Benchmarks, Performance, Temperatures and Power Consumption

OK, so that you have a reference point, check out my system specifications.

Aside from the specifications I provided earlier, here are the GPU-Z readings on the card.

Here is much of the same from Sapphires own utility TRIXX.

I will utilize this program a little later to unlock some extra performance a little later in the review, so stay tuned.  You can grab a copy of Sapphire’s TRIXX 3.0 here:

You can see that the Sapphire Radeon PULSE RX550 4GB is based off of the Lexa chipset, part of the Polaris line.

Before I get started with the benchmark, I wanted to show you how my testing process goes.  First off, during all of my benchmarks, I capture the power consumption of the entire system.  For these tests I capture the minimum power used, average power used and the maximum power consumed.  To get all of these readings, I use “Kill A Watt” by “P3 International“.

I use the following applications to benchmark

OK, let’s get started benchmarking

Now, I did mention early on that this is a budget GPU, but it still does relatively well under the “Gaming PC” section of the Run Details.  Sadly, the card has not yet been entered into FutureMark’s database, so I entered it myself, but I guess it takes a while to register.

During this test, the lowest amount of power consumed by the PC was 71.1 Watts, that is amazingly low, for an entire PC with the specs listed previously, I am very surprised honestly.  The hottest the video card got in this test was 54°C.

Part of the 3DMark Suite, let’s see what TimeSpy says.

Well, not really high but again, think budget here.  How does it compare to other results?

Performing better than 7% of all other results.  TimeSpy heated up the card to 68°C at an average of 105 Watts consumed.  Keep in mind, 3DMark and TimeSpy are only benchmarks, not games, so let’s check out this card performs in games.

For the rest of benchmarks, I have adjusted the settings a little different than I usually do for the higher end video cards.  I will use this benchmarking method for all budget video cards moving forward and of course I will be listing below all the settings I have made.  For performance cards, I would use ULTRA settings, for budget PC’s I will use Medium/Normal settings which might be a little more realistic for more people.

So now let’s get on to Metro Last Light.

Here are my presets for Metro Last Light.  The only thing I will change for each test is the resolution and each test will be run in 1680×1050, 1920×1080 and 2650×1440 and presenting you with the results, power consumption and temperatures.

For a budget GPU, this really does amazingly well at 1080P, 14.26% above the almighty 60FPS and best of all at normal settings.  At 6850 x 1050, the FPS count was very high at 78.73FPS and I have that in there just in case you don’t have a 1080P monitor, but know that you can turn up the eye candy just a tiny bit more.  At 2K though, it was a bit rough at 42.87FPS though kept on par with the 1080P’s thermals at 74°C and only slight higher power consumption at 138Watts over 1080Ps 131Watts. The sweet spot here is 1080P, and there is nothing wrong with that.

I am a bit surprised how well this card handled Metro, again with the newer settings, but now let’s see how Thief handles.

Here are the settings I used throughout the benchmark, only changing the resolution in between.

It looks like Thief is having no part of this card, chugging a bit even at 1680 x 1050 coming in at 52.2 frames per second at a chilly 67°C.  Bumping up the resolution somewhat at 1920 x 1080 we drop down to 42.5 FPS, performance dropping by 18.58% though the wattage did drop 9.15% which is always welcoming.

In Thief, the “Normal” preset leaves some things on High and Default and leaves Tessellation enabled, which would explain the performance degradation but I don’t want to change the presets.  The only thing I do change on presets is the disable v-sync.

Well, I hope Laura can help a little here.

The settings look very similar to that of Thief, but the normal preset is a bit different, but let’s see what the results tell us.

Wow, Laura took the power back here, stole it out of Garret’s hands, see what I did there?  So at 1920 x 1080, we see an Average FPS of 84.3, 12.19% lower than that of 1680 x 1050 performance, and at a 1.63% decrease in power consumption oddly enough and the temperature dropped by 4 degrees.  At 50.5 FPS on 2560 x 1440, it was almost at the magic number… what can we do to make that score improve?  I will get to that in a little bit, but let’s make a pit stop over to New York in “Tom Clancy’s” “The Division”.

Here are the newer presets

The only thing I changed from the Medium default was to disable V-sync, just in case it does pass 60 frames per second.  So let’s see the results.

Well, it doesn’t look like this card is cut for The Division, though I had high hopes for it.  At 1080P, we can see the average FPS is 34.3, a 12.72% decrease from the FPS on 1680 x 1050 which was 39.3 FPS.  1080P did come in 1 degree cooler and consumed 5 Watts less on average.  At 2560 x 1440 though, it was very choppy coming in at 21 FPS, that’s a 38.76% decrease from 1080P’s and came in consuming 10 Watts more as well.

I will work on getting this to run better with those settings I have listed above, but let’s see how the card runs on a few other popular games.  Those games are League of Legends, Quake Champions, Counter Strike: Global Offensive and Paragon.  Check out the next page.

Iggy Castillo
I have spent many years in the PC boutique name space as Product Development Engineer for Alienware and later Dell through Alienware's acquisition and finally Velocity Micro. During these years I spent my time developing new configurations, products and technologies with companies such as AMD, Asus, Intel, Microsoft, NVIDIA and more. The Arts, Gaming, New & Old technologies drive my interests and passion. Now as my day job, I am an IT Manager but doing reviews on my time and my dime.
Iggy Castillo