Will Engine Prime utilize the multiple threads in multicore processors like Serato does? For instance, if you have a dual-core processor, you can analyze (4) songs at a time…same w’ quad-cores being able to analyze (8) songs at a time. If so, this will indeed be a game changer over the competition’s offering.
I could be mistaken, but analyzing 4 tracks with a dual-core CPU will only work on selected Intel CPU’s, the ones that support hyper-threading (HT). On others each core will just do one task.
AMD CPUs don’t have hyper-threading technology I think, so the number of cores is the number of tasks. That said, AMD sells 6- and 8-core CPU’s.
that’s correct. serato (or other programs) will use twice the number of jobs compared to your core count only if your cpu has hyper threading. however, keep in mind that in many cases the bottleneck is the storage device. there’s no point in analyzing 8 flac files at once on an usb2 stick for example.
anyhow the most important factor is the accuracy. if you have a fast analysis but have to fix half of the tracks manually afterwards it’s useless.
Amen to that
I don’t work for Denon and can’t say for sure, but since they are re-writing this app from the ground up, implementing SMT(Simultaneous multithreading) should be really simple and therefore there isn’t a reason why this feature shouldn’t be present in the code.
My guess the reason for this feature not being put in vanilla(meaning plain or original version) Engine is that it would of require a major re-write of the product, therefore a major undertaking (like re-writing the app from the ground up X). Worst case, Engine Prime can read Serato meta data, so you could allays use Serato for scanning big chunks of files before bringing them into Engine. I doubt you will need to do this.
Also, in one interview there was a call out of a i5 proc being in the sc5000 decks (4 cores available) and most preview videos call out a quad core. There would be no reason for Denon to put in a quad core if their software did not support SMT. What is really exciting though is if there really is a i5 in these decks, that would mean the decks and Engine Prime would share the same back end code, meaning all features between Engine Prime and the decks should be in line and released at the same time and not require a huge effort. Basically updates will be fast and quick and the i5 will be able to offer a long long shelf life. Trust me when I say there would of been a lot of cheaper options out there then using that i5(if true) but this will pay off for everyone in the long run.
Once again, these are my opinions based off what I am reading and what I know of the IT/Dev Opps industry and I do not work for Denon. I do work in IT though and I have a really strong background HW and SW releases.
Hope this helps.
No point? Really? I’ve been doing it for nearly a decade. I’m gonna’ have to disagree with there being no point to it. The analysis has been accurate all along. It’s a real time saver and I think Denon should consider including it.
In this case he is correct, unless you have a USB stick that is fast enough to keep up with the sustained IO. If the drive is the bottleneck running two operations at once will slow analysis time. A really good why to test this is to take a few really big movie files take one file at a time and put it on the disk vs 4 files on the disk at once. If disk IO is the bottleneck putting one movie on at a time for all 4 movies will be a lot faster then trying to put all 4 movies on at once. Basically you are creating a huge que-ing process where the CPU has to wait for the storage before continuing.
I would agree with you though, there is no reason to have this feature in there. especially for those that are analyzing on a local SSD then transferring the files to USB or whatever.
I would also like to state that a class 10 USB2 drive would be able to keep up with the load needed to analyze 4+ tracks at once. Most people just buy a drive and don’t look at IOPS the drive can handle though.
Also on a side note, all USB processing is done through the CPU so that is another reason why you would not want to analyze a large group of files on a USB stick at once as the CPU will take a hit for running this device,vs a PCIe device. I think this is why Serato makes you unplug a controller before you do a large analysis.
Just to be clear, I want this functionality for the Engine Prime software that’s on your laptop when transferring metadata to a flashdrive. From my experience using Serato for the past 12 years, 9 of which offered support for analyzing via hyper-threaded Intel multicore processors, I can tell you without a doubt, that building metadata on a flashdrive is faster with this functionality than without. Even w’ USB 2.0 drives.
that entirely depends on the size of your files. in theory usb2 gives you around 35MB/s. in practice however with the average usb stick it never goes that high and writing is much slower anyway. so it’s a simple calculation: as soon as reading the files of your anayzing jobs would need more than 35MB/s it’ll be actually slower than running single jobs one by one.
While you are correct, we don’t know for sure, and it would be nice to have a official answer or access to the people who could give us the official answer, you can take a look at the information that they have provided and decipher where they are going with the tech. You are not going to put in a more expensive x86 processor into the decks unless your end game was to be able to share code from other x86 platforms. Reading between lines again, if they are sharing code, and putting in multiple cores into the decks so that we can run multiple processes at the same time off one deck (as in having the ability to have two tracks playing at once), we are using SMT and there for Engine Prime will support the feature Blackavenger is asking about.
I think you nailed it. One of the things Luke is saying is that anything he and his brethren Denon ambassadors throw in as requests get in there really quick (days), which would support the “single platform” assumption you mention.
Engine Prime dedicates 5 processes/threads to quickly process each track, rather than only applying 1 process/thread to 5 (or more) tracks.