrottleStop Guide". In my music workflow, low latency is most important so I do not undervolt iGPU these days.

You should give it a try on the 8th gen to see if it helps reduce temps and power consumption. Those are big issues for the 9570 so every bit counts...assuming it does not cause other issues.

 

So it looks like the DIMM in AIDA64 actually corresponds with the first Ambient sensor in HWINFO. Not sure which is labeled correctly, but the DIMM in HWINFO seems to correspond with severe throttling. My guess is that's actually a VRM sensor of some type.

 

I thought there was a post and photo here or at Reddit from someone who confirmed DIMM was throttling at low 60s and used some compressed air can upsidedown to make sure that was the sensor. I can't confirm

Of course, the 9550 VRM sensor was mislabeled in a few monitor programs when released (GoNz0 had Aida64 fix that)

 

There was a guy on reddit with a 9560 that padded his DIMM sensor to keep it from throttling, but no one I know with a 9570.

 

benchmarks and stress tests updated: https://www.ultrabookreview.com/20435-xps-15-9570-review-live/

no significant PL throttling after 20 minutes of unigine heaven 4.0 loop; had to run TS bench concurrently to cause PL throttling. DIMM sensor didn't go past 62C, so I wasn't able to trigger VRM throttling either.

 

Good to hear, and I agree with your comments about keeping the bigger battery and adding another m.2 SSD slot, makes sense to me.

 

My 9570 (first XPS 15) arrives this week. Curious for those that have owned other models whether the iUnlock method of channeling fan airflow + heatsinks or the "stacking pads to the use the backplate as a heatsink" (https://www.ultrabookreview.com/14875-fix-throttling-xps-15/) method is the better approach for cooling the VRM. Thoughts from those who have done both? Perhaps we could test both methods in this thread and put numbers around them?

 

@pressing thanks for sharing your knowledge with the 9550 and 9560. Now we need to make sure which sensor is which on the board and if we can cool that area.
I really didn't expect to have 6*C threshold on the RAM.

Has anyone tried padding or blasting the problematic areas with a fan?