Geek Benchmark Download
Karla Cassone
I'm on iOS 14.5.1 with my iPhone 11 Pro Max that has 87% battery capacity. I've noticed that when I do a Geekbench 5 benchmark test while connected to the charger, it gives me around 1340 for single core score and 3270 for multi-core score.
Now, if I do the same exact benchmark test on the phone without having it connected to a charger, it gives me around 760 single core score and 1250 for multi-core score. It's the weirdest thing but I think it has something to do with the battery, maybe?
And big Apple YouTuber Nick Ackerman also has the issue caused by 14.5.1 where his iPhone XR is outperforming both his 11 Pro and 12 Pro Max in every benchmark test, getting exactly the same iPhone 12 scores as me which are half the Geekbench average for the iPhone 12. It seems to be only affecting the iPhone 11 and 12, older iPhones are not affected:
For our AYANEO Geek review we will be comparing the difference between it and the AYANEO 2 handheld gaming PC. We will be performing some system and game benchmarks then comparing the results. Lastly, we finish by playing some games recommended by our viewers and then our final thoughts if you should buy the Geek instead of the AYANEO 2.
In our tests whilst running the Cinebench benchmark we got a highest temperature of 45C, and highest fan noise of 69dB. With the 50.25Wh battery, we got exactly one hour at full load at 28W TDP, and four hours thirty-five minutes idle on the desktop at 28W TDP.
As part of our AYANEO Geek review we perform both system and gaming benchmarks. We start our benchmarks with some system tests to give us a look at the overall performance which we can compare with other 6800U based handheld gaming PC at 28W TDP.
We continue our AYANEO Geek review with some gaming benchmarks. For the gaming benchmarks we will be performing them at the highest resolution of 1200P as well as 800P, across a mix of 11W, 20W, and 28W TDP. This gives us a good range of resolutions and TDP levels for gaming on for best quality through to battery life.
Full disclosure here: We are comparing the benchmarks with the AYANEO 2 gaming handheld which were reviewed in early November 2022. There are some differences in the results depending on the test due to different drivers as well as game versions. So these benchmark results are only for reference and should not be taken as absolute.
We get decent scores across the four benchmarks, and a frame or two above the AYANEO 2 on some. You can play at 1200P and limit to 30 or 40 frames, or go 800P and increase a couple of graphics settings.
We review a lot of hardware at Ars, and part of that review process involves running benchmark apps. The exact apps we use may change over time and based on what we're trying to measure, but the purpose is the same: to compare the relative performance of two or more things and to make sure that products perform as well in real life as they do on paper.
One app that has been a consistent part of our test suite for over a decade is Geekbench, a CPU and GPU compute benchmark that is releasing its sixth major version today. Partly because it's small, free, and easy to run; partly because developer Primate Labs maintains a gigantic searchable database spanning millions of test runs across millions of devices; and partly because it will run on just about anything under the sun, Geekbench has become one of the Internet's most-used (and most-argued-about) benchmarking tools.
"I'm really glad that people seem to have latched onto it," Primate Labs founder and Geekbench creator John Poole told Ars of Geekbench's popularity. "I know Gordon Ung at PCWorld basically calls Geekbench the official benchmark of Twitter arguments, which is the fallout from that."
Geekbench's cross-platform compatibility is part of its appeal, which has been baked into the benchmark since its earliest versions. It began at the height of the PowerPC Mac era when Apple's hardware was exotic and niche and apps that ran on Mac OS X were relatively rare.
"I just switched over to the Mac back in about 2002," Poole told Ars. "So I was getting used to that ecosystem. And then the [Power Mac] G5 came out and I thought, oh, this looks really cool. I went out, bought one of the new G5s, and it felt slower than my previous Mac. And I thought, well, this is really strange; what's going on. ... So, you know, I grabbed what [benchmarks] I could download and ran them and got really confused, because what the benchmarks were saying wasn't jiving with my experience.
"So I actually went and I reverse-engineered one of the popular benchmarks and found that the tests were, for lack of a better word, terrible," said Poole. "They weren't really testing anything substantial, you know, doing really simple arithmetic operations on really small amounts of data, not really testing anything. And so I thought, how hard can it be to write a benchmark? Maybe I should write my own."
The LearnGeek Innovation Cycle is an informal assessment. I blend my 20+ years of field experience with ongoing peer conversations, literature reviews and gut instinct. I'm not trying to predict the future with 100% accuracy. Instead, I hope to provide a simple, insightful, practical perspective on how L&D is evolving so organizations can benchmark their practices and make smart decisions as they prioritize limited resources.
Before performing any benchmark, ensure that nothing else is running on your computer. If an app is crunching away in the background, it will slow down the benchmark and skew the results. And plan to run your benchmarks when you won't need your PC for a while, since some of these tools can take a while to run their tests. We'll let you know about how long you can expect each tool to take.
Prime95 offers to perform a torture test right off the bat. The torture test is ideal for testing the stability and heat output of your CPU, and is particularly useful if you've overclocked it. If you want to perform a torture test, go ahead and click the "OK" button. Note that the torture test can take quite a while to run. If you just want to perform a benchmark instead, click the "Cancel" button.
When it's done testing, NovaBench displays an all-round NovaBench Score---where higher is better---and it also shows the results of each individual benchmark. Click the "Compare These Results Online" button to check out how your score stacks up against other computers on the NovaBench website.
After downloading and installing, go ahead and run 3DMark. On the home page, click the "Run" button to benchmark your PC. The benchmark you see will vary depending on the version of Windows---and DirectX---that you're running. For Windows 10 PCs, the default benchmark is "Time Spy."
And like with 3DMark, you can get each version of PCMark as a free, basic edition or paid, advanced edition ($29.99). The free version includes video playback, web browsing, image manipulation, and storage benchmarks, as well as some 3D graphics and gaming performance benchmarks. The paid version does add additional benchmarks and fancier result graphs.
After downloading and installing the edition you want, go ahead and run PCMark. We'll be using PCMark 10 here, but most of the options will be similar in other versions. On the "Home"page, click the "Run" button to start the benchmarking.
The benchmark can take a while to complete---nearly 15 minutes on our test system. PCMark shows you the progress of the tests at the bottom of your screen, and you'll see additional windows pop up as it tests video playback and graphics. When it's done, you'll see the results and, as usual, higher scores are better.
After downloading and installing Sandra, go ahead and run it. In the main window, switch over to the "Benchmarks" tab, and then double-click the "Overall Score" option. Alternatively, you can run benchmark tests against specific components.
The Overall Score benchmark includes benchmarks of your CPU, GPU, memory bandwidth, and file system performance. Make sure that the "Refresh the results by running all benchmarks" option is selected, and then click "OK" (the check mark button) to run the tests.
Sisoft also offers the ability to customize your ranking engines, which is free but requires you sign up via email. If you prefer not to do this, you can just hit the "Cancel" button to start the benchmarks.
Geekbench scores can be compared against others via the Geekbench Browser. On the benchmark result page, set a result as the baseline for comparison by pressing "Set Baseline" on the right-hand side of the page. Use the search bar to find a model that you would like to compare it to, then simply click on the device to view a side-by-side comparison of their scores. Pressing "Remove Baseline" on the right-hand side of the page will allow you to set a different baseline to make other comparisons.
Geekbench can only run benchmarks for the Compute APIs which are supported by your device, if any. If you believe that your device should support a Compute API and it does not appear in Geekbench, please check our website to confirm that that API is supported by your version of Geekbench. It may be necessary to update your GPU's drivers or to install the Compute API toolset on your device.
If you are using scripts to run the Vulkan Compute benchmark on an AMD system with no display, you may need to update your drivers to ones released in September 2019 or later so that Vulkan is available as an option.
Over the years, manufacturers have implemented various techniques to increase computer performance, like increasing the cores in a CPU and allowing multiple threads to run simultaneously on a single core. However, this means that statistics like gigahertz or core count are no longer a good way to compare the performance of two different laptops. If you want to know whether a laptop can process photo edits, run physics simulations, or compile code quickly enough to suit your needs, you can look to a Geekbench benchmark. These measurements are a good way to obtain comparable results among laptop models, helping you get a better idea of the kind of performance you can expect when running day-to-day tasks.
f5d0e4f075