tbh what resolution it's set at just to get a base score for comparing the diff shouldn't matter imo. it's not like you're actually playing anything to care about how it even looks while testing. just make sure the same settings on each test is all that really matters.
otherwise an in-game benchmark is your best bet.
Comment has been collapsed.
MSI Kombuster isn't really a benchmark, it's more of a burn test to check OC settings are stable.
Comment has been collapsed.
I use sites (where you can input what card you have and what game you want to run) to compare cards.
A lot depends on the CPU and RAM too.
http://gpuboss.com/compare-gpus
there might be an equivilent one of this http://www.notebookcheck.net/Computer-Games-on-Laptop-Graphics-Cards.13849.0.html for pcs
as you can see you can enter specs and it benchmarks for you
I also have a FPS counter in games and I tweak settings at first, but quality is subject to how you play (for example I prefer running on high but I don't like things like motion blur).
Comment has been collapsed.
No I mean my GPU is not worth benchmarking. I know most recent games run bad on it, I don't need a benchmark to see that.
Comment has been collapsed.
Yeah I know that people overclock their hardware extremely just to get a high score in benchmarks.
Comment has been collapsed.
I really like FFXIV Benchmarks. :) There has been 3 so far, you can get the latest one from the official website. It's surprisingly convenient for benchmarking as it shows the real-time graph of both your current performance as well as your Best / Previous / Specific score file.
Forgot to mention it's also very pretty, of course, has nice music and 5.1 sound.
Comment has been collapsed.
Glad you liked it. :) The ending summary only shows the average framerate, but it saves the entire framerate flow data for a particular test in one of the .dat files stored in the "data" folder (used to build the comparison graph). Sadly, I don't know of any custom result browsers able to display those graphs outside of the benchmark itself.
Comment has been collapsed.
The built-in benchmark of Arkham City to see how it runs when sabotaged by UE3, then the Resident Evil 6 benchmark tool (separate and free on Steam as well) to see how it runs on a semi-decently optimised game. If I had it, I probably used the latest Hitman as well. Or Doom 4(I have it thanks to a generous person, just haven't started it yet to see if they have a benchmark in the game or not).
I really don't like 3Dmark because it is an open secret that all video card manufacturers cheat on it by releasing drivers that run it better than a standard 3D application. This is also why cards are tested with real games and synthetics are there just for the pretty score.
Comment has been collapsed.
Valley is my choice for benchmark.
https://unigine.com/products/benchmarks/valley/
Comment has been collapsed.
furmark is great
http://www.geeks3d.com/20160826/furmark-1-18-0-gpu-burner-opengl-benchmark/
(the download is the big yellow box )
Comment has been collapsed.
Out of curiosity, what model is your RX 470? (As a point of reference, your benchmark scores almost reach the numbers for my R9 290 [Sapphire Tri-X], which I find interesting).
Comment has been collapsed.
Sure thing. Just wanted to know the model, thanks.
Comment has been collapsed.
Not when it is the last generation. Despite the numbering, the only difference between the R9 200 and 300 series was a slightly higher base clock; they were rebranded 200 cards in any other regards. The 400 series made quite a large jump.
Same with NVidia, the 1000 series almost doubled the performance of the 900 series. Granted, there they had it easier, as the old architecture was in production for waaaaaay too long.
Comment has been collapsed.
Going from 28nm to 14/16nm certainly helped, but I was a bit underwhelmed by the current crop from AMD, to be honest. Interested in seeing what they can pull off with Vega but I'm keeping my expectations in check.
Comment has been collapsed.
769 Comments - Last post 3 minutes ago by OwieczkaDollyv21
1,960 Comments - Last post 15 minutes ago by MeguminShiro
7 Comments - Last post 30 minutes ago by FEGuy
19 Comments - Last post 1 hour ago by Vincer
37 Comments - Last post 3 hours ago by GarlicToast
49 Comments - Last post 3 hours ago by Chris76de
11 Comments - Last post 4 hours ago by PoeticKatana
48 Comments - Last post 2 minutes ago by lext
559 Comments - Last post 3 minutes ago by Aldcoran
438 Comments - Last post 6 minutes ago by steveywonder75
245 Comments - Last post 8 minutes ago by imminiman
187 Comments - Last post 9 minutes ago by Kolisz
2,214 Comments - Last post 28 minutes ago by Mayanaise
172 Comments - Last post 29 minutes ago by genkicoll
I just bought an RX 470 to replace my HD 7850 on my main PC. The old card will join another HD 7850 in CrossFire in the HTPC. I'd like to benchmark both machines before and after the upgrades. Any recommendations? I prefer free tools or ones that come with games. I tried running the built in benchmarking tools of Metro 2033 and Crysis 1, but both crash for some reason or another. I'm using Windows 10 64 bit, so this might have something to do with it. I tried 3DMark Basic Edition, but the included free test (Time Spy) seems way too challenging to my HD 7850 to be of any value (compared to actual games). Any other suggestions welcome!
Thanks!
Update: Thanks for all the suggestions. Here are the test results with my main PC (Windows 10 64 bit, Intel i7 4770K, 16 GB RAM, 1920x1080 screen resolution)
Comment has been collapsed.