Benchmarks, Performance, Temperatures and Power consumption
Before we get into the benchmarks, you have to know what I am running maybe if anything to help you compare with your own machine.
- Enermax Equilence chassis: http://azon.ly/2r4s
- Intel Core i7 7700K Processor: http://prourls.co/8vdm
- EVGA Z270 FTW K Motherboard: http://azon.ly/3nbv
- Arctic Freezer 33 eSports One: http://azon.ly/tblg
- Thermal Grizzly Kryonaut Thermal Paste: http://azon.ly/aokd
- Patriot Viper 32Gig 2800MHz DDR4 PVE432G280C6KGY: http://prourls.co/57lc
- Sapphire NITRO+ Radeon RX Vega64 8GB HBM2 Limited Edition: http://prourls.co/ayeq
- Western Digital WD Blue 500GB: http://prourls.co/57lc
- Patriot Ignite 480GB SSD: http://prourls.co/57lc
- Kingston HyperX 240GB SSD: http://prourls.co/254o
- Samsung 850 EVO 500GB SSD: http://prourls.co/1k1s
- WD Black 512GB M.2 2280 PCIe NVMe SSD WDS512G1X0C: http://prourls.co/7ue2
- Cooler Master Silent Pro Gold 1200W Power Supply: http://azon.ly/Umwm
- Creative Sound Blaster X Pro-Gaming H7 Tournament Edition Headphones: http://azon.ly/g201
Here are the clocks and specifications of the Sapphire NITRO+ Radeon RX Vega64 8GB HBM2 Limited Edition as reported by TechPowerUp’s GPU-Z.
The driver I used in this review for testing was AMD’s Adrenalin 18.2.3, a Beta driver. I usually go with the latest and greatest WHQL or Beta driver.
For benchmarking, we will of course see the performance the video card provides, but we will also check the temperatures as well as the power consumption. In the tests I report on minimum, average and max power usage using the “Kill A Watt” by “P3 International”.
I have changed up some of the games and programs I have used with more demanding titles, here is the list of games and programs used for benchmarking.
- FutureMark’s 3DMark Fire Strike
- FutureMark’s TimeSpy
- Metro Last Light
- Rise of the Tomb Raider
- Tom Clancy’s The Division
- Tom Clancy’s Ghost Recon Wildlands
This scored very nicely, and over all scored 93% better than all other results.
Breaking it down a bit, we can see the complete score was 18,049 while the Graphics score was 22,779, Graphics test 1 was 111.17 FPS, and test 2 was 89.30 FPS. The overall Physics score was 14,035 and the test was 44.56 FPS. Finally, the combined score was 8,482 with a combined test score of 39.45.
GPU-Z reports 2 separate temperatures for the GPU, namely GPU Temperature and GPU Temperature (Hot Spot), the hot spot is the hottest of the 2, so I will report its findings as it is a sensor inside of the GPU Silicon. During these tests, the hottest the GPU pumped itself up to was 70°C with GPU itself drawing 264Watts while the entire system drew a max of 428Watts an average of 348Watts and at its lowest with 3DMark running 69.5Watts.
Coming over to DX12, still on 3DMark, we will check out the performance on Time Spy.
The results here, 80% better than all other results, that’s not bad at all. Now don’t get dismayed, the results captured on 3DMark and Time Spy do not take into account if the system and/or GPU’s are overclocked, so only take it as a reference point and maybe throw in some of your own magic into it. I will do a little of that for you, so stay tuned and I will also provide the overclocks I applied and how it performed as well as some comparisons.
The Time Spy score was 6,868 while Graphics score was 7,184, Graphics test 1 was 49.72 FPS and test 2 was 39.18 FPS. The CPU score was 5,498 and the CPU test was 18.47 FPS. On Time Spy, we hit a max of 69°C with the GPU itself taking up 266Watts far less than 3DMark consumed. Here the total system consumed a max of 407Watts, on average ate up 383Watts and the minimum she took was 97.6Watts.
Getting away from synthetic benchmarks, which are still good to gauge performance we are going to check out how this card performs in games. I changed up the line up a bit this time, to have a few more current games, but still some of the older games that tax cards. Check out my list.
Through these benchmarks, I keep the settings the same, changing only the resolutions. Here are the presets I used.
Here are the results from all 3 presets below at 1920 x 1080, 2560 x 1440 and 3840 x 2160.
At 3840 x 2160, Metro did its duty, even with SSAA disabled and brought down the performance of the card. On average, the card provided 32.6 Frames per second while consuming an average of 162 Watts heating the card up to 71°C. From 4K to 1440P, there was a frame rate increase of 58.81% at 59.76 FPS, very playable though the temps did bump up slightly 6.8% though the average power consumption dropped 24.22% to 127 Watts. The trend continuing, from 1440P to 1080P there was an increase of 29.32% in frames per second, a 1.32% decrease in temperature and another decrease in average power consumption of 2.39% at 124 Watts on average.
So far, this is a pretty decent introduction into the card, the numbers look manageable. 4K might be able to be saved turning off some of that eye candy, but you don’t lose too much at 1440P and you get to keep all of the candy, so for me 2K is where it’s at so far. We need to make sure these results are not lost, but if they were, Laura would come in and find them for us. With that, let’s see what kind of performance we get out of Rise of the Tomb Raider.
Here are the settings I used, again only changing resolutions keeping the presets at Very High.
This one a little more difficult to provide results without drawing it out some, so I will only work on the “Overall Score”, since it encapsulates all results. I do provide the other results so that you can compare with your own.
At 3840 x 2160 we receive 47.65 frames per second drawing only 387 Watts on average heating up to 71°C. At 2560 x 1440, our performance increases by 60.77% at 89.25 FPS and with that the average draw is 393 Watts, only a 1.54% increase and thankfully the temperature dropped by 1°C. 1080P again increases performance by 36.26% coming in at 128.78 frames per second at 75°C at 404 Watts on average, an increase of 2.76% over the previous wattage results.
As of right now, it looks like 4K with the eye candy is out of the picture, but maybe results later in this review will change things.
Laura has gotten deep in the trenches recovering the results but might need some help getting out. I think Tom Clancy’s crew over at Ghost Recon Wildlands might be able to help out.
This one I do have to apologize on, now after performing all of the tests, I come to find I had benchmarked everything at high in Tom Clancy’s Ghost Recon Wildlands and being that I unfortunately have to return the card, I cannot retest everything. I won’t let the cat out of the bag just yet, but there has been a lot of testing in this review. If things change on that mater of having to return the card, I will update all results.
Keeping with those settings, here are the results.
At 3840 x 2160, it seems to not pop up here to that magical 60FPS, and this is not even on very high. Here we can see it hit a peak of 79°C at an average of 406 Watts. At 2560 x 1440, it performed 52.41% better scoring 78.31 frames per second at 75°C with an average wattage of 417, that’s 2.67% higher than at 4K. At 1080P, it hit 101.23 FPS at a temperature of 75°C with an average power draw of 397Watts, 4.91% lower than at 2K and 2.24% lower than at 4K UHD.
Up next, all hell has broken lose in New York after Laura returns the relic and it’s up to the team on The Division to bring some semblance of peace back to the city.
Here are the settings at which on which I benchmarked.
The story here seems to hold true with the other results, 4K seems to be a little too much. Here at 4K, we can see the average FPS was 47.3 with the temperature peaking at 76°C and the average wattage at 408. At 2560 x 1440, we can see the average was 56.12% higher at 84.8 FPS with the average wattage and temperature dropping 1°/Watt. At 1920 x 1080, the performance increased by 29.73% at 113.6 frames per second while the temperature rose another degree and the average wattage dropped 2 Watts.
With all these tests performed, 4K still could not keep up and 2K was just above the sweet spot but 4K is not looking that great on the RX Vega right now, would you consider that a bad thing?
All of these results might not give you much to go off of, unless you had this card already but I want to give you a treat, well really 2 treats, but before I go there, I want to show you a little gameplay.