Let’s find out how the new Intel 10th gen 8 core i7-10875H compares against the older 9th gen 8 core i9-9880H and see what the performance differences are, they aren’t quite what I was expecting.
These are the key spec differences between these processors, they’re specced quite similarly despite the i7 and i9 branding. Both are 8 core 16 thread parts, they’ve got the same amount of cache, and same base clock speed. The single core turbo boost speed of the 10th gen chip is a bit higher, and the 10th gen also makes use of faster DDR4-2933 memory.
The laptops I’m testing with are the MSI P65 Creator for the i9-9880H and the Gigabyte Aero 17 for the i7-10875H. There’s a bit of a size difference, so the Aero may run a bit cooler. I’ve also limited both to a 45 watt TDP in Intel XTU, the stock power limit for each of these CPUs.
Starting off with Cinebench R20, I’ve got the older 9th gen 9880H up the top, and newer 10th gen 10875H down below. The 10th gen chip had a 7.5% higher single core score, which makes sense as it can clock higher there, however I didn’t expect the multicore score to be lower. It’s not much of a difference though, the 9th gen processor is just scoring 1% faster than the newer 10th gen, so kind of margin of error stuff and effectively the same. I’ve also got some older Cinebench R15 results for those that still use it, the single core score from the 10875H had a similar 7.5% higher score, but again the multicore score from the older 9th gen processor was scoring 2.7% higher this time. Handbrake was used to convert a 4K video file to 1080p, then a separate 1080p file to 720p. In this test, the 9880H was a little ahead in both instances, just 1% faster in the 4K test, or a 2.5% boost in the 1080p test. This is a multicore test and it seems to match what we saw in the Cinebench multicore results. Blender is another core heavy workload, and the results were very close together here. The 9880H was just a few seconds faster in the BMW test, then the 10875H gets its first multicore win in the classroom test, though again it’s only a few seconds difference, at the end of the day these are performing extremely comparably. V-Ray is another multicore test, this time the 9th gen processor was in front once more with just a 2.5% lead in this test, so a similar margin to many of the other multicore results covered so far. The Corona benchmark also uses the processor to render out a scene, this time the 10th gen 10875H was completing the task a few seconds faster, giving it a 3.5% performance lead over the 9880H. 7-Zip was used to test compression and decompression speeds, and the newer 10th gen chip was faster in both instances here, with a 4.4% faster decompression speed and 2.6% faster compression speed, so a slight edge there. Veracrypt was used to test AES encryption and decryption speeds, for some reason the 9880H was ahead in both of these tasks, granted I’ve found the Veracrypt benchmark a bit inconsistent, even when I take the average from 10 runs like I’ve done here. Geekbench 5 was hitting a 4.5% higher single core speed with the 10th gen chip, which makes sense given the higher single core turbo boost speed, however for some reason the 9th gen chip was scoring 10% higher in multicore, the largest gain seen out of all tests with the 9880H. I’ve used Adobe Premiere to export 4K video, and the test utilizes quicksync which both systems are able to take advantage of. Like many of the other multicore tests, I was seeing better performance from the i9-9880H which was completing this task around half a minute faster. I’ve also used the Puget Systems Premiere benchmark, in their test though the 10875H was scoring 0.6% faster, so not much of a difference, but at least it’s ahead unlike in the previous test where I actually exported one of my own projects. Adobe Warp stabilizer uses a single core to smooth out a clip, so as a less core heavy workload the 10th gen i7 was completing the task a few seconds faster, giving it a 1.3% lead in this test. On the other hand, the Photoshop benchmark from Puget Systems was scoring higher with the older 9th gen i9 processor. I had the discrete Nvidia GPU disabled for this test so we should only be looking at CPU differences.
These are the differences between the Intel 10th gen i7-10875H and the Intel i9-9880H when both are capped to the stock 45 watt power limit. I thought the results were quite interesting. As expected, all of the single core results were ahead on the 10th gen processor which makes sense as it has a higher single core turbo boost speed. I didn’t expect the 9880H to come out ahead in most of the multicore tests though. Although the 10875H has a higher all core turbo boost speed, with a 45 watt power limit it’s unlikely these speeds are actually being hit anyway.
Once you average things out there’s not too much difference at all, but still I was expecting the 10875H to come out ahead more than this.
When we look at the total system power draw from the wall while running the Blender benchmark, the 9880H was using about 5 watts more than the 10875H, so it would appear that the 10th gen system may be a little more efficient. Keep in mind that this multicore heavy test was performing almost the same on either CPU, however there would also be some differences due to the laptops themselves being different too.
When we look at the clock speed differences in this same test they’re very close together, however the 10th gen chip was just slightly faster despite also using less power. I’ve added the temperature differences because people would riot in the streets if I didn’t, however these are far less comparable as the 9880H was in a smaller 15” chassis while the 10875H was in a larger 17” one and it also ran the fans much louder as well, so yeah not too fair of a comparison here.
Unfortunately there are no gaming results at this time as the MSI P65 had a 2070 Max-Q while the Aero 17 had a 2070 Super Max-Q, so I wasn’t able to fairly compare games. No undervolting was done in the previous tests, many 10th gen laptops seem to have this locked, and this was the case with the Aero 17. I’m guessing this has been done to mitigate the plundervolt vulnerability, so technically we could also squeeze out some extra performance from the 9880H by undervolting it.
Some 10th gen laptops do support undervolting like the GS66, but yeah if you plan on doing that with 10th gen just do some research in advance to see if it’s possible.
Prices are a little hard to compare, both because there aren’t a wide selection of 9880H laptops, and because 10875H laptops are still launching. The P65 that I’ve tested here with same GPU and 32gb of memory and 1TB SSD goes for $2700 USD. The new MSI Creator 17 with 10875H with same memory and SSD but with newer 2070 Super graphics is about $300 more.
It’s hard to compare fairly as 10th gen also launched with the new Super GPUs, you could of course get the same MSI laptop with 10875H for much cheaper though, but in general I’m not expecting the prices to be that much different. Even with 10th gen, 8 core Intel based laptops still seem to come at a price premium, something that I hope will change in future as AMD’s Ryzen 4000 starts applying pressure.
When we look at the prices Intel are selling these to their customers for, the 10875H is apparently cheaper. Given the i9 is harder to come by, the 10875H is probably going to be the way to go simply due to availability. The key take away was that there really isn’t that much of a difference going from the 9th gen 8 core to the 10th gen 8 core. Makes you wonder if 10th gen is perhaps suffering from security mitigations or something, that might help explain some of the lower multicore performance, otherwise I’m not too sure given they have the same core and thread count, same base clock speed, same cache and same power limit. I don’t seem to be alone in these results either, Tim from Hardware Unboxed showed a similar difference between these two chips.
Anyway let me know what you thought about the performance differences between the last gen i9-9880H and new 10th gen i7-10875H processors down in the comments, which would you get and why?