 Tom here from LARN Systems, and we want to definitively answer the question right here in February of 2023 of which is faster, TrueNAS Core 13 U3.1 or TrueNAS Scale 22.12.0. And this is February 15th, and I know a new version is coming out soon for scale, but we tested what was available and hey, we'll run these tests again and update you if there's some big differences when the new version does release. But to answer the question, it depends. And I know that's not what anyone wanted to hear, but it did turn out that, well, depending on your workload and depending on whether you're using 4K writes, 4K reads, or whichever other variations of one meg block, two meg block, or a lot of other factors, there's not a clear winner. Each one had its strengths and each one had its weaknesses. All that's linked down in the forum post down below, but let's go ahead and take a look at that forum post and talk about where the winner and loser is depending on your workload. Now, before we get to the test results, let's talk about the testing process, which I outlined right here at the top of the forum post. That is also linked down below. Basically we loaded TrueNAS Core, set up the shares, connected it to an XCPNG host, then ran the series of benchmarks from a VM that was running Pheronics. We also created an SMB share and on the same network we had a Windows server that we connected to that share and we used CrystalDiskMark to create the benchmark. Now, all this was done with TrueNAS Core 13 U3.1. Then we did an in-place upgrade, which means we didn't have to reload the VM or recreate any shares. All the settings copied over that way. We weren't at risk of making any mistakes that would skew the benchmarks and just went through the same process of running the numbers again. That way we could eliminate as many variables as possible to make the test as fair as possible to really trying to dive into both of these systems. Now let's talk about the results, which by the way, there's also a link to the Pheronics results down at the bottom of this forum post, but this is where the question is not as easy to answer. It's not black and white because there was a lot of variants based on what the block size was you were testing with and whether it was a read or a write action, whether it was a sequential or random action. Those all produced a wild variety of numbers with no clear winner. For example, let's jump over here and say, if we have a random read block size of one meg, we would say that TrueNAS scale 22.12 is a winner. It's not a lot faster, but it's somewhat faster, but IceGuzzy clearly doesn't perform well at all for both random read block size of one meg under TrueNAS 13, but it does perform pretty good here and pretty evenly across these and maybe still a little faster than NFS, but let's jump down to one that's even more extreme. If you're random write block size of four kilobytes, you're gonna wanna have a system running TrueNAS 13 with IceGuzzy because it can score substantially higher, double the NFS and well, just about double NFS on each of these. It's a pretty big difference there for that particular workload, but one result that definitely puzzled me is if your random read block size is 128K, TrueNAS scale just performs substantially better. We've got about double the performance here for NFS versus NFS on 13. So once again, it's not a clear winner. It's a, if you're using this, this will perform better. Now let's talk about SMB performance. We actually have two different benchmarks we did, one with one gig, one with 16 meg. I wanted to test larger writes or smaller writes and reads and figure out which one was the best and it turns out that, well, they're pretty much the same in some categories. So 161 read, 161 read here for the sequential, but when you get down to the random, there's a pretty big difference there. It's about half the performance at the small random 4K mark here. So you're losing a pretty substantial amount of performance with TrueNAS scale if you need that particular write. On the other side, if you're looking at the write speed, the write speed was consistently faster over on TrueNAS scale. We thought that was interesting. Even across different block sizes, you're still able to sustain higher writes. Now we got 330 here versus 388, but only 289 here and 399 here. So once again, some differences. We ran these a few times and results kept coming up within a few percent of each other and this is a repeatable thing we could do with this system. So hopefully this helps you make a decision between TrueNAS core and TrueNAS scale in terms of using it as a NAS, but when it comes to applications, the clear winner is gonna be TrueNAS scale. There's a broader array of applications available with the third parties that you can add to it as well and Docker things and all the tinkering. And I know it's still relatively new. I've done a recent video and talking about how to map host path storage into some of the applications. It's a new product, but there is still a lot of support and going for going to be more support for those applications. So if applications are your deciding factor, not down to some nuance of how fast can the ice-gussy target perform or how fast of an NFS target can perform or your SMB, if the small rates matter to you, usually the apps is gonna be the winning and deciding factor and scale is the clear winner when it comes to that. So leave your thoughts and comments down below or how do my forums for a more in-depth discussion or even go visit the TrueNAS forums to engage with the team and members over there and see all the opportunities and fun things that are going on with the TrueNAS and the whole ecosystem over there. Thanks.