Nvidia's Trying To Make Benchmarking Easier [Update: It's Out!]

free benchmarking tool nvidia amd

Anything that simplifies the arcane process of benchmarking and PC analytics is, generally, a good thing. So it's nice to know that in the near future, everyone will be able to download a simplified benchmarking tool from Nvidia that works on AMD and Nvidia graphics cards.

Update 10/07/19: FrameView has now been released to the public, and you can download it for free here without any logins or middleware software. Some shots of the program's interface, and how it logs data, can be seen below.

Called FrameView, the tool is basically a piece of software that Nvidia described as a software version of their frametime capture tool, FCAT.

Now there's plenty of software benchmarking tools out there. FRAPS has been a standard for a long time, and there's been lots of alternatives introduced to accommodate DirectX 12 and Vulkan games, like PresentMon or AMD's OCAT (Open Capture and Analytics), the latter of which was updated only just last week.

Not all of these tools have great interfaces or work with the same amount of reliability, and in some cases the information spat out can be indecipherable to novices. FrameView offers a pretty straightforward view: it'll display an overlay with the average frame rate, as well as frame rates for the 90th percentile, 95th percentile and 99th percentile, as well as power used per watt and the amount of dropped frames.

nvidia FrameView

The big advantage here is that FrameView has integrated power management, something FRAPS/OCAT and other software tools don't do. All of this can be logged in real-time (although the beta software won't display the overlay while logging), and Nvidia stresses that there's no noticeable performance impact or difference by using FrameView compared to the supremely expensive, hardware-based FCAP solution.

One kicker is that FrameView won't be able to report the amount of power used by AMD boards, because of differences in how those APIs work. But all other functions, including logging for rendered and displayed FPS, should work just fine.

FrameView works with DX12, Vulkan, DX11 and older APIs, although the pre-logging overlay currently only functions in DX11 and DX12 games. Jeff Yen, Nvidia's technical marketing director, said that Nvidia is working on making the overlays visible during the logging process as well.

A beta version of FrameView has been released to the press, with Nvidia still determining how they'll distribute the program more broadly. I asked whether it would be released on Steam (like other benchmarking/overclocking tools have been), although at this stage it's likely that it'll be a separate download on the GeForce website.


Comments

    Not to mention that FRAPS hasn't been updated since February 2013!

    This looks really nice, and it's great to see 95th and 99th percentile built in. What's the support like for frametime tracking?

    I definitely trust it to be objective. Nvidia would never do anything suspect.
    Now can we have intel develop a CPU benchmark, that would be great.

      Came to say the same thing. I may be paranoid but I'd never trust a manufacturers own bench mark.

      I wish I could upvote that comment more.

      If it was a tool developed in cooperation with AMD then maybe I'd trust it. But as it stands, not so much.

      I expect someone will check out how legit it is. Toms Hardware is usually all over that sort of thing so if theres anything dodgy going on, you'll hear about it. And I expect Kotaku to be one of those reporting it. Until then, its another tool to use.

      Really though, when you're talking about the differences the "optimising" makes, its so small the average user isnt going to see it anyway. Most people vote with their wallets anyway.

        Yeah i hope some good analysis is done on it.
        But as you say even then it will just be a generally handy tool. So for my own use, i am happy about that.

        Well i mean, with the differences you can see in games depending on how much work they do to optimise for each vendor, as well as the vendors drivers doing the same. I could certainly see it pretty easily having a big difference between AMD and Nvidia cards using this tool depending on how its coded.
        The 'voting with wallet' issue comes in if a bunch of people use this tool to post benchmarks online and its unfairly favouring Nvidia making them seem better than reality. Most high quality benchmarkers would surely not be so absentminded not to keep an eye on that but there are enough randoms posting becnhmarks and such on youtube and everything that it has the potential to mislead people.

    Yeah.
    Really i'd prefer some 3rd party to work with them both to make a good becnhmark suite.

    Though i have to say i wrote out my post before i fully read this article. This software is really just a monitoring software to check performance stats, not really a benchmark.
    Still could make comparisons more difficult if this has less overhead than other tools, because AMD will still have to use the other tools which may be more impactful on performance.

      True, but it's software to help people benchmark their systems against whatever games they'd like to use; not all games have inbuilt canned tests. But anything FrameView can hook into, you can also use FRAPS/OCAT/PrecisionX/other software, so it's not a complicated matter to gauge if there's more or less overhead with FrameView.

        True, but for noobs or people not too knowledgable, they might see comparisons just on this tool across AMD/Nvidia (or AMD card using a different tool) and not know that there could be a bit of a margin introduced by the capture technique.
        I mean that is literally the only reason Nvidia would spend time making this, they hope in some way it will get more people to buy Nvidia cards.

        But yeah as an individual i will say this is a helpful tool ill probably use sometimes. i'll probably play around seeing how OCs change performance in different games and such stuff (in an easier to check way).

        What will be interesting will be comparison articles that check whether it's doing anything underhanded in benchmarking cards between the two companies. Some side by side comparisons with FCat or similar will be interesting.

          Don't expect many FCAT comparisons (it's a hardware based rig that costs thousands to put together, so few outlets use it). More software based comparisons would be likely, but nobody's going to do it after all the Super/Navi/AMD hardware arrived within days of each other.

            Don't need a heap of comparisons from different sites. Once Anandtech or Toms Hardware does it that'll do the trick.

Join the discussion!

Trending Stories Right Now