Apparently the Dell G5 Special edition gaming laptop has been improved with BIOS and driver updates. A lot of you have asked me to retest this machine with the latest updates in place, so I’ve fully updated it and run all my tests again to see what sort of improvements Dell have made to gaming performance and thermals.
I originally tested this G5 in June 2020, so about 6 months ago now. The games I’m testing on this machine have also had updates since that time, so some of the performance changes could be due to that.
Temperatures
The temperatures were an area where the G5 did poorly in the past, running in excess of 100 degrees Celsius. 
Now while the updated G5 still had peaks above 100 at times, when running a combined CPU plus GPU stress test over the course of an hour, it was actually a fair bit cooler with the updates in place, which is shown by the December result. The CPU was around 8 degrees cooler, while the GPU was 10 degrees cooler, a substantial difference. Both were tested with a 21 degree Celsius ambient room temperature, so it would appear that this difference is due to updates as I’m running the same test under the same conditions.
Clockspeeds
When we look at the clock speeds during these tests, we can see that the CPU was about the same, just a 1% higher speed in the updated December testing. The RX 5600M GPU on the other hand wasn’t reaching as high clock speeds, it would seem that they may be running the graphics with a lower power limit to prevent it getting as hot.
Cooling + Smart Shift
The G5 has shared heatpipes, so by not smashing the GPU as hard it would also affect the CPU. The G5 also makes use of smartshift, so it moves power between the CPU and GPU as needed based on the workload, and I suspect this is most likely what’s been tuned over time. So if the GPU isn’t clocking as high anymore, then the next question becomes, how is gaming performance affected?
Copyright TechPlanet.today
11 Game Compared
Let’s take a look at 11 different titles and see the differences.
Shadow of the Tomb Raider was tested with the games benchmark tool. 
I’ve got the older results from June shown by the purple bars, and the newer results from December in red. At the lowest setting preset the GPU is likely not being hit hard yet compared to the CPU, so it makes sense that the frame rate might not be different. As we move up to higher setting levels though, the updated system starts performing a little worse. At the highest setting preset the older config was only around 5% faster, so not a big difference. Don’t worry, there were instances where the updated system performed significantly better.
Take Battlefield 5 for instance. 
The difference in average FPS wasn’t that big at the higher setting levels, however the 1% low improvement is quite large comparatively. 1% lows generally get boosted in laptops with better CPU performance, so this may be a result of the CPU running cooler or possibly by it being able to take more power with SmartShift. If the GPU is using less, the CPU may be able to get more. 
The Witcher 3 was also showing huge improvements with the updated system. With the ultra setting preset the updated G5 was reaching 15% higher average FPS, and at low settings it had a huge 35% lead, one of the biggest differences seen out of all games retested. 
CS:GO was slightly worse with the updated system at all setting levels. The average FPS was closer together compared to the 1% lows, which had a larger drop off with the updates in place. 
Call of Duty Modern Warfare was performing almost exactly the same with all settings at minimum, but with all settings maxed out the older result was about 13% faster than the newer updated result. 
Borderlands 3 was tested with the games benchmark. Basically the same performance with the lowest setting preset, then in terms of average FPS the updated system in red was a little behind my original test results from 6 months ago. 
Control had similar performance at both low and medium settings, then with all settings maxed out, which is fairly GPU heavy from my past experience, the updated system was again behind the older results. 
Far Cry New Dawn was another game where the results were better with the updates regardless of setting levels in use. This one typically depends more on CPU than GPU performance, so that may be why the updated system was ahead here.
Assassin’s Creed Odyssey was also a little ahead at all setting levels with the updated system, and again this is a test that I’ve typically found to perform better with increased processor performance than graphics.
Metro Exodus was about the same at low settings, then the updated system was behind at the rest of the setting levels, a common pattern that most of the games we’re looking at here seem to exhibit. 
Overwatch was similar, though the average FPS was higher at low through to high settings, then the updated system was only behind at ultra and epic presets.
Game Differences - Max Settings
These are the average results out of all 11 games retested with the highest setting preset.
On average the updated system was around 3% slower than the older one, only a small number of games were actually performing better with all updates in place. I think these results make sense based on the clock speeds noted earlier, the GPU isn’t clocking as high, and at higher setting levels this is where the GPU is going to have more of an effect.
Game Differences - Min Settings
If we instead look at the differences between the same games but with the lowest setting preset instead of highest, the updated system is actually performing more than 5% faster now.
Processor performance generally matters more in games at lower setting levels, and if the GPU is using less power then according to the way smartshift works the CPU will be able to use that power and perform better. Combined with lower temperatures and less potential thermal throttling and I think this is why we’re seeing better results here.
CPU Only Performance
Outside of gaming, CPU only performance appears to be the same. 
So this definitely seems to be a change to the way smartshift works and how power gets assigned to the different components when both the CPU and GPU are active, like when gaming.
Conclusion
So basically the updated Dell G5 special edition seems to run cooler than when I originally tested it closer to launch 6 months ago, but at the expense of some gaming performance at higher setting levels. Games at lower settings on the other hand did tend to perform a bit better with the updated system, even worst case I don’t think we’re losing that much performance with the updates in place.
I think it’s probably worth it for the cooler temperatures, which may improve device longevity, but let me know what you think down in the comments. Would you prefer a hotter machine and better performance, or do you think the new updates offer a better mixture?
If you want to find out more about the G5 special edition then you can check out my detailed review.
No comments yet