• Happy holidays, folks! Thanks to each and every one of you for being part of the Tom's Hardware community!

[SOLVED] Laptop Outperforming Desktop (Help Please!)

Vlad25

Honorable
May 13, 2016
7
0
10,510
Hey guys,

I would really appreciate a bit of help here, as I just can’t understand what’s happening.
I own an Omen 30L GT13-0027nq (1 year old) with the following specs:
  • Intel® Core™ i9-10900
  • GeForce RTX 2080TI 11GB GDDR6
  • 32GB RAM (HyperX Fury 3200)
  • 1TB SSD NVMe (WD BLACK SDBPNTY-1T00_
I also just received my Lenovo Legion 5 Pro 16ITH6H with the following specs:
  • Intel Core i7-11800H
  • NVIDIA GeForce RTX 3070 8GB
  • 32GB RAM (SK Hynix HMAA2GS6CJR8N-XN)
  • 1TB SSD (SAMSUNG MZVL21T0HCLR-00BL2)
  • WQXGA -165Hz
When searching online, most benchmarking websites agree that the CPU & GPU on my desktop should outperform my laptops, or that they are quite close to each other:
Given the above generic benchmarks and the fact that one is a desktop, before buying my Legion 5 Pro, I assumed that the desktop will surely still perform better. I was very wrong.
Here are benchmarks ran on my desktop and laptop.
My Desktop benchmarks:
My Laptop benchmarks:
As you can see from the above results, my testing differences are quite significant and not in line with the generic benchmarks, especially with regards to the GPU 3D Mark (7281 points difference) and CPU (1294 points difference).

Also, there’s a huge difference between the desktop 2080TI and the laptop 3070 GPU’s (the laptop GPU outperforms the desktop one by 6852 points in 3D Mark).

Is there something wrong with my desktop GPU? How else can we explain these big differences?

On my desktop, I've also ran HWiNFO with Prime95 and Furmark. These are the desktop results.
 
Last edited:
Solution
"Performance Limit - Thermal = Yes."
That 2080Ti is getting ~TOASTY.
Gpu core temperatures running in the upper 80s(C), with hotspot running up to 106C.
Memory reading not available without using equipment, from the looks of it. It MIGHT not be as bad though, since its frequency never dropped.

FYI:
The gpu core's default limit - fans ramp up, Gpu Boost stops trying to boost - is 84C. The critical limit(shut off) is 88C, but it has to sit there for a bit for a shut off to happen.
Hotspot has a critical limit of 110C.



What's probably going on here is that the gpu can't breathe.
Card needing a repaste? Maybe, but it's not unusual for gpu core and hotspot to have a 10-20C gap.
Easy check:
Get a fan,
nI6vx5v.jpg
...
Userbenchmark... BAH, I'll take passmark as more accurate.

In passmark SSs, it all looks ok, except Gpu Compute.
The why or how? IDK, there's not much else to go on... Did you try to look for anything abnormal beforehand - such as, gpu core thermals, gpu hotspot, Vram thermals... perhaps gpu boost clock stuck at a low clock?
 
"Performance Limit - Thermal = Yes."
That 2080Ti is getting ~TOASTY.
Gpu core temperatures running in the upper 80s(C), with hotspot running up to 106C.
Memory reading not available without using equipment, from the looks of it. It MIGHT not be as bad though, since its frequency never dropped.

FYI:
The gpu core's default limit - fans ramp up, Gpu Boost stops trying to boost - is 84C. The critical limit(shut off) is 88C, but it has to sit there for a bit for a shut off to happen.
Hotspot has a critical limit of 110C.



What's probably going on here is that the gpu can't breathe.
Card needing a repaste? Maybe, but it's not unusual for gpu core and hotspot to have a 10-20C gap.
Easy check:
Get a fan,
nI6vx5v.jpg
2GBempv.jpg


and take off the Omen's side panel. [Opening it up to do this shouldn't be voiding any warranties.] Aim the fan to blow full blast inside the PC and test thermals again.
If core and hotspot drop more than 5C, then you likely have an airflow problem.
If they barely change, then it's the cooler, paste, or something inside it(like dust).

What to do after that? If it's still under warranty, it'd be best to take that up with HP first...
 
Solution
Thanks a lot for your input, Phaaze88! I've uninstalled all my Nvidia drivers and it's actually working great now. GPU compute jumped to around 11k now and all seems generally better.

I guess it was just a software fluke after all. I will try to look at termals though as well, if you say they might be of concern.