Starting off, under the hood of the processor, we have an upgraded CPU. The Pi 4 has a Cortex-A72 clocked at 1.5 GHz out of the box, whilst the Pi 5 has an A76 at 2.4 GHz stock. That is blazingly fast. The Pi 5 also has a VideoCore 7, the generation up from the 6 in the Pi 4, with the GPU in the Pi 5 being clocked at 1 GHz, up from the 600 MHz of the Pi 4. Both Pis come in the same size memory models, 1, 2, 4, and 8 gigabytes, but the Pi 5 rocks LPDDR4X memory. This should be a bit more efficient and a little bit faster on paper than the LPDDR4 memory of the Pi, or so we thought. More on that later.
In terms of display outputs, both Pis have two micro HDMIs. I think that format is here to stay, but the Pi 5 has an improved HVS and display pipeline, meaning that it now supports dual 4K60 display outputs. The Pi 4 could only do that at 30 FPS. Both rock the same wireless and ethernet. We tested these and found that they were pretty much the same performance-wise, and they also have Bluetooth 5.0 and Bluetooth Low Energy. They also both have the same USB interfaces, but the Pi 5 now supports simultaneous gigabit operations on both USB 3.0s, and this is thanks to the new RP1 I-O controller, which is a nice little piece of custom silicon on the Pi 5.
That RP1 also enables the new four-lane MIPI connectors. Whilst the Pi 4 had one DSI for displays, one CSI for cameras, the Pi 5 has two DSI slash CSI connectors, meaning two displays or two cameras. Both still take a micro SD, but the RP1 on the 5 brings much greater performance. More on that later. And the RP1 chip also enables a new feature that is not present on the 4, a single-lane PCIe 2.0 interface. The Pi 5 is a bit more power-hungry as well. More on this later as well. Both take five volts, but the recommended Pi 4 power supply supplies three amps, while the Pi 5 chews through five amps. Oh, that's what the power supply supplies, not what it actually chews through in day-to-day use.
And there are many features that the Pi 5 has that the 4 doesn't, including a dedicated UART, a real-time clock with RTC battery, a dedicated fan header and heatsink mounting points, and also a power button. Just a very nice little addition. We're very happy about that. Ah, but what is a spec sheet without actual performance benchmarks? You can make anything look good on paper if you try hard enough. Well, we've briefly touched on these benchmarks in the review video of the 5. Check it out if you haven't. But the Pi 5 not only beats the 4 on paper, but in every single benchmark we performed as well, and by quite a significant factor at that.
Okay, first some ground rules. All benchmarks were performed running an early alpha build of Bookworm OS. That's the new OS coming out for the Pis. The Pi 4 was cooled with an appropriate heatsink and fan. We don't have it on this one. With the Pi 5 being cooled by the official active cooler, both of which resulted in no thermal throttling in our air-conditioned office. And of course, both Pis were powered by their respective official power supplies.
So starting off with the CPU, the Pi 5 came out on top of every benchmark, but by how much depended on whether we were doing a more synthetic test or a real-world test. We started with everyone's favorite, Sysbench, which isA prime number calculator. In both single-threaded and multi-threaded tests, the Pi 5 came out about 50% quicker. If these numbers on our graph seem a little higher than your Pi 4 benchmarks, that's because Sysbench performs much better on a 64-bit OS.
We then booted up StressNG. It is a more informal benchmark, but it was one of the few ones we had no issues on Bookworm. We had a lot of roadblocks we hit trying to set up some of these benchmarks. StressNG stresses the CPU, but counts how many operations are performed over the stressing time. We ran 60-second tests and saw that the Pi 5 scored about 75 to 80% higher on single and multi-threaded tests. And yes, BOGO ops per second is the official unit of this test. I am not making that up.
For our more real-world benchmarks, we saw an even greater increase in performance. We used a timed BZIP2 test, which compresses a non-repetitive random binary data file. We compressed 100 megabytes and found that the Pi 5 was able to complete this task in under half the time as the Pi 4. In a similar result, we benchmarked GIMP, the photo editing suite through Pheronix test suite. It performs some common operations on a sample image. And again, the Pi 5 was able to do these in under half the time, about 2.3 times quicker.
While technically not solely a CPU-based benchmark, we also gave Speedometer 2.1 a whirl. It tries to simulate real-world web browsing application scenarios and measures how quickly your browser can perform certain tasks. In this benchmark, the Pi 5 smashed the Pi 4 with it being able to perform these tasks just over three times faster. So all in all, that new Cortex-A76 brings a huge leap in computing performance, with less on synthetic benchmarks and more on real-world applications. You can make that as you want.
GPU performance was about the same story again with the new VideoCore 7 coming in at about double the performance of the previous Pi's VideoCore 6. We ran a default GLMark2 benchmark at 1080, which renders various scenes to test texture handling, shading, lighting, et cetera, some very practical graphical tests. We did encounter some bugs and weird issues with this one, but Veronik's test sweep came to the rescue again, and the Pi 5 scored more than double the Pi 4. Running a time demo in Open Arena, the open-source Quake 3 clone, we saw that the Pi 5 again came out much higher, with it running at triple the FPS of the Pi 4. This was at default settings, 720p. There may be some other software and OS-related side things besides the GPU giving this three times number, but both of these tests indicate that the Pi 5 is much quicker in the graphical department.
In our memory tests, we got some astounding results that are a bit hard to believe. We ran IntMark on RAM speed to measure the maximum possible cache and memory performance while reading and writing one-megabyte blocks of data. Whilst the Pi 4 was able to perform this at about 4,000 to 6,000 megabytes per second on read and write operations, the Pi 5 came in at nearly 30,000 megabytes a second. That is about a five-fold increase, and we have no idea why. These are early days for this board, whether it is something related to Bookworm or whether they are doing some big brain stuff hardware-wise on here, we don't know. It's not that the Pi 4 is running slow on Bookworm. We got the same speeds on Bullseye, but the Pi 5 just seems to stomp here, and it was a similar result across different block sizes.
And finally, rounding off our benchmarks, we tested microSD speeds. One of the biggest additions to the Pi 5 was the new RP1 chip to handle IO capabilities, and the effectiveness of this new chip is very apparent, with the Pi 5 being capable of double the sequential read and write speeds of the Pi 4. We attained these results with HD Palm for the read and DD for the write with 50,008-kilobyte blocks. We also booted up IOZONE and found that, while 4K write speeds were about the same across both, there was about a 50, 60% increase in 4K random read speeds.
We also did some testing on our Audi Arc and saw that the Pi 4 draws about 275 milliamps on idle, and the Pi 5, about 500. This is with no monitor or peripherals plugged in. We then plugged in a monitor, keyboard, and mouse, booted up StressNG, and we saw that the Pi 4 consumed about 1.3 amps under load, and the Pi 5, 2.15 amps. That is a bit more power hungry than the 4. We also analyzed boot times and saw that the Pi 5 booted in just under 20 seconds, considerably quicker than the 35 seconds of the 4. So I think the Pi 5 is leaps and bounds quicker than the Pi 4. I think at least a doubling of performance across the board is a good ballpark figure.
So what does this mean in terms of experience? Well, over the last couple of days we've been using the 5, it has been great. And I think while the Pi 4 can be used as a day-to-day computer for light activities, the Pi 5 can do this without being frustrating. The Pi 5 is just way more responsive. Just opening things and manipulating windows in the desktop environment was smooth and snappy on the Pi 5, and a bit clunky on the 4. There was a lot less waiting for things to load, less browser crashes, all the little infuriating things. We had only some minor frame drops here and there while watching 1080p 60 FPS YouTube, whilst the Pi 4 struggled with it.
In writing this video, circumstances meant that I had to get out my five-year-old laptop from uni, and I would actually try and do as many tasks as possible on the Pi 5 because it was just quicker. My laptop probably does need a repasting though, but you get the point. I'm excited to see how this power will be used. I think RetroPie is going to be a great experience when it supports the Pi 5. The new features are also amazing. I think one of the first things I'm gonna be doing is replacing our media PC at home with a Pi 5 strapped with an M.2 through that new PCI interface, as I think it's now fast enough to be a replacement for my media PC needs.
And in our brief experience with the Pi, I think that's the biggest thing that jumps out, that it is a very viable daily driver for light loads and general web browsing, something that I wouldn't consider the Pi 4 to be. The Pi 4 is a few bucks cheaper, consumes less power, and has that headphone jack, but the trade-off of these features for a huge leap in processing power and a wealth of new features, that's a value proposition that is really hard to beat.
Till next time. Take care. I'll see you later.
Makers love reviews as much as you do, please follow this link to review the products you have purchased.